Morning Session
Chairman Pai:
Good morning and welcome to the FCC forum on open radio access networks. This event originally was scheduled for March, but i hope and expect it will be worth the wait. Thank you all for joining us, especially our conference participants.
We have quite a lineup. Among those speaking, all five FCC commissioners, experts representing leading technology companies and upstart firms, top talent from the academy, the commerce department director of policy and strategic planning, Robert Blair. The former ranking member of the house intelligence committee, Jane Harman, who now leads the Wilson Center.
For the first time i’m aware of, the U.S. Secretary of State is participating in an official FCC event. Usually, the FCC could only get a turn out like this by offering free food. Poorly, that is not happening here.
What is the big deal? The big deal is 5g. These networks will be embedded in almost every aspect of our society and economy, from businesses to homes, hospitals to transportation networks, manufacturing to the power grid. Over the past few years, the fcc has aggressively executed our 5g fast plan. This strategy teachers three key parts — freeing spectrum for the commercial market, promoting wireless infrastructure, and encouraging firewire deployment. This strategy has yielded big results. We have already completed multiple spectrum austrians — auctions that have repurposed huge swathes for 5g, and have seen record-breaking investments in infrastructure essential for next-generation networks. But our focus is not limited to promoting networks that are strong. We are also committed to making sure they are secure. For years, u.s. Government officials have expressed concern about the national security threats posed by certain communications equipment providers. To address this concern, we have aimed to protect the integrity of the communications supply chain — the process by which products and services are manufactured, distribute it, sold, and ultimately integrated into our communications networks. Specifically, the fcc has prohibited the use of money from our universal service fund to purchase or obtain any equipment or services produced or provided by companies the fcc determines pose a national security threat — namely, huawei and zte. We’ve also started a process to identify and catalog insecure equipment, usf funds in communication networks, so we can remove and replace it once congress appropriates funds for this purchase — this purpose. Looking to the next generation of wireless technology, much of the equipment at the heart of 5g networks currently comes from a few suppliers. Three of the most prominent are sweden’s erickson, finland’s nokia, and south korea’s samsung, but the largest is the chinese way — huawei. Carriers worry that equipment could expose them to security risks. Huawei’s marketing power, aided by the chinese commonest party’s subsidies, might seem to make it the cheapest and best, but u.s. Intelligence law requires companies like huawei to cooperate with and keep secret state intelligence work. The law also creates opportunities for chinese intelligence agencies to compel access to an organization’s facilities, including communications equipment. In short, many are recognizing you get what you pay for, and that the long-term cost of using insecure equipment is likely to outweigh any short-term savings. In addition to the security issues, carriers may be concerned by relatively consolidated marketplace. Some have told me, both here and abroad, that vendor diversity is useful in terms of price competition, avoiding the lock-in problem, and assuring a backup supplier, among other things. Technological innovation has opened a new path to address these concerns. That technology is the subject of today’s forum — open radio access networks, or open ran’s. They can transform 5g network architecture, costs, and security. Traditionally, wireless networks rely on a closed architecture in which a single vendor supplies many or all of the components between the network base stations and the core. But open rin — ran’s could disrupt this marketplace. We could see diversity and suppliers, cost-effective solutions, and could see the keys to security in the hands of network operators, as opposed to a chinese vendor. All of this might explain why some telecom companies are beginning to develop and deploy open interoperable standard based in virtualized radio access networks. As an added bonus, many of the leading firms in the open rin space are based in the united states or in countries aligned with our vision of 5g security. How this marketplace will evolve is hard to predict with certainty, but here is what i can say with confidence. Innovation and competition make for a stronger, healthier telecom ecosystem. That is why so many are excited about opener in’s potential — open ran’s potential. In this form, we want to encourage innovative solutions. One way to do that is by convening top experts in the field to discuss the benefits of open ran, the challenges of implement it, and the lessons learned from deployment’s thus far. That is exactly our task today. I’m grateful for our impressive roster of public and private sector experts for taking the time to discuss with us the current state of open r.a.n. -Related technologies and the path ahead. I want to thank the fcc staff who helped make this event possible. Among the 31 staffers who brought us from concept to concrete — ken baker, connie days — diaz, charles, robert, wes, michael, becky, janet, and others from her bureaus. Evan and shawn, and as always, jeff friordan and the av team. It is my honor to present this message from the u.s. Secretary of state, my friend, mike pompeo.
Sec. Pompeo:
i first want to thank chairman pai, a fellow kansan, for inviting me to speak . As secretary of state, i’ve spent a lot of time addressing the china challenge. The chinese communist party is leveraging its technological prowess to erode freedom and democracy at home and around the world. That is why the united states has called on our allies and partners in government and industry to help us protect people’s freedoms and data from the ccp. I’m very pleased to say that a growing wave of countries and companies have and — banned h uawei and chosen clean vendors. It is about 30 countries that are clean countries, and many are the world’s biggest telecommunications companies and nations. They are using technology that makes them clean telcos. We have also rolled out the clean network, a coalition of like-minded countries and companies protecting cybersecurity from the line actors. Network maximizes connectivity without risk from un-trusted enders, and stops ccp censorship of americans. Today, each of you will determine how you or your company can join or support the clean network as well. On a similar note, i support the fcc and chairman pai’s strong efforts to free as much possible spectrum for 5g as possible, and to do so as quickly as possible. I hope each of you will do that. When americans compete and innovate, we will win. Freeing the spectrum will drive a fast build out of our own 5g networks, and will stimulate innovation that will drive economic both to push 5g technology into everything from when you factory floors to telemedicine to autonomous vehicles. And it is critical. It is critical for our national security and that of other countries. We want our friends to choose trusted 5g vendors for their network needs, not vendors tied to the chinese communist party. The world does not want china’s communists hiking — hacking into self-driving cars or medical tools. If innovators create open networks and deliver the best solutions, we will help citizens around the world avoid these threats. But we need to be open. The spectrum needs to be open, and we need to do each of these things if we are going to achieve any of these goals. Look, it is very simple. Technology must advance freedom. That is what we believe in the trump administration. We hope you will join us in these efforts. Thank you again for having me. Good luck.
Chairman Pai:
my thanks to the secretary for those are marks. And now we are honored to hear from mr. Rob blair, who works at the commerce department as the director of policy and strategic planning. I’ve seen him hard at work on these issues here and abroad and i am glad he is able to participate today. Mr. Blair, the floor is yours.
**Mr. Blair: **thank you, ajit. Can you hear me?
**Chairman pai: **i can, yes.
Mr. Blair:
there were good. I want to thank you for the invitation to speak today. You and your staff have been a pleasure to work with and i know the work that went into making this event happen. This is at a critical time for us to get together and talk about this issue, and i’m very much looking forward to discussing oran a bit. As you all know, i think it is going to be a key part in promoting more secure and democratic telecommunications systems. Your comments were perfect. They are a great introduction to the situation we find ourselves in and truly the promise of oran and a more diverse telecommunications infrastructure. Before i turn to oran, i want to note at a high level some of the work department of commerce is doing to support the national strategy to secure 5g. In the station, all of us, continue to strongly defend our economic and national security. One big focus is securing our telecommunication system. Back in march, the president signed the national strategy to secure 5g, which i think is an incredible achievement. The department of commerce has been leading or supporting a number of work strands within this strategy. I don’t have time to go through all her work on this front, but i want to highlight a few key points. Nist, our national institute of standards and technology, is increasing focus and investment on advanced communications, identifying security gaps in the supply chain, and coordinating participation in standards development activities. Our international trade administration is developing a strategic remark for promoting — framework for promoting u.s. Tech companies around the world, working with the state department and with our trading partners, to achieve the goals of our national strategy. Ntia, our national telecommunications and information administration, is working to make more spectrum available for 5g. It also launched a program to supply information about supply chain risks with industry partners. This will be a key point on establishing clean paths and clean networks overseas, as sec. Pompeo just alluded to. Ntia is also working with international and private sector partners to develop principles of promoting a vibrant and diverse 5g ecosystem, which i will return to in a moment. I want to turn to the challenge we face, our vision for a more secure and prosperous future, and our plan to achieve that, working in coordination and operation with private sector partners. The current situation presents our nation and most of the world with a serious challenge. There is only a few companies at this point building 5g networking equipment. Some of those companies are subject to the control of authoritarian government that among other things are not governed by the rule of law, do not have independent judiciaries, and have no respect for personal privacy protections area some of these governments use our very freedoms against us. This situation is bad for countries, bad for companies, and bad for citizens. Countries cannot build strong societies when they have on trustworthy and insecure telecommunications of the structure. Companies — if a national monopoly is held by incumbents, there are such high barriers to entry that new companies are not able to break into the marketplace and provide better, less expensive products and services. Finally, the situation is bad for american citizens, and citizens around the world, who cannot enjoy new, innovative services that never reach the market, and risk having limited or needlessly expensive access to networks that would continue to be insecure and on trustworthy. Given these challenges, we have a vision in the united states government of harnessing 5g to increase prosperity while simultaneously increasing national security. This vision includes a couple points. One, improving our and our allies’ national security by building telecom systems without any un-trusted vendors, and harnessing the power for innovation that improves the function of these systems, but also delivers new and better products and services that improve the lives of people around the world. This outcome will be more secure networks, and markets that reward innovation and provide lower cost products and services companies and consumers. The real question at this point is, how are we going to achieve this vision? As part of the national strategy to secure 5g, administration, the department of commerce, and specifically ntia are focused on promoting vendor diversity. By increasing vendor diversity, we can reduce or eliminate a reliance on un-trusted vendors and increase competition in this important marketplace. Based on stakeholder feedback, we know industry is eager to build and deploy open and interoperable 5g networks. We at the department understand the move toward open architectures can and should be industry-led, but the government has to help. One way we are helping is working with like-minded governments to identify ways to collaborate. A second way we are working is to build off the success by exploring the develop of principles for open and interoperable networks. These principles could identify further ways to foster vendor diversity and infrastructure resiliency, as well as ensure open and competitive markets. They could drive rapid innovation, collaborative design, and interoperability, and facilitate the transition from our current situation and the challenges it presents to open networks and benefits to consumers, countries, and companies. So, with that — if i could ask everybody to mute their lines, i’m getting some feedback. I like to thank everybody for your work at attention on this critically important issue. This is one of the great challenges and opportunities for virtual conferencing, and i think we are working through some of the kinks sometimes, but we have a chance right now to build a telecommunications system that will protect the principles of freedom and openness that this country was founded upon, and i’m looking forward to continuing our partnership with the fcc, with the rest of the ministration, because we really are at a seminal point in the development of these processes. A gene — ajit is a great leader on this and it is a pleasure to work with him.
Chairman Pai:
thank you for rolling with the punches and for those insightful remarks. It has been an honor to work alongside you here and abroad, as i mentioned. Now it is my pleasure to introduce a good friend to those in the technological innovation space, the honorable jane harman. The former ranking member of the house intelligence committee, leader of the wilson center, and a dedicated public servant who has dived into these issues for a long time and has worked on a bipartisan basis to advance the public interest. Madame congresswoman, the floor is yours.
Ms. Harman:
i appreciate the introduction and the 20 minutes of remote learning i have just done on this very complicated and fascinating topic. I’m jane harman, recovering politician, think tank ceo, and digital adapter. Let me salute you, ajit, for your leadership at the fcc, for your friendship, and for your bravery in inviting this grandmother of eight to address a hot tech topic. I hope my grandchildren think i am really cool by doing this. Even as a digital adapter, i am not a total novice on ig. The open ran concept, or even the work of the fcc. I can remember when 3g and 4g were just joining the alphabet. My congressional district, as some of you may know, was the satellite-producing center of the universe in los angeles, just south of santa monica, so understanding communications satellites was a necessity. And by the way, serving on the intelligence committee was, and i was proud to serve therefore eight years, in my 400 years in the united states congress. As a member of the house intelligence and commerce committee, where michael pao — mike pompeo and i overlap for one month before i left to run the bipartisan oasis at the wilson center — my colleagues and i worked to improve radio interoperability after 9/11, for first responders. By the way, at the wilson center, our science, technology, and innovation program has just developed a micro site with deep dive interactive’s and briefs on key 5g policy challenges — both security and economic. Check it out at 5g.wil soncenter.org. I’m not kidding. There will be a test at the end of my remarks, so you better take a look. Here are my remarks from my experience in congress in conversations with the many geniuses at the wilson center. First, and you have heard this from other speakers — the u.s. Should be properly proud of her history of innovation, especially in technology. We sequenced the human genome, and today most anyone can access that technology for a few hundred dollars. We have not always been and have not needed to be — we underscore that — have not needed to be first movers. Europe beat us in inventing the jet engine, but america was able to eventually lead the industry with efficient production and distribution, which is why the word race in the context of 5g is not very helpful. The question is not who is winning, but how we can put ourselves in the best possible position to take advantage of current and future economic benefit while addressing security concerns. I totally agree with secretary pompeo — associated with this fifth generation of cellular networks. We have major advantages, the world’s best academic training, significant sources of funding in industry, well-versed in research and distribution. Second, it is boring to talk about setting standards, but they matter. Firms need an industry standard so they are not at a competitive disadvantage. Plus, if we are not setting the standards, someone else is. We were traditionally the leaders in sitting in the front row of all the meetings, and now we are not. It is a big disadvantage which will only get bigger if we don’t change course. Third, policymakers need to think outside the box. That is why the fcc is opening today’s session on open r.a.n. I first saw a data center or vpot built in the cloud — i actually did this — on a study tip with congressional staff of the west coast three years ago. I have some serious privacy issues with ai voice assistants, but was astounded to see one take a command and watch the code build and run. Even my grandchildren cannot do that. I admit i did not really know what i was watching at the time, but later it became an aha moment. The open r.a.n. Concept is not that different — is not that different. Rather than needing hardware, we can take advantage of software to do many of the things we used to need screws and motherboards to do. In some cases, software is better. And we in the west to software very, very well. In addition, the open in open r.a.n. Moves 5g away from proprietary systems dominated by a handful of companies to a diversity of hardware and software players. Put simply, it allows carriers to shop around and then plug and play. While some worry that more software means more problems, others suggest that open networks will mean more eyeballs where it counts. I have always found that more eyes and more brain cells are better. It also means that we even the competitive playing field and give more startups a chance to create the next disruptive innovation we never expected. Fourth and finally, the elephant in the room — michael pao mentioned it — china. Our discussions should not only be dominated by concerns over chinese companies such as huawei and cte. If it’s a broader. Open r.a.n. Is not a geopolitical silver bullet. It is one of a broader set of steps necessary to shape a robust and secure telecommunications ecosystem. And there is more to do. But in our discussion of open r.a.n. And future conversations, we can and we should dig deep into the logistical and technological weeds. We have to answer questions about integration, roles, and responsibilities ross a diverse network of players. We will also need to figure out — and i think this is a big part of the day — how to help u.s. Companies enter the market with open ran — r.a.n. If possible. Open r.a.n. Is not the end of the 5g conversation, but it is a great getting. It is a huge opportunity for the u.s. And the west. Let’s seize it. Thank you very much.
Chairman pai: thank you so much, jane, for taking the time. I suspect you are going to be teaching your grandkids a fair bit more than you are letting on. I appreciate your willingness to participate today and look forward to executing your vision of a more prosperous, more secure future. Now it is my pleasure to turn to a friend to all of the commission, commissioner jeffrey starks.
Commissioner starks: thank you, mr. Chairman, and thank you for organizing this important discussion. As many of those gathered here today may recall, the united states was once a worldwide leader in the telecom network hardware. Companies like motorola led the world in the design and manufacture of telecom equipment, and for radio antennas. Unfortunately, american leadership in this market has disappeared over time. While many factors were at work, american telecom carriers were eventually left without a reasonable domestic option. The timing could not have been worse. As the chinese government’s made in china 2025 strategy tilted the field toward its own manufacturers. Chinese government supported artificially and lowered huawei’s and cte’ — zte’s prices and undercut competition. This was not free-market competition, but part of a strategy to leverage economic power into geopolitical dominance. Throughout this unfair advantage, the equipment produced by these companies has become pervasive around the world, and even reaches u.s. Networks. This is not just economic gamesmanship. According to our intelligence agencies, in exchange for the subsidization, chinese companies have allowed backdoor access to state agencies, enabled functionality for our networks’ disruption. As a result, the technological foundation of our communication networks has been weaponized. Congress and the fcc have recognized problem of untrustworthy equipment in u.s. Networks, and are working to assure its removal and replacement. Open rem networks are part of that solution. Most exactly a year ago today, i published an op-ed advocating for the development and use of software able to virtualize 5g infrastructure to replace suspect equipment. Our country has long been a technology leader in software and wireless technology, growing our capability to make secure infrastructure makes sense, promote the security and economic standpoint. And so we need to invest in this technology, and we must do some deep and proactive thinking on the best policies to effectuate our goals of promoting secure telecommunication networks that benefit our shared future and get the best value for the american taxpayer where we need to rip and replace insecure chinese men. Here is a new idea. I recommend that we explore the each rip and replace carrier rebuilding its network required to consider solutions offered by him oran provider — by an oran provider. That would achieve many goals including global competition, capitalizing on u.s. Software advantages, accelerating the development of oran as a product and as a business case, and allowing for alternative vendors to enter the market and offer specific network solutions. While no carrier should be forced to adopt it, it would encourage carriers to consider technology that might otherwise have been overlooked. Oran holds tremendous progress. — Promise. Its growth could advance american technological leadership, enhance competition, and reduce our reliance on foreign vendors across the world, all while bringing down replacement costs. It deserves serious consideration, and i look over to the conversation we are having today. Thank you, everyone.
Chairman Pai:
thank you very much, commissioner, for your remarks. Look forward to working with you on this and other issues under the fcc’s bailiwick. We will kick off the first panel, an introduction to open interoperable and virtualized networks. We’re starting with a hanging in — with a bang. We have the general manager for the 5g infrastructure management network platform group at intel, the chief and ministry of officer of record 10 movable — raukuten mobile, a member of the technical steering committee, and the executive vice president, chief of strategy, and product manager at lts star — altiastar. I have to offer a disclaimer for this and subsequent panels. The presentations you are about to hear were voluntarily submitted by the panelists and do not imply any expected benefits to the presenters. By hosting this conference and providing a forum for these presentations, the fcc does not thereby endorse any device, manufacturer, vendor, or company described in the course of this conference. With that legal requirement out of the way, i thought we would jump right into it with a foundational question, which i will post first to caroline. Let’s start by defining the topic of today’s event. What is the radio access network? What is an open radio access network? How does this oran or open r.a.n.
Ms. Chan:
Thank you, chairman pai for participating. We will say what the traditional architecture s. It mostly consists of purpose built solutions on 3g p.p. Standards with custom similar to optimize network functions. Both software and hardware are verified, integrated. Typically delivered by single vendor. Before i go to open r.a.n. Let me touch on virtualized r.a.n. Solution. About 10 years ago a group of ecosystem intel part we started developing r.a.n. Functions that ran on software on standard serveers like today in some cloud and implement t. — You i.t. All software was s.d.n. Allowing all additional facilities. Some as r.a.n. Alliance which cat cattle is the co-chair. Between various components of the radio s.s. Network and referring to as open. And you have several flavors the traditional network still and you have the virtualized and going forward you have the open r.a.n. Under the o-ran specifications. Back to you.
Chairman Pai:
thank you for setting the stage for this conversation. I will turn next to sachin katti. One narrative we player is there is not a u.s. Based alternative and some played that point as well. Can you discuss how o-ran might diversify the marketplace and leverage america’s traditional strength in software?
Sachin:
as carolyn said it is pushing the envelope virtualizing the radio network and opening up by specifying open interfaces. This will open the door for a tremendous amount of hardware inno innovation. By definition virtualized we can run a variety of hardware and essentially an app, a software on top of it. This plays to american strengths on hardware innovation. Look at the innovation in cloud and a.i. There is a mushrooming of start-ups trying to have similar. The same innovation will apply to the network because we can virtualize it. They don’t have to worry about the underlying hardware. Because there are open interfaces i can mix and match radio similar from window was software from different windows. Previously it was hard for a new hardware innovation to come in and to get your foot in the door you had to not only build the harsimilar — the hardware and now opening up you open the door for innovative hardware to focus on that, get that right and move the market in that place and you will be able to get customers because they are adopting open interface us. The transformation happening because of that and open interface between the radio hardware and rest of the software we will see an amazing growth by station hardware and that is a tremendous opportunity for american industry to take advantage of.
Chairman Pai:
Certainly an exciting forecast for the future and one we would like to secure as well. I know one company in this space that has been actively doing not just thinking has been rakuten and we are pleased to have tareq. I know there was a time when huewei was set to be your vendor. Why did you go in a different direction?
Mr. Amin:
thank you for inviting me to speak and i want to congratulate you on an executed auction. We have been watching this from japan as well. I think to understand the present it is important to understand a little of the past. I had an opportunity in my life and career to work for a very innovative company that wanted to change the fabric of society by disruption that they could do in connectivity. I moved to is r india i learned the value of one u.s. Dollar. The average revenue there was one dollar and i learned importance of diversity in supply chain and criticality of having the right component vendors. Then i met the chairman of rakuten and the bill r pillars were three big areas. The first area i call it courage to explore which is an explore. Do we continue to build networks in the same fashion that it was built for decades. Huewei was i think almost was ready to be selected as a vendor. Or do we look in a brave world in which cloud is part of the essence of how we think of future networks. That is the first pillar, courage to explore. To enable there we realized the united states continue all the necessary ingredients. You just need the right for lack of better word shift to put it together. We invited world class leading companies. Qualcomm. Intel. Cisco and we decided this willing partners could innovate into the space. We didn’t want to take the risk and our chairman and c.e.o. Decided if there was any risk that could be will by selecting huewei that is something we would not want to do. I have one slide i would like to share with everybody. Can you display that one slide? When we put the network in play, i think in the beginning a lot of people will many doubts about the ability of sunday and architecture open r.a.n. Can it work and should it only be a technique knowledge secluded to the network and i had a picture in my office that shows the two rockets that spacex has landed on a barge and i said if a human being has managed to put rockets on a moving ship we certainly could do an innovation around cloud technology. What you see is a picture. What we committed to the equivalent of fcc in the u.s. Is a build number we have doubled and every single site in tokyo today is sitting as a cloud november open r.a.n. An platform not only have wes realized quickly benefits and evacuation of such architecture but because of simplicity our ability to roll services and deliver is unparalleled. So today we have exceeded north of 6,700 imagine row bay stationless and — macro bay and by march of 2021 we will be 70% pop coverage a and i’m carrying double the traffic in japan but what excites me most is security. This is what i’m so thrilled about. When you have full visibility the possibility of what you can do with transparency and transparency end to end, not just software but where do the components come from is an unparalleled opportunity for the industry. I want to commend you personally for the leadership that you r, the united states government are doing. We are through about this future and eager and willing to find a way to collaborate and support and create the linkage needed between the japan and united states to advance this technology and show it is real and i invited everybody to come to tokyo hopefully we the covid situation is out of us and i would love to show you the vision we talked about being a reality in tokyo.
Chairman Pai:
thank you very much above the for the hospitality last year and your words just now. I will like to focus on one of your comments concerning security. Do any of the rest of you have a thought as the implication of r.a.n. For security. Do you find it potentially more secure architecture?
Mr. Amin:
The other comment is open interface means that i can pick the best of solutions for each part of the network and if you go with the innovated architecture you trust that one vendor and they have to get it right. Here you have that ability to test and inspect and you have that ability to replace when you find something not working. By definition given the ino vase that we have seen in the rest of the world in the cloud how quickly things can get upgraded bringing the same cloud flavor and replace and upgrade in an agile fashion that will lead to a more secure and clean network than today.
Chairman Pai:
terrific. I did want to turn to tear tear. One thing from tarek’s slide about tokyo was the nature of deployment. We have incumbent networks already built and we are seeing the emergence of stand alone networks discussed and deployed. How dots o-ran play into with it is a brown field deployment or green field deployment.
Mr. Maupile:
good morning, everyone. It is something we have seen in other industries where you have innovation and you have entrepreneurs who want to take advantage of the innovation and it has created an opportunity for entrepreneurs as we have seen in japan and same innovation adopted in the u.s. And we expect to see the same in other markets such as germany and europe. So, there will be a few. But the real opportunity for open r.a.n. Is brown field meaning operators with an existing network and what is very interesting with open r.a.n. Because it is built on the stand out so you can insert open r.a.n. Not just even for 5g but take advantage in 4g networks and there are different motivations to do that. We call it green field or brown field. Any operator with an existing can do green on brown. You can insert. There is a learning curve. It is not easy. I don’t believe building green field is easy be task but for brown field there’s a different set of skills. It is about adopting what we have seen with people who are familiar with the clouds and develop the i.c.d. And the brown field to move forward is really because the 5g requirements are going to require that the network being flexible, being more intelligent. That is what it does require this new architecture. There’s another motivation we have seen and hear from major global operators and i will call on one of them it is a thesity. It is a necessity because you need to have not just a diverse supply can chain, you need more choices and you need a strong supply chain. So we need to change the economics of this industry as has been mentioned before. Many companies have disappeared so as we are working with 5g and the next g we have a strong ecosystem. That is where for a brown field operator it is not just to take advantage of a new architecture that will drive a lot in readiness for 5g but also to have access to more choices, very strong ecosystem. This is the promise of open r.a.n.
Chairman Pai:
can i follow up on that in terms of green on brown or think deployment for some who have already built those networks. What do you see as some of the challenges of the widespread adoption of o-ran. ? Is it scalability or capital for you or anybody else that would like to comment. As part of that what do you see as the transition period? Is it a question of years to execute that sort of transition into a more texture d network or is there some other time frame?
Mr. Maupile:
Like any innovation which also some disruption with it, there is a lot of denial and skeptdism before you — skepticism before you see acce acceptance. What has been happening is that rakuten has been bold to build this network the open r.a.n. As proven that it does work. It is fantastic technology in a very challenging, demanding environment so the level of confidence we have been able to create is helping other operators to embark on this journey. That is what we see the major change in the last three or four months. With the covid-19 we are all experiencing is that many operators now not just evaluating but planning to deploy open r.a.n. In the network.
If you ask me what is the last two or three years learning experience in rakuten for us it is not a green field. It is more like green mines. We focused a lot on the organizational fabric and structure. By the way if you converse with web scale companies, amazon, facebook, google you will find the same caliber of talent we focus on building such a platform and architecture. Maybe this is a challenge. You have asked what are the challenges on scaling this. I said scaling the organization is a challenge and the second challenge i would see and this is something i have talked to various u.s. Agencies because the question was asked what is the role of government that could play in enabling such disrurpgs and i look at the — disrufpgs and — disruption and i see the united states having a momentous opportunity and talking about the leadership it had in motorola and lucent. I would say the u.s. Could seize that opportunity again if it rel realizes how to put the pieces of the puzzle together and you have everything you need and right allies to support this journey. I know what the challenge is to make this successful. For me it is about encouraging a good supply chain of components in silicon and we have enough good software companies that can help us out. So maybe the challenges i see is slightly simpler it is skill set organization and creation of a healthy ecosystem around silicon for the remote radio head and products that need to be created and most of the silicon would come from the u.s.
Chairman Pai: ok.
Mr. Maupile:
I want to add, green on brown. Recently verizon the most — one of the leading ones in the united states announced a pilot of 5g and network built on virtuization including the r.a.n. Piece. You can’t get more proven than that. It is surely in the united states green on brown so you can’t beat that. I think that is what everyone is trying to say.
Chairman Pai
If i could follow up, as i mentioned interoperating be — interoperability is one and how important is the standard setting process in establishing or enabling interoperability among suppliers and what part do the telecom project play.
Mr. Maupile:
They are foundational. They are the foundation to drive interoperability between different suppliers. We have seen that time and time ag again. Especially with the open r.a.n. Both organization o-ran alliance and tip play a key role. While the o-ran alliance define the difference the tip focus on total typing, integration and testing. T-mobile is hosting a tip 5g n.r. Open r.a.n.land right in your home state of kansas. So, o-ran alliance and at this point you announced a partnership to develop interoperable 5g solutions. One of the first is open ran requirements document. It provides a blueprint for the suppliers to test against and it is not just a one macro solution. People can add different modules to that. It truly leverages the fact that the u.s. Ecosystem has a proven software innovation and we have the cloud technology for a long time and it is the standard to put all of the hlego pieces together.
To add to that they are involved one of the lessons i want to highlight is this was it is foundational because these are new organizations that are giving operators a voice in the standard setting process. So historically the standard setting process has been dominated by the incumbent vendors and what do telco operators want to build on and build out the new 5g networks so both bodies were about how do we give operators a voice and get them to specify what would, how would they like the platform to change and i think that is they are critical going forward.
Chairman Pai
That is great. I will note with all due respect to silicon valley i’m tpnot surprised. Kansas prides itself on being on the edge of technological innovation. You said picking best grade components. That raises the question of integration. What do you see as some of the challenges and opportunities in terms of integration for you or anybody on the panel.
Mr. Maupile:
I would like to share a slide.
Chairman Pai: sure.
Mr. Maupile:
Great question, Mr. Chairman. The dilemma is the following. You want the benefit of choice without the pain of integration because historically you have been able to get one solution and have that deployed everywhere. I think this is where there’s an opportunity for innovation for u.s. Companies. If you look how cloud services are delivered today there is a developer you mix and match several services and build a new service in days. Why can’t that be true for telco? This is what we are trying to realize in the u.s. With dish and this is going to be a great point for an open r.a.n. 5g network in the u.s. Which will be at scale. But you can build an automated way of continuous integration and deployment. The best of cloud tools and have essentially an app store on site. So you talk about network this is collection of network functions delivered as software and each of them is an app. And really the vision is that you can build an automated deployment and let cloud technology manage the deployment of these functions and software across all of the different clouds in the country and figure out how to stitch them together and deliver a mission critical service like a 5g network. So this system integration shouldn’t be seen as a challenge. It is an option for u.s. Innovation because we have mixed and matched things a best of your knowledge of things and produce new products. This is the approach for 5g. That is what we are doing for dish and i think it will be a fantastic proof point for u.s. Innovation.
Chairman Pai: : tareq, were you wanting to jump in?
Mr. Amin:
as jurors to be — as engineers, it is fun. In the early days if there is any definition of what we went through as a really out of this world science project in the beginning the idea of taking a hardware based base band and all of a sudden you move this into a virtual machine and the magic you discover once you move that is what this industry doesn’t understand about what is the benefit of open r.a.n. Luckily as we understand it because we are living it every day and i say in every keynote we are scratching the surface of what is possible because the level of automation you realize when you discover how you could manage these v.m.’s and tomorrow it will be containers, it is elegant. Elegant when you think 4g and 5g insurance have the potential to be like managed wi-fi service. That is what i would consider to be the north star we should be working towards. We have taken an advanced, an entire telecom app store concept because that’s where we are headed and our approach to rakuten communication platform is driving there but we need collaboration and industry support and we have really solicited not only industry support but we think private and government could really work on this initiative hand by land to drive such — lhand and it is not evolution flair — evolutionary. Once people understand the enthusiasm and excitement i would say wait until you operate a true cloud network you will discover the benefits.
Chairman Pai:
that is terrific. One thing i find interesting about this space is in a way it exemplifies what clayton chri christianson called disruptive innovation and small firms to enter and driver some change. Thi thierry, altiostar is one of the smaller companies compared to larger ones. What do you see o-ran portending for smaller players and how does this enable their participation that would not be in the traditional network architecture space?
Mr. Maupile:
we consider this will be more diverse and be able to attract other companies to participate. It is all about innovation. And we are able to build the next generation of networks on software and lueverage, the cloud is creating a new entry point. We are removing barriers that have been in place from the traditional vendors. Having this open supply chain and it is open because we are using open interfaces we have confidence that what you get can be deployed with others and you can participate. It is clear not one single company can claim to have the best technology for all elements of the r.a.n. So it is important to unleash this innovation and it is very good news for this country because we have a lot of assets and intellectual property we can leverage and that is very much what you build with the new content, not just where you manufacture the product, it is is where the intellectual property is coming from. Now it can be real put in the market and it is because the network is becoming a platform. If you want to really realize the promise of 5g, you need to attract a lot of applications who can write and add their value on this platform. That is a very fundamental aspect of open r.a.n. Is becoming a platform for the operator and for application developers. It is essential for the success of not just open r.a.n. But for the industry and to deliver for all different 5g use cases and we are just scratching the surface but we need to create the foundation.
Chairman Pai:
it sounds almost like the mobile app developed on the 4glte. All kinds of apps sprung from that. You talked that a lot of these intellect all property and other properties are u.s. Based. I was hoping to pose to caroline how do you see open r.a.n. Solutions leading to american leadership in 5g and other wireless technologies?
Ms. Chan:
the extent of interface by r.a.n. Creating opportunities for innovation to come to the market. You alluded to the 1990’s and dot com. We have been investing in the diversification technology and we have a thriving e-commerce and cloud and i.t. Business. In the last 10 years intel for example has been investing in the open r.a.n. Technology over a decade. We surely started having a lot of intellectual properties like thierry mentioned the server technology and software innovation. Because 5g marks the convergence of innovation and computing tprorpl many verticals like health care, precision agriculture, retail and so on. We need software innovation to create great use cases and technology tailored to the requirements of the use cases. Intel, t-mobile, nasa, microsoft we are part of pga open inthough — 5g open innovation in seattle. That is a good example when you make the platform open or agile you are inviting the innovation that comes in. So while open r.a.n. Is not a panacea yet i believe it sets a really good foundation for leadership in 5g and wireless technology. Being, let’s not forget the great use cases are special so we need f.c.c. Leadership under the chairman.
Chairman Pai:
r sachin: if you look at the last decade there’s been a tremendous amount of innovation and application that run on top of the network. I can that the tphnext decade will be innovation on applications that are programming the net network. The network is allowing the application to deliver the differentiated like the sever driving car — self-driving car. That is what open r.a.n. Will enable because it is nining program ability and open independent faces so applications can program the network and deliver capability that you could not do before. This is critically important for the economics of believe 5g because if it is another unlimited data plan i think the praeurtsz will struggle — operators will struggle to make it work so how do you build services they can monetize. It is giving programers the ability to program the network and have new solutions and i think that is what open r.a.n. And open program believed and interfaces are a critical building block.
Chairman Pai:
in terms of programmability and do you see a role for adjacent technologies and artificial machine learning to play a role in optimizes these networks?
Sachin:
how do we build a self-driving network like a car. You tufrpld on this question on innovation earlier. This will lead to more complexity because you have more things to tune and put together and we want to obtain that technology like lurching technology but a.i. Won’t be that useful that applicable if the network is closed. It is two stepsment you open the network and have interfaces and programmability and a.i. Figures out how to program the network to deliver on this open disaggregated vision of the new services so a.i. Is a critical building block to obtain the new complexity and leverage the new capabilities for delivering the new services. Skwra
Let me idea one more comment to that. I want to give you a practice example of the power of when we talk about open r.a.n. And open independent faces in our network. We have teen r taken an approach there’s a smaller entity 11 people in san diego that provides what is called in this industry sever organize — self-organizing network in any traditional legacy network can you ask a big vendor with i independent great a small one it would be impossible. So that is open a.p.i. A he is we have taken a very powerful a.i. Machine learning algorithm that does, that auto configuration for the network the zero touch configuration and self-healing integrated through a.i. And 11 man on spectral efficiency gains into are unprecedented by being open. We have taken altiostar software through a.p.i. Integration through the apps that the industry dreams about there interoperability. You talk about the power that you don’t have to create everything by yourself. You just have to provide the platform and on top of the platform the innovations will come through many start-ups that exist already in the industry.
Chairman Pai:
that is incredible. I notice that we have about five minutes so i want to be respectful and close with the following question. This is a very exciting look at where we are going and where we are in some cases. What do you say as the role of government, the fcc in particular but not exclusively, in encouraging the growth of the open r.a.n. Environment?
If i may start, we have a responsibility as be a industry — as an industry to educate. People don’t know what they don’t know and we have discovered in the last few years that open r.a.n. As an emerging option needs to be explained. We need to explain what is open r.a.n. About and there is a lot of we are living in the world of fake news so we understand a lot of information is communicated that is not true and the real network being deployed is the best way to say this is real. What is important here is especially from a u.s. Point of view is we are looking at the government and policy makers to facilitate the adoption of open r.a.n. I have heard this many times when we are competing outside of the u.s. Everyone expected u.s. Being a short case. A lot of technology is coming from the u.s. Obviously we have a lot of trusted partners participating in the broad ecosystem which is great. But it is a unique opportunity for the u.s. To take advantage of the innovation. There is tremendous innovation not just from our staff but other companies. We can mention that v.m. Wear and cisco we are not able to participate in this part and now they are so this is important that we have the opportunity to show and dish is going to deploy, there is great. There’s a unique opportunity where the fcc is very much driving this transition for the rural carriers. Wouldn’t it be great that the carriers can take advantage of this innovation? And wouldn’t it be great with the dollars you use from the u.s. Taxpayer we can do more and we can reduce the divide. So that is the kind of thing the u.s. Government and policy makers can help us. And when we compete outside we need to have a level playing field. When we compete against untrusted vendors as a start-up we can only compete on our own merits that we have better technology. It is competing with innovation and we have to compete with innovation. But we have to make sure there’s a level playing field and we need the support of government to do that.
Chairman pai: terrific.
Sachin:
this may be in terms of what role the government can play it has an especially important role to play to follow up on how we make green grow in a brown environment. I think as new spectrum comes out and in have been so many successful and carriers will deploy new networks incentivizing them, nudging them to deploy open radio interfaces because one of the big khraels if you deploy it now you are stuck the next decade because it is hard to replace radios on cell you towers and virtualize because you gain the flexibility of making right choices and not locked in because similar locks you in for the next decade. So the government could play a great role to make sure these as we transition to 5g we get the deployment architecture right. That’s is terrific. I’m afraid we are out of time but i want to thank all of our panelists for lending us your expertise. I have learned quite a bit in the last hour and i’m sure there is were more under the surface of the water for me to follow up on. Thanks to each of you for taking the time to be with us today even in a virtual environment i thought the conversation was quite lively and hopefully some of the viewers thought so as well. With the that, we will begin the second panel after a short break. Stay tuned for the folks watching on the live stream. We will be back in a few minutes. Thanks to all of you again.
Welcome to our second panel of the day we have a great lineup of speakers for the benefits of deployment/driving innovation. Before that we will hear remarks from fcc commissioner brendan c carr.
Commissioner Carr:
thank you. It is great to participate in this event. Thank you to chairman pai for convening such great academics, researchers and professionals amount to secretary pompeo for the remarks and leadership you have been providing around the world in furtherance of clean secure 5g networks. This transition to o-ran is a big deal. When i think of the transition i think to one of the great unbundlings we saw which of course is the p.c. Market in the 1980’s. Some one born 1979 i remember those early days. You think become to that period of time i.b.m. Was dominant in the p.c. Market and in a way that was completely vertically integrated everything from the hardware to what we think of as o.s. To software was packaged. With respect to the i.b.m. Of today which is a different company and we will hear from some of them to come the up shot of that back then was a product that was clunky, expensive and really was only affordable from highly profitable corporations. In the 1990’s microsoft came in and completely disrupted that vertical integration. It unbundled the p.c. Market and separated the hardware from the o.s. From the software. Flash forward we know what that produced. It dramatically drove down the cost of laptops and enabled competitiveers to enter from dell to gateway and for consumers could afford to buy a p.c. That unbundling of the p.c. Market coincided with the build out of intel infrastructure. You put those two together i think that is why we saw americans getting on the internet for the first time in the 1990’s. That same dynamic and type of unbundling is about to take place in the wireless infrastructure market. So, for those of us from policy perspectives i think that is attractive for three main reasons. One is improvement in service. The other is improvement in security. And the other is jobs. From the service perspective this unbundling means that everything from consumers to wireless carriers themselves are going to be able to mix and match components and pick the best ones that meet their needs. We will see competition across what used to be a vertically integrated market and that will improve performance. We are seeing pieces of that today. Right now there is one rural carrier that has been working with a o-ran vendor and seeing a through put increase by about 40% and seeing their costs cut in half with a virtualized core plus o-ran. That is some of the reasons this will be important. We have been working at the f.c. For a number of years and i have been leading the effort to reform the wireless infrastructure resumes to make it easier to swap out up grade and add no equipment and i think that will coincide with this desire to build out the o-ran architecture. The other piece is jobs. By unbundling we open the marketplace to smaller providers, ones that are expert on software and that happens to benefit a lot of u.s. Companies which i think is a great thing particularly when we are talking about build-outs that might be supported by universal service dollars. A lot of the companies are here from altiostar to wireless, many others. So i think this unbundling will help spur investment in high paying jobs in the u.s. The last piece is security. More than two years ago the f.c.c. Launch the the effort to make sure universal service dollars are not going to insecured network equipment and smaller rural carriers felt they had no option on expensive bespoke pieces of huewei gear and that was relatively affordable including subsidy assist from the communist party and what we conceived as security threats. The question is how do we make sure they feel like they have more choices and more affordable choices for secure equipment and that is where a ran — o-ran will help. It is something outside vertically integrated insecure huewei gear. The challenge for us and opportunity from chairman pai is to accelerate this to o-ran. We want the private sector to have confidence they can put this equipment and software into their networks. One backstop we have in mind is the rip and replace we may engage in costing one to two billion dollars to have all subsidized huewei gear out of the network and we as policy makers need to accelerate the transition to o-ran so when the smaller providers feel like they have a choice and o-ran vendors have a fair shot at competing for their business. That is one back sstop and we are really excited about the disruption that will come from unbundling the network and look forward to hearing from the panelists today. Thanks.
Thank you, commissioner carr for those remarks.
We have a great lineup today. Panelists please turn on your cameras and i would like to introduce them we have cristiano amon eqqualcomm, craig farrell i.b.m., peter gammel from globalfoundries, morgan kurk from comscope, diane rinaldo from the open ran policy coalition, john roese from dell, mariam sorand from cable cable and soma velayutham from nvidia. If you would like to read more about them, you can go to f fcc.gov/events and scroll to the bottom. We want to take time to focus on questions today. Let’s kick off. We have heard a lot about o-ran in the previous panel and we got an introduction to the topic. Morgan, what changes does open ran introduce into the current business model for telecom infrastructure?
Morgan:
thank you for having me here today. The current business model has not changed in at least 20 years fundamentally. This is limited the opportunity for innovation and global competition, and in the current system competition is really limited to a once a decade competition during a transition of a cycle between 3g and 4g, 4g and 5g. After which the small winners, there are usually two are locked in for the full cycle. This concentration provides a level of stability, but it has a protracted and incremental limit to innovation. There’s little opportunity for cost reduction. What open ran would really do is change all of that. If i could have my slide pop up it is probably easier to discuss based on this. What you see in the slide is effectively a number of interfaces being defined which allows for competition in each of the subcomponents. Whether it is the central unit, distributed unit or the remote unit or active antennas. In addition, what you see on this is an opportunity to change something which has previously been monolift thick spwaoeurbl in hardware or software from a single company to being open. That is really the key to this. You can have a variety of competitors competing both on cost as well as on innovation in each of these components. As we know, when companies are able to compete they are able to innovate more rapidly. I have here a r an example so people can see of a comscope remote unit with integrated active antenna. It is an indoor remote but it fundamentally is compatible to o-ran and would work with ot otherer party’s d.u.’s or distributed units or centralized units thus enabling a company like comscope to compete in an area we are good and have somebody else compete in an area they are particularly good at. When you have this full industry that is competing in all of these various areas, you will end up creating a more stable and fast moving and more dynamic infrastructure that ultimately not only benefits the consumer but the industry and u.s. Manufacturers like comscope. I like to say that if companies in the united states are given the opportunity to compete on an open playing field like o-ran, we do very well across the globe.
Thank you, morgan. Diane, i want to move over to you. Morgan talked about diversity, competition, his slide shows how when you disaggregate the network equipment from the core to the edge, from hardware to software, you necessarily are going to introduce new names into the mix, maybe names not previously associated with network equipment. Your organization the open ran policy coalition has a diverse member that represents that dynamic. What does your organization say about the diversey of the radio access market now as opposed to 10 or even five years ago?
Diane:
it is always to be the first one to leave it on mute. Thanks, i appreciate the opportunity to be with you. Just to say to the chairman’s remarks i hope we can all have lunch at the fcc building. It would be good to connect face to face. Absolutely, we are all here today and you have assembled such an impressive group of government fox and — folks and private sector to drive diversity and ensuring a more robust so supply chain. Wea first launched — when we first launched play 52 was to educate policy. Rakuten, dish, they are building out their networks to o-ran specifications. It is here and occurring. So what can we do as private sector companies. How can we work with the government to promote policies to ensure that we have scale a as we continue to promote the innovation that is driving at the component level. While open ran is a specification it is not a technology, the technology and technical evolution will occur at the kpoefpbt level. The radio, the hardware, software. So it is great to bring together such a diverse group of folks to talk about what is happening today and what we foresee for the future.
Evan:
speaking of names, john, your company as a household name, but not necessarily because of oran. Well i hope the majority of americans know about oran, they may not think of oran when they think of dell. You are here for a reason and you have talked about the changing makeup of this panel. How do you see these changes evolving?
John:
hopefully, people are familiar with dell. We are the largest infrastructure technology provider in the world. We have a large service provider business, but we are specifically a player in the legacy of telecom, the radio side. What attracts us to this new ecosystem — and i will talk about that in a second — is that it is not just about building the next generation mobile network in the image of 4g, which was primarily driven by mobile broadband consumer consumption of packets delivered over the air. It is to us the digital fabric that is ultimately going to sit underneath almost every digital transformation, whether it be smart manufacturing, transportation, health care, education ecosystems. Those are areas dell has been extremely visible and present in, helping those industries transform. When you think about where they are going, their users or distribute it. Data is distribute it. Connectivity is critical. If you look at 5g, the full promise of it is the fabric underneath. The challenge right now is if you look at other big i.t. And cloud transformations, the ecosystem in 5g is relatively small. If you compare semiconductor companies in the ai space versus the number of new semiconductor companies in 5g, it is anemic. If you look at software companies, there are great small startups go — doing great work. If you compare it to the cloud ecosystem, it is one 1000th of the companies participating. For 5g to really hit its promise of open, integrated architecture that is this underpinning of the digital transformation, we have to attract the entire cloud and i.t. Industry into this industry — semiconductor companies, system companies like dell, software startups, investment in new ecosystems. Academia has to scale. Across the board, whether it is supply or demand, the ecosystem has to get larger. One of the things we like about open ram — r.a.n. Is by allowing replace ability in the architecture, it introduces white space where innovation can happen, whether it is a low-level interface or a high level abstraction, like virtualization. The minute the slayers exist, it attracts technology companies to try to find a better way to create value. For us, it is a strategic decision to be part of the 5g ecosystem. The entire i.t. Industry depends on it. It is also an opportunity to help bring that ecosystem together and turn it into outcomes, which is something dell does better than those people in the world.
Evan:
thank you, and based on your background, it looks like you have literally merged with technology, so congratulations on that. Speaking of, we have mariam joining us from a soundproof studio, a videoconferencing super move. You’ve been following this ecosystem for a while, so i figure it would be great to piggyback on some of the things john said. When we hear about open r.a.n., we almost necessarily associated with 5g cellular networks. But wireline and wireless are inextricably linked, and there is a reason the fcc’s 5g fast plan has such an emphasis on fiber. What do wireline networks, whether cable or fiber or other wireline networks, see in oran?
Mariam:
hank you for inviting me to speak at this panel. It’s certainly exciting. First, as you correctly alluded to, a wireless network is what transports traffic, the total network is what transports traffic from the wireless part of the 5g radios to the 5g core. Therefore, the network must be looked at into and — end to end to design for future use cases. Take low latency, which is one of the pillars of 5g. That needs to be viewed and to end — end to end. Let’s not get stuck on 5g. 6g is around the corner, and other wireless technologies. So in general, when we are looking at the future network and the future of innovation, we must take into account the convergence of the wired and wireless networks. So what open r.a.n. Along with virtualized r.a.n. Does is to come as my colleagues have started — stated before, to decouple software and hardware, break down the radio components into smaller pieces, and therefore this makes it easier to bring together and interoperate the elements of a wired and a wireless network, through the use of common hardware platforms. This is exactly where our members will play a critical role and will leverage open and virtual r.a.n. To deploy wireless very efficiently and surgically on top of their existing, very robust wired networks.
Evan:
we have heard why open r.a.n. Might be good for the network. The title of this panel, of course, is benefits of innovation. What does that mean for the consumer? What is the end-user get out of this, as well as the network operator? Soma, i would like to turn to you. One of the benefits we associate with 5g is network slicing, the idea you can have multiple networks within a network — maybe one network for gaming, one for smart cities, one for public safety, all with the same radios producing different networks, different needs in terms of geography and density. To open networks or particularly the virtualized networks, make network slicing easier? What does that mean for the consumer?
Soma:
thanks for having me in the forum. Definitely. As my previous speaker spoke, network slicing basically is about software declining operations, right? Virtualized r.a.n. Makes it easy because you are basically creating network functions that express software building blocks. And they are abstracted from the hardware. But the three components that make up network slicing — one is you need to have software unbundled from hardware. Second, you need to have really good orchestration. You need to be able to make that orchestration. It is already unbundled, so you need to coordinated. The third thing is you need to really have very high compute to be able to orchestrate this complexity. It is almost like you unbundle it, but you need to be logically one, but physically disintegrated that is a lot. That is what network slicing is all about. Network virtualization — it is not that it makes it easier. It is a requirement. It is a prerequisite to slicing. Oran takes it to the next level because oran proposes certain interfaces, like 01, a1 — it is what is called a radio intelligent controller. These interfaces unbundle to another level. You do not have to be tied to a network vendor or someone specific, but you can really use the open interfaces and r.a.n. Deep learning to further orchestrate it. Let me recap. I think network slicing, in fact, virtualization is a prerequisite. But oran takes it to the next level. It unbundle the logic from rules based to potentially ai-based logic.
Evan:
another benefit the fcc looks for when they look at novation in networking’s spectral efficiency. The fcc has pushed aggressively to make as much spectrum available for commercial use as possible. Chairman pai today touched on the 5g fast plan. It is no secret there is very little greenfield spectrum left, spectrum not currently being used. That creates challenges. Part of the technologies you just described help with spectral efficiency through deployment of things like software-defined networking and virtualization? The goal is essentially to squeeze more juice out of the same lemon. How do we do it?
Soma:
i think that is probably the best expression of structural efficiency, to squeeze more out of the lemon. I think the number one technology that is being looked at in 5g and beyond is massive my more — mimor, antennas. It is very promising. But at the same time, it does have front hardware requirements. Massive mimo alone is not good enough, because you need a lot of rules in order to coordinate that. The idea is if you can have virtualization and open interfaces to your antennas, if you are open, then the rules engines can be very advanced. Today the massive mimo is based on certain logic and rules, probably understood by a handful of people, no more than 118 people in the world who understand beamforming and mimo coronation. But by opening this up, you are able to use ai and deep learning to do beamforming algorithms, to do multipoint coordination. These algorithms are very advanced. One analogy i would use is if you think about cellular — cellular is an organic east, right? It is more of a chemical reaction than a logical expression. It is like making beer with yeast. You don’t have logic. You need to be reactive and optimize your network based on that. That is very well-suited suited for ai deep learning. That is why we are very excited at nvidia because we have a huge heritage of ai and deep learning. Bringing this into telecoms and 5g, and getting the best of the spectrum, is fantastic. I would say two things. One is beamforming. One is massive mimo as a technology. Beyond massive mimo, using ai and deep learning to make the best out of that — that is the most exciting thing.
Evan:
that tracks with some of the things we heard about artificial intelligence being enabled by open and virtual networks in the previous panel. Craig, we have also heard a lot today about the importance of integration. Integrators will be unimportant part of oran in ensuring equipment and software from diverse suppliers can work together. Can integrators replicate the ease and convenience of essentially a network in a box that traditional companies offer? Do you see ibm supporting smaller startups that do not have the scale of the larger companies we are more familiar with question mark craig: a number of the speakers today have talked about the importance of the end to end integration of a solution, and this will be a critical component to the solution made up of a number of components from diverse suppliers. The whole idea of interoperability testing is going to be critical to seeing the widespread adoption of oran in general. Suppliers like ibm, we have a long history of doing end to end integration testing for services in i.t., so as oran opens up for us and we market, we look forward to bringing this sort of expertise, and we have entire practices around testing and certification. As a number of people have talked about today, we have security testing. May have practices around this sort of thing already. We really do look forward to this sort of opportunity to bring all that expertise to a whole new market. This is going to be very, very important for the integrity of end-to-end solutions to be maintained, because we all want component vendors to be able to innovate and bring the best of breed solution for their particular area. The idea of bringing best-of-breed solutions it is something ibm has been doing for a long time, so we want to do that. But it is critical that system integrators take the components, do the testing, and ensure the integrity of the end-to-end solution. Just an answer to your second question about will ibm be working with smaller companies to do this, the answer is absolutely, yes. In fact, we already are. There are a number of examples. Some of the companies on the call today, we are already working with. So yes, indeed.
Evan:
it is now my pleasure to introduce my co-moderator, charles mathias, associate bureau chief of the wireless bureau. He will have some questions for you all. Thank you.
Charles:
good morning, everybody. It is a pleasure to see you all. Thank you for participating. I’m glad to emerge from the cloud for this purpose. One of the things that we are looking at as a potential benefit for oran is innovation in the chip market. And we have heard a lot about the importance of chips. I’m wondering whether for the u.s. Economy and from a national security perspective, you have some comments and thoughts about the role that chips will play, the chip market will play in oran. Also, how do you see this evolving as the technology develops? I think you are muted. It happens to us all. Still.
Christiano:
can you hear me now? I’m in a studio, so i had no control of the mute. Thank you for participating in this event. This is an exciting time for the industry. I want to go back to some of the remarks made by commissioner carr at the beginning. In this industry, what we have seen in everything that has got scale, especially when you think about consumer electronics or anything that has digital processors in it, it really scales to disaggregation. I would say that american companies did well overall by embracing innovation and disruption in the moment we have in front of us. An example was provided. I will give you another example very near and dear to us. When you look at the phone industry, the phone itself was vertically integrated. You look at companies in the beginning of cellular, like nokia, ericsson, motorola. Qualcomm created a platform that you could change computing from desktop to the palm of your hand. The silken conductor platform — you look at what the smart phone is today, and in many cases, the smart phone has been the first time many people on the globe had access to the internet cannot even measure, i think, the amount of innovation it created. The smart phone today is mankind’s largest develop and platform. But the internet went exactly the same way. If you look at the data center today, it is the same thing. I would say by american companies embracing innovation first, i think american companies have done very well. Look at the role of american companies in the horizontal personal computing platform, both in the u.s., as the semiconductor and processor, you looked at the phone space, you looked at the cloud, and it will be no different in network. The network is going through a process of disaggregation. I think what is great about this panel and the fcc initiative — it is about not only bringing awareness, but a huge drive for american companies and companies around the world to embrace innovation, and go into infrastructure 2.0. With that, i will start answering your question to say it is the mental for you to create a horizontal semi conductor platform for that to happen. We are excited about this opportunity, especially for a company like qualcomm that has been super focused in wireless communication. Assets that today you only see companies like — companies like erickson, companies like huawei — have on the digital side and the rf side. A company like qualcomm that has done that for the device space — it is our number one core competence, and we can create a horizontal semiconductor platform that touches on not only the baseband for digital units, but all the way on all of the solutions for remote units, and allows the entire industry to innovate by building new hardware and software to drive the transition to what we call the infrastructure 2.0. We are very excited about this. We have started early with this vision. I think rakuten, you saw in a prior panel, was the world’s first public network, which is based on vran and open r.a.n. Architecture enabled by our technologies. There is more coming from qualcomm very, very soon in building a very competitive horizontal platform to enable many companies to innovate in the space and drive this disaggregation of the r.a.n. In the same way that was successfully done within personal community — computing and the smart phone.
Charles:
thank you. Peter, i wonder whether you share that perspective or have a different you. Mute.
Peter:
thank you, charles. First of all, i think with the speed of innovation, i look forward to joining the next panel by my virtualized avatar telepresence. I know we can do this. In fact, even though we are talking about open r.a.n. Today, the innovation speed is not where it needs to be. Our current forecast is only 10% of the network will be open r.a.n. By 2025. We need to work together to accelerate that. I know that is what is common to a lot of the panelists. We are seeing an explosion of companies optimizing their hardware by the specific function in the network. Many of them are represented today. Specific examples we already heard our beam formers where there are a variety of u.s. Companies looking at both analog and told beam formers, and on the networking side, the explosion of silicon photonics to interface directly to the cloud. At global foundries, that allows us to optimize the technology to meet the segmentation in the network, and the second thing we need to do where we are not quite there — we talk a lot about interface standards, but in many of our chips, the interface standard itself is provided by a third-party ip provider, and the stability of those interface standards and availability of them in specific technology nodes also needs to be accelerated. That is something we are very eager to work with our partners on. Once you have interface standards, one of the things we can offer and hardware companies can offer is a route of trust, or physically unclothed bull — unclonable functions. With route of trust, you can have those functions be more secure. The final thing i would like to mention is also the opportunity and the role of the u.s. Government. We talk a lot about security. Many of the hardware companies are u.s. Companies. We have many of them represented today. But if you take it one step lower and look at the chipmaking industry, right now there is only one u.s. Supplier of chips. That would be global foundries. The rest of the chipmaking industry is offshore, and we have been working with the u.s. Government to ensure that there is a robust supply of silicon from onshore suppliers.
Charles:
that is great, and thank you very much. All of this sounds like this is going to require a staggering investment. Diana, i’m curious, at least recently it is fair to say that telecom equipment has not attracted a lot of attention from the venture capital market, especially when you compare it to the kind of money we see going into software startups in the valley and other leases around the world. Do you see oran changing that? You did it again.
**Diane: **sorry about that.
[Laughter]
absolutely. With a closed proprietary system, the barrier to entry is just so high that we have not seen much venture capital money in this space. To be a start up company and to make the investment of time, brainpower, funding — to get a product ready and to only be able to sell to a handful of companies — you know, it was just too risky work. When you open up those interfaces, you are injecting that additional competition which is going to drive innovation. It will bring more vc money back to the telecom space. We are moving to the next generation of networks waste on the open interfaces and protocol, so it is pretty amazing to watch unfold in such a short period of time.
Peter:
can i comment on that as well question mark in the foundry, we are seeing u.s.-based started but is focused on specific segments of the network infrastructure. As was said here, it was made available because now you are allowed to optimize your solution for a particular piece of the segment, so many of the networking companies, beamforming companies, are very much startup companies, and we are seeing quite a you entran ts from the u.s. Venture-backed community within some segments of the network, again accelerated by the open ran — r.a.n. Standards.
Charles:
mariam, would you agree from your perspective? Is this what you are beginning to see as well?
Mariam:
look, i mean — let’s say there are three things that today are bringing us all here together. That is where we are getting all this diversity everybody is talking about. It is disaggregating the network into smaller pieces. It’s opening interfaces. And the third one is the decoupling of the hardware from the software. So the makeup of this panel and all of today is a testament to that, right? Basically, you know, i am very encouraged by what i see so far. I see the smaller companies joining the ecosystem. This will definitely foster healthy competition and drive innovation. This will definitely — you will see a lot of movement in the economics and investments around this. And i would like to see more focus on converged networks. But yes, very encouraged by this.
Charles:
john, as you know, one of our jobs at the fcc is to help facilitate 5g network deployment. This is one of the things we spent a lot of time on, so that consumers can reap the benefit of the new use cases that will run the network. I think you are in a unique position to tell us whether you see a startup environment developing around 5g, and what the role — what do you think the role of oran will play in that?
John:
yeah, i think — not to be negative, maybe to be realistic, we are not where we need to be. The u.s. Vacated the wireless industry, with the exception of some semi conductor companies and the actual operators, over the last decade. We have now woken up to that and realize this is a strategic technology. I think everybody has got that message and put in comments. Now we are in the process of rebuilding in industry. We are not just tweaking it. We literally have areas like large-scale systems integration or the ability to put all these pieces together into a system that quite frankly existing companies and new companies have to step into those roles. There is quite a long journey in front of us. I think we are moving a lot faster than previous generations, but in order to do that, we have to recognize there is kind of two paths. There is a path to rebuild the legacy of telecom — wrong path. I was there before at nortel and other companies, and it is not going to win. The other path is to take what we have learned in the cloud and i t ecosystem, and build the ecosystem that way. If you look at that ecosystem, it requires a number of pillars. It requires at-scale players to put the pieces together. That is what companies like ibm, dell, microsoft, and many big companies do. By definition, we thrive when there is an ecosystem of innovation that can eat into those engines. The second piece is we still have a lot of work to do to reestablish the startup ecosystem. Two years ago, i went out and talked to vc firms and asked, what are you doing in 5g? The answer uniformly was "nothing. — Nothing." there were no exits. I can invest in the company, but even if they build a successful radio resource algorithm, who do you sell it to? How do you use it? I think oran has started to open up some of that avenue. But it is insufficient by itself. We will need anchors, meaning large companies that can deliver the systemness of 5g. We will need access points that mariam described. Those are happening, but it will take time. Finally, we have to catalyze an industry of startups that will take advantage. There is promising early indications. We have seen a number of very interesting radio resource management companies. Everyone of them we ran into were so happy to talk to somebody other than incumbent, because they felt like there are only three types of networks i can use the algorithm in, and the interfaces are not open, so i can’t do this stuff around advanced nema forming, holographic team, new algorithms — fascinating stuff. Not a lot of them, but now they are getting the message there can be more. We track about 20 open radio suppliers around the world. About half of them are what we would call friendly countries. That’s going to be insufficient. If the fcc continues to do the job it’s been doing, which is pretty good, about reforming and reallocating spectrum, we are going to need more radios. Say we figure out what to do with 12 gigahertz. That is a new adventure. We will need a new radio ecosystem around it, and startups that provide innovation to take advantage of those dams. — Bands. We have just begun this journey and it is quite hard for a rural operator to assemble a turnkey outcome in the open ran ecosystem today. The good news is we have pretty much everybody at the table. The bad news is we have a lot of work to do. The opportunity for government is to help us do that, to invest in that outcome as opposed to the individual components, to really focus on the prize, which is when we are all done, there is a u.s. Ecosystem that is dominant around the next generation of wireless. We don’t have that today. But if we all focus on that, we will get to a very interesting place, and all the tools necessary to do it exist.
Craig:
i can speak up on john’s point. One of the dangers we face is if a number of small, highly innovative companies get together but they are only loosely compliant with the oran standards and specifications, we run the risk of making the interoperability problem become enough that we do not get the widespread adoption we all absolutely need. It is very important that we all look at this. Look very hard to not allow this market to fracture into a bunch of loosely compliant component vendors. Then, what we all want simply won’t happen.
Charles:
so that is an interesting and i think important point as we look back at you makes — unix and where that went. What do you think are the key u.s. Technologies are areas of leadership that are likely to be most advanced by oran?
Craig:
there are a number. In fact, in all deference to my fellow speakers, a lot are working on really important technologies that i think will advance. From my perspective, i think we will see a lot of advancement in oran cloud technology in particular. We are already seeing things like data plane development and single wrote i/o virtualization. We are seeing memory access things. And we are seeing a number of security features being added around virtual private clouds. I think you are going to see cloud platforms, from my perspective, which is a key u.s. Technology — i think you are going to see that really involve to support the requirements that networking on a cloud, and indeed virtual r.a.n. Networking, is going to place on cloud platforms — the idea of high-speed packet processing and high i/o data rates. Think we will see cloud in particular evolve to support the sort of requirements.
Charles:
thank you. And i’m curious, are there any particular barriers — we are going to be interested in seeing what barriers to oran adoption you have seen, and what you think. We have alluded to this in different contexts before. What is the role the u.s. Government, other governments, can play in promoting it, if you think there is a need to do something?
Kurk:
there is a lot to be done before oran becomes the defect oh deployment methodology. What was said earlier by john was right now we are on a run rate of only 10% of networks being on compliant by 2025, halfway through the decade. This really has to do with incumbency, maturity, and continuing development costs, i think. Networks cost billions of dollars to roll out. They take a decade to roll out. They are kept for at least two decades. And incumbency position is incredibly powerful, and keeps people within this sort of ecosystem. The other challenge that people have is that — the standard being not quite the standard, this breaking up of the standard, so you don’t truly have interoperability. I think for that to be successful, we will have to have plug fests, system integrators we have not had before, the likes of ibm and dell on this call. A way of de-risking the situation for the operators or the network owners in general, because they are going to be operators that sell to others as well as folks who run networks for themselves. I think that in order to be successful, governments can do a couple of things. They can put their money where their mouth is. So as was discussed earlier, we can have federal agencies. We can have dollars that are taxpayer dollars required to be part of an oran-compliant network, to start this cycle going. The second thing government could do is offer investment credits — d risk — de-risk the risk of developing these different pieces, to jump start the ecosystem as a whole, from small companies to larger companies, because it is a very expensive proposition to develop a lot of this, and if you are going to have relatively small market share for years, that makes it challenging to have the return on investment that you need. I think that other countries will typically do the same thing. They can support open r&d investments and encourage the adoption of open r.a.n. Compliant networks in their jurisdictions, like we are seeing in japan, like we are likely to see in india and europe, along with the u.s. I think make oran analogous to the promotion of the free market. Each country can stimulate this free market, this oran, slightly differently, as long as everybody is working toward the common goal of a level playing yield. What we have seen with the free market, and i expect to see here, his extraordinary results as long as the playing field remains open, because you create an ecosystem that is better than any type of monolithic ecosystem that can exist. Charles: thank you. You know, it’s interesting because a lot of the time at the fcc we hear people say government keep out of our business, and what you have described is a little bit of tension, because you are saying take a more proactive role, but also step back. Is there consensus on the panel about this?
Kurk:
if i could add one thing before you put it to the rest of the panel, i think it is because the playing field is not level with a variety of countries supporting specific vendors. If no one was supporting anybody, i would totally say your former case — stay out, in effect.
Soma:
i would add to that. I would definitely agree that the fcc or government needs to take a more active role. One, it is a strategic industry. It’s not just any industry. Right? This is the fabric of the next industrial revolution. Everything we do is going to be connected. It is a strategic industry. The second thing is, spectrum is a national asset. Whether we like it or not, government is already involved. It is a national asset. We want to utilize the best. My position as if fcc and the government would have an incentive tied to the spectrum, saying that if you want to have the spectrum, either you get credit for using oran or some kind of incentives, because the level — it is not a level playing yelled, as kurt was saying — kirk was saying. You are against 40-year-old incumbents, and you need help to get over that hump. Once you have that, the ecosystem is hard to beat, but it is also hard to create.
Charles: cristiano?
Cristiano:
this is an excellent conversation and i would like to make a few points. It is consistent with some of the conversations i think john mentioned, and also this issue of having a system that is consistent. Governments can play a very important role. This is a different thing than building a new consumer electronics device, right? The nature of the telecom network is no different than the power grid. More important than ever, 5g has enabled a lot of mission-critical capabilities. We need to look at it the following way, especially to drive investment to the sector, and for the companies, the key public operators. They have to operate in network. I think we do this in two ways. One, private deployment of 5g vran — that to happen. It is already happening. It can proliferate to the enterprise. No different than the enterprise building its own access points. It drives a lot of cases. You start with that right now. And then government should put a milestone and seek alliance with a number of countries that have the same vision, in the process of upgrading their telecom infrastructure — the u.k., other european countries, japan, korea, and many that align with this view, to put a milestone that will drive operators to specify a system or platforms that are going to enable those oran things. The risk of doing it prematurely, if you want to do it today — the risk is you end up with a noncompliant open ran — r.a.n., and you start, because those ecosystems take time to develop. The way we look, usually, if you want to change that stat, to have more than 10% of base stations and radio units with open ran and vran capabilities in 2005 — i think when you start thinking about as soon 2023, it is to basically drive requirements and timelines and incentives for the public operators to specify right now, as they get rfp out, their next generation infrastructure that will come in with — and then with all the right interfaces. And that gets done by the tier one rubble operators in an organized and harmonized way. Governments can play a very important role aligning that timeline and driving the requirements, because then, as john outlined, you are going to create a huge ecosystem. It is going to be investing toward 2021, 2022. Wish turning to launch the systems in 2023. There will be a lot of act — exit opportunities for start up companies to build the pieces together, and then you have the major operators in the world driving this. If we don’t do this in an organized manner — and i think we can start with a very fast innovation out of the private sector, private 5g networks. That is going to be incredible, starting right now. But we need to use this as a subsidy to drive next-generation networks, and then we are going to end up with a true — the private sector is very good about doing standards like the 5g standards grid example, a true open r.a.n. And vran innovation. Very important i think to have that innovation and coronation and drive that, because this is like the energy sector. The airplane is flying and you have to do these changes midair. And we can get there by acting in that way.
Charles: go ahead, yes.
Peter:
i think that is a fantastic timeline. I think even 2023 is probably the fastest you could do, but it is going to be challenging. What we have not discussed today is the wide range of companies who are starting to deploy non-terrestrial networks, including low earth orbit satellites and other varieties, the need to get those into the open r.a.n. Infrastructure is imminent. Usually people talk about open r.a.n. Within traditional wireless networks, but the excitement around nonterrestrial networks is quite palpable in our industry.
Charles:
cristiano, would you agree? I noticed your focus was on working with major national network carriers, but we have been making a lot of spectrum available at the fcc that analysts say could be used by private networks or for internet of things. Is there a role therefore pushing oran principles? Or is that really too sliced, as you might say, to make it a relevant force?
Cristiano:
absolutely, it is happening right now, that i think that is what is going to drive a lot of the development and innovation, as i said in a prior comment, which will then enable this technology to be mature for major shifts of the tier one operators to their next generation open r.a.n. I think they keep that technology filament is to accelerate into private networks right now. I think governments around the world are making spectrum available for private use. I think the fcc has done a great job with the spectrum. I think we have the pal, the g aa. Those will enable private deployments. Private deployments — it is the true promise of 5g for digital transformation of enterprises. It is easy to see, when the pandemic happened — we all went home and started to get connected to teams, like we are doing right now. We are very fortunate in our company because we drink our own kool-aid, but many other companies realized a lot of their assets — people were not connected. There is accelerating digital transformation where you connect everything. That happens as you bring 5g to enterprise. I will give you a simple example. If we are doing teams right now and people are collaborating in one drive and you don’t have access to your papers, you are seeing more and more that the hard drives move into the cloud. The i t infrastructure entirely is moving to the cloud. You are connected 100% of the time. You don’t have access to a workstation. You use high bandwidth to do that on the cloud. We are going to see 5g going not only to manufacturing robots, but to the enterprise in itself. And that is going to happen to private networks and spectrum available for those private networks.
**Charles: **thank you. Mariam, are you just gesturing?
Mariam:
sorry. I could add to that. I completely agree. That is my earlier point that networks need to be looked at beyond what we are calling 5g today — whether it is the satellite, wired networks, antennas. I have been a fan of open r.a.n. And virtualized r.a.n. Since the early days. When i came to cable a year ago, the first thing i realized was that cablelabs has been at the forefront of developing multiple generations of open and interoperable communications, so this is not a first. It is just a first maybe for r.a.n. The core network was done before. We have done this before and we know that open r.a.n. Is not only possible, but it is an evolution of the radio access network. I just want to mention that. Every innovation is going to have its early stage. This has its early stages. It will definitely happen, and we need to look at a broader view of how open and interoperable infrastructures and specifications will help the entire communication system.
Charles: thank you. Diane, we have been talking — john, would you like to follow-up up on that point?
John:
just one maybe snarky point. Not only have we done this before, but the telecommunications wireless industry is the last one we have done it to. We have done this to compute, storage, networking, virtualization, software development, application delivery. The one outlier that is not in this modern format is telecom. It is things like wireless. Not only have we done it, but if we did not do it, this would continue to be an anomaly. It is very likely we figure out how to do this, we will do it, and we will have what i describe as wireless as a first-class citizen of the cloud era, which it is not today.
Charles: it is not too snarky.
John: some people might feel that way.
Charles
diane, we have talked about all the positive things. I wonder whether you have a sense of the things we would lose out on if we don’t promote oran and move in that direction.
Diane:
sure, and i think one thing we hear time and time again is we are jumping to that next generation of networks. It is the optimization. It is the customization. It is the things we have not thought of yet — where there you are building out a network in new york city are in my small town of rutherford, maine, population 5000. You need a network that will be best for the customer base. We are talking about opening things up, driving competition, driving innovation. We are taking a leap ahead. Just to go back to your last question about her governments and the private sector can work together — that is why the coalition came together. We came together to promote policies and initiatives that help take the next step, whether it be public and private partnerships, research and development money, or education events that help bring together a wide swath of the private sector to discuss what they are doing in this space, and how they are taking networks to the next level. Go ahead.
Kurk:
if i could add to that, i think that if this is not done, we are missing this giant opportunity. The next generation of wireless is all about efficiency. It is driving down the cost per bit her area, if you like, covered. Without driving that down, all of those use cases we have that we can consider today, or the things we might not of considered yet, simply want exist. I think this is necessary beyond what we are thinking about today. Can you imagine 20 years ago if we would have thought everybody would have been streaming today, and all of the opportunities for additional channels like the one we are on today being available to sort of — democratization of television content? We have so many things that get enabled by an opportunity to innovate, and an opportunity to really take an ecosystem that was closed and fairly slow moving, and create a very, very fast-moving ecosystem — i just think it is — it is an absolute necessity to happen if we want to enable what the future of wireless is going to be.
Charles: soma, you had a point?
Soma:
yes, i do. It is not just about cost and efficiency. It is really the new innovation that is going to open up. This is the iphone moment of cellular networks, right? We are just about to cross over. Just imagine the things we can do, all the things in the iphone, and the same things we can do with the network. You say, what would lose out? I believe ai is the next industrial revolution. I think everybody agrees to that. This is about national productivity. How can you be highly optimized? When you have ai and you do not have to outsource call centers outside of the u.s. — Bringing ai and 5g is going to go hand in hand. It is the next level of productivity and of them is asian. It’s very dependent on ai and 5g . It is not just about the next level of connectivity. It is the next industrial revolution really depends on ai and 5g. We really have to take a proactive approach with this.
Craig:
just to pick up soma’s point, we heard an example from to wreak — tarik about spectrum efficiencies. If you see how it can be smuggled into different components along the end to end solution, we have barely touched the surface of what ai is capable of doing and every one of these components. If we miss out on that opportunity, yes, it will be expensive. It will be less efficient. Frankly, we won’t realize the potential so many of us are looking forward to with this.
Charles:
this is a great way to end our conversation. We are almost out of time. To conclude this panel — this is an all-play opportunity. What are the top potential innovations that each of you see resulting from oran? Go ahead.
Mariam:
i will go first. So, i talk about convergence, ok, earlier. What does that do? It solves both practical and interesting use cases as i was talking to you guys and realizing what is happening. The practical problem is around the seamless user experience. Simply, if i can take this call from my desktop right now, transfer it to my mobile device seamlessly, so i can run upstairs and tell my kindergartner, who is sitting through a virtual class, once again, how to unmute her audio and seamlessly come back down here and talk to you guys — that is practical. That is not anything fancy. But also, this convergence will enable the future of things like we talked about today — autonomous vehicles which require ultra reliability. A.r. Nvr, which require low latency. That is what excites me about it. — A.r. And vr, which require low latency.
John:
as we make devices intelligent, they have finite power and finite compute capacity. To do spectacular things, they need access to abundant resources. Today, those abundant resources are somewhere else, and public clouds far away. They are out of the real-time boundary. As we build cellular environments and have mobile age compute, and we are open and can innovate in that ecosystem and add functionality to the network that is literally within one to five milliseconds of the compute device, we can do things like have collections of cars and roads work together to see around corners, to do smart agriculture were backhaul is not the problem because the ai is near the devices. All of these things are exciting. We are the largest provider of compute vices in the world, i think, today, and we are excited by the idea that if we have this kind of network innovation going on, and we can land compute close to all of those devices, they can exceed their raw potential in really interesting and innovative ways. That is a super demanding requirement, but it is super important for the 5g environment to be part of that ecosystem, not an independent ecosystem.
Charles: who else?
Peter:
i will start by saying the industry 4.0 is one of the things we see as a tremendous opportunity in the semiconductor industry for open ran and 6g. Already, our demos in the clean rooms with operators using a.r. Goggles to free up their hands and service equipment — it is unbelievable what you can do, and the opportunity to bring manufacturing back to the u.s. As you put more intelligence on the factory floor. That said, i will quote the previous chairman of the fcc, tom wheeler, who said really our goal is to deploy an agile, fast, and low latency network. No one envisioned were — uber when 4g was being deployed. I think the opportunities are going to be in places we cannot even imagine. Once you put this flexibility in the hands of app developers, software developers, and hardware developers.
Kurk:
i think this is an opportunity to help close the digital divide. Making the systems available to you wherever you are, whenever you are. Anymore future sense, it kind of ties in with where you are in creating an environment. I say this not just because i love star trek and my last name is kurk. I think there is a step well beyond avr or a.r. Goggle to white fields, to hollow rooms, to things we have not invented yet today, that will fundamentally change the way we work and interact with each other, and collaborate not just within the wireless industry, but across all industries, in new and meaningful ways, specifically when we are geographically separated, as we are in these times.
Charles: thank you. Cristiano?
Cristiano:
i would like to answer that question maybe with a different angle. I think there is an incredible amount of new services, opportunities with 5g. When we develop a technology like 5g, we know it has to last a decade at least. There is going to be huge potential for innovation. I’m going to answer your question — we saw a lot of innovation came to the telecom space in many cases through applications. I feel that what is going to happen with open r.a.n. — It is an opportunity for the telecom sector and telecom players to innovate. The thing open r.a.n. Will do on innovation — it will change how development is done. It will change the paradigm shift from having maybe one software release per year that all features get prioritized by the vendors, and once you finally get the feature, you cannot commercialize because the billing system does not support it, right? That is what happens in the telecom sector today. Open r.a.n. And vran will change that and allow innovation to happen like it happens on the cloud. Now it will happen like pc’s and smart phones. Great news for the telecom sector. All of those features and the full potential and the vision of 5g will be fully realized.
Charles: soma?
Soma:
i would end with saying 4g was about smart phones, cloud computing, and of course 4g connecting to mobile broadband. 5g is about iot — not just smart phone, but smart everything. Smart factories, smart robots. Smart everything, 5g connectivity, and ai — these three things, the whole ecosystem is going to take it to the next level of productivity.
Charles:
we are cutting into your lunchtime, and i am conscious of that. Thank you all for a very stimulating and exciting panel discussion. Appreciate all of your input. I would like to turn it back to evan to conclude. Thank you all very much.
**Mariam: **thank you.
Diane: thanks charles.
Soma: thank you.
Evan:
thanks for participating in a fascinating panel. I did not think i could be more excited about this technology than i already was, but you have achieved that. You have at least one fan here. Please enjoy your lunch. I hope it is a physical lunch and not a virtual lunch. This is one thing we should not virtualize, which is food. There is your dad joke for the day. At 1:40 p.m., we will have another fantastic panel — lessons from the field, where do we go from here? We will hear from operators currently deploying this technology in the field, what they have learned, and how we can take those lessons going forward. That panel will be moderated by chairman pai and will begin with opening remarks from commissioner O’Rielly. Thank you all.
[Lunch break]
Afternoon Policy Session
Chairman Pai:
Welcome back, everybody, to the fcc’s 5g open radio access network forum. We have had excellent panelists and speakers already today and we are kickstarting the post-lunch speaker series with another series. I want to start off by recognizing my friend and colleague, commissioner michael o’rielly, for some remarks about this topic. The floor is yours.
Commissioner O’Rielly:
Good afternoon. Many thanks to chairman pai for establishing this forum and for inviting me to provide a few comments ahead of it — ahead of the next exciting panel of experts. This has been a fascinating discussion so far, and i have no doubt this upcoming group will provide further insight into upcoming technological innovation. When the pandemic disrupted the original plan for this forum and the suggestion was made to convert it to an online event, i enthusiastically welcomed the move. Nowadays, there are few if any commission functions that must be done in person. I hope future commissions readily embrace this reality and cut back on forcing people to travel to the new headquarters for each of these opportunities. As a technology-centric agency, it is imperative the fcc reside on the leading edge of technology. Technology adoption starts at home. Maximizing our own use of advanced communications will lead to more some policy outcomes. Turning to the topic at hand, there should be little doubt about the importance of wireless communications and devices in today’s society, both for consumers and businesses. Almost every american uses in interacts with wireless services in one form or another in their daily lives area the question becomes whether these wireless communications, especially those prevalent in the envision future of 5g licensed services, are sufficiently protected from certain nationstates, rogue organizations, troubled individuals, or a combination thereof, each of which may intend it to engage in nefarious or harmful activities against americans and the rest of the world. The advent of open r.a.n. Provides one path of potentially minimizing exposure points. In its simplest conceptual terms, open r.a.n. Can be considered analogous to secure interoperability, by breaking wireless networks into components and moving away from end-to-end product lines. Overall security can actually be improved, whether it is reducing reliance on foreign manufacturing or providing incentives to harden physical infrastructure, protecting corresponding software from intrusions. Open r.a.n. Can reduce threats to overall network security if done properly, and give users the necessary confidence to transmit even the most sensitive data at any time, from any location. Let me lay down three conditions i believe are essential to secure the success of open ran — r.a.n. First, it must be done without technological mandates from the u.s. Government or any intergovernmental body. The fcc and other government entities like the capabilities and requisite knowledge to impose specific network design requirements or other such directives on the private sector. At this point, it has been proven time and time again, but bears repeating here. Second, we must maintain vendor neutrality. That means no single company or select set of companies should be blessed or favored by the government over others that provide similar functionality. We must not pick winners and losers, especially since doing so can stymie the advancement of ideas and innovation. Third and related to the first two, the process must room involuntary. Certain companies may have nuanced views of how to develop and implement the new technology, and they should be permitted to proceed as they see fit. The market will sift the best ideas and ultimately determine which approaches work best. With that said, i want to extend my thanks to all the expert panelists, and express my appreciation for being part of what is truly the best panel of the day. Over the next 70 minutes, panelists will explore ideas and discuss the current state of the marketplace, and what is being done in the field. The discussion will likely examine the practical experiences and case studies of equipment manufacturers and software providers in the development and advancement of open r.a.n. What is being done today? What have been the significant successes and failures? What are the major obstacles going forward? And what does this mean for the future of open r.a.n.? I think everyone for their attention, and i turn the floor back over to chairman pai.
Chairman Pai:
thank you very much, commissioner, for your remarks, and for framing the focus of the next panel, which is indeed lessons from the field. We have talked already about an introduction to oran. We moved to potential benefits. Now we are going to move ideally to some of the practical — what are some of the benefits that companies here and abroad are seeing in terms of the deployment of oran technologies? To help eliminate that topic, we have a truly incredible and broad subset of speakers. In f political order, we start with john baker — alphabetical order, we start with john baker. We have lori bigler from at&t. We have stephen bye, executive vice president and chief commercial officer at dish. We have marie-paule odini from hpe. Mathe oommen joins us from reliance jio. We have steve poppa, the founder of parallel wireless. And marcus weldon, the chief technology officer and president of nokia bell labs. Thanks to each of you for taking the time to speak with us about these important issues. I thought i would kick it off with matthew. Obviously, jio has done incredible things in india to dramatically disrupt the wireless ecosystem. Arguably, nothing has been more interesting and pertinent to our deliberations today than what you have done in the virtual open radio access network space, india being a very vast country with densely populated urban areas and sparsely populated rural areas, similar to the united states. I thought i would turn to you to see, was virtualizing the network a challenge for you in both urban and rural environments? Why did jio choose the standard? What have you learned in the process of deploying this new network architect?
Mathew:
thank you very much for having jio. It is a privilege to have myself here. It is always a pleasure, chairman pai, to hear from you, as i would say you continue to push the ecosystem to a path that is unlike the traditional, if you will. When we hear ourselves talk about open — oran, open r.a.n., i think each of us have strong perspectives, if you will, of what these are, where this is, and how and what is the right approach to deliver toward operationalization. I think it is important to get to a common reference definition . Very important to get to one standard that is clearly interoperable, because we have oran, open r.a.n. — Many different formats. Having said that, when you look at jio, everything we have done and will continue to do will be certainly unorthodox. What i mean by that is the whole world is talking about open, and i am not yet sure of any other network yet, i would say — and i know it could sound a tall statement — that is all ip and all iot, and all ip version six, anywhere in the planet. Why i say this is important — because without all of it — and it is not about downloading an application onto the phone. This is natively available. That means it devolves to 5g, with a fundamental building block. You need ip version six for security. I believe these are essential and critical building blocks. In terms of where we were or are in terms of delivering not just the lowest costs or lowest bid per cost — for any other network, we also believe we have the most economically viable db, because we are delivering about 44%, which i think is pretty good in the industry. When we launched our network, we launched it in one city. That one city happens to be india, the whole of india. It was not about rural to dense urban or metro. It was all across, the — because we lived in inclusive india and i can — and in inclusive launch. There were right of ways that needed to be done differently for each demographic, each area, as well as from morphology perspectives, etc. But the biggest challenge continues to be site acquisition. How did we resolve that? We were doing 700, 780 sites per day. The only way we could do that per day — the only way we could do that in a democracy — you know, it is not the easiest to bring a site up. When we did that, we built our own ground-based mass. Why essay that is very important. For technology to have your own efficient and effective infrastructure, the reason is because you need to make sure that when you use massive and my ammo — massive mimo, without that, oran does not give you the benefit you need. Massive mimo is important. The utility cost is important. Having the right utility solution, from manhattan to mumbai to kashmir — the power is critical. In low-cost infrastructure, via a highly efficient backbone, is very important. Without the right wireline or without the right bone, 5g, along with low-cost infrastructure, i think is very important. One more point i would add is a challenge as well as an opportunity is handsets. We knew when handset was a critical component, especially when you go to the rural market. We solved that problem when the world told us the lowest handset we could get into the market was $84. We came up with a $22 handset. Today, more than 105 million of that single sku. In the 5g ecosystem, it is not just about the oran from a network, but we also need to push hard in terms of getting the device ecosystem not just from cost-effectiveness, but the type of devices as well.
Chairman Pai:
that is terrific. Thank you for that background and for your vision. Certainly in terms of inclusiveness and reach, that is something we share here in the united states. Speaking of vision, i thought i would turn to stephen bye at dish. Dish has been involved in the oran space, leading up to and including today. I was wondering if you could speak about how oran informs dish’s vision for becoming a wireless competitor in the space, and what lessons learned from building a standalone network have you learned thus far, and do you expect to learn into the future.
Stephen:
thank you for the opportunity to be part of this form and your questions. We are very excited at dish to be in the wireless business with the acquisition of boost. We have talked about being in wireless for a while, but now we are in the business. As you know, we are committed to building a greenfield nationwide network that is clad native and 5g, based on oran capabilities. We are well underway toward meeting the buildout obligation. Very excited to be here and very excited to be part of this panel. When we look at oran and innovation around oran, we think it is important, but it is really only one step, and i think from the earliest sessions as well, what is really important is the openness. It is an open — open ecosystem. We think that unlocks the potential for u.s. Vendors and u.s. Technology companies to innovate around open architecture and an open platform, so we think that is really important. I think commissioner carr’s analogy to the pc industry is important. There are a lot of lessons we can look back on as we look at the opportunity going forward. For us, it really starts with 5g . 5g is a great platform, and oran as part of that. Fortunately, we are in a position where we don’t have any legacy. We don’t have 3g. We don’t have 4g. We don’t have antiquated oss and bss systems we have to rich — retrofit into this architecture. We look at this is a tremendous opportunity. With oran comes the disaggregation of hardware and software. Really, i think a lot of things we see unfolding around cloud-native networking are really vital to unlocking the potential that oran provides. To us, oran is an important step, but one step, and it is really this cloud native capability and the open environment that really is important. And timing is everything. We are coming into this environment where we are building a network, able to take advantage of the great work that has been done to date to get us to this point. We are combining all of these capabilities to do something very unique at dish. I think with this platform, it’s going to unlock a lot of innovation, going forward. I think when you talk about lessons, there is probably three lessons i would touch on, and there is a lot of talk about integration and interoperability. I think one of the challenges with any network is every network has multiple vendors, whether you start with the antenna, through the cabling, through the radios, transport, which mathew touched on, in terms of switching routing or oss, bss — there is no network that is one vendor. It is always a combination of many vendors and partners. Integration is just part of the job we do. That’s part of what we do as a network provider and service provider, is to integrate all those components together. When we look at oran, there is nothing that makes that any more challenging than the work we undertake everyday as we integrate these systems. One of the things i would say is the architecture with oran and open plan actually gives us much better visibility and transparency into the entire stack, into each of the components. We are seeing that as we do all our testing today. It really allows us to drill in and understand what is going on, and where those pain points are when it comes to integration. I actually see it as a tremendous opportunity to integrate more effectively and we have done in the past. I think one of the things that is also unique — the opportunity comes with the open architecture — we have really gone from sort of closed systems to an open system. That allows us to tap into a workforce and skill set that has a lot more software talent, that can bring that software expertise from other industries and bring it into the network and the telecom space, which i think is fantastic. It really allows us to tap into resources and a talent that is less reliant on what i consider sort of unique system-certified systems that have the secret keys. We are able to tap in and i think even develop a much more technically proficient workforce as well, which i think is exciting. The second lesson, i would say, is around security. There is always security challenges with any network. And we have taken sort of a zero trust model. I think as secretary pompeo pointed out, the focus on a clean network is vital. I think there are so many tools and capabilities we can bring to this open environment that come out of other platforms, that are now accessible and available into the telco environment or the wireless environment. I think we are able to leverage that in a much better way than we were in the past, and it is always easy or easier in an open environment to use those tools than a closed or vertically integrated system. It is a lot easier to find the cockroaches when the night — the lights are on than fumbling around in the r train to find something. That is the terminus opportunity that comes with an open environment, with the security architecture. The third lesson, i would say, is teamwork. Everything we do is tough. Building the networks is a challenge. There is no easy button. Having the best team and the right partners is essential. Those partners who share in that vision, who are committed, are really the partners we are working with. We announced a number of those today. They share in the vision of the future and a vision of this open platform that we believe really unlocks a lot of innovation.
Chairman Pai:
thanks to you and to mathew for those comments. As you were talking, it reminded me of an interesting post by ben thompson about india, jio, and the three internets. It talks about the opportunities for scalable, agile, all ip oran-based info structure. I commend it to those who are watching. If you have time, certainly read it. It gives you a hint of where companies like reliance and dish might be heading in terms of their vision for the future. If i might, let me turn to john and asked, what can we learn from the deployment in europe? And to what extent do you think the lessons from those deployments are transferable to the u.s. Market?
John:
it is such an honor to be amongst his team of panelists. Great wealth of experience. Let me start off — there has been a lot of discussion on open ram — r.a.n., which we see a somewhat the correction of the standardization process that took place over the last 10 plus years, and really it is all about open interfaces that have full specifications that allow a product to be developed. When we talk about open interfaces, there are a couple of other interfaces in the network that should be open at the same time. It goes under the guise of the x interfaces. There was some discussion about around field versus greenfield. But in you also get the x interfaces open, such that they can talk with each other, avoiding the hard handoff or dropped call issues — you have still got issues that need to be addressed, and i think we see that the operator should not be having to pay for standardized interfaces to be open. There are issues around some of that. You know, as we have gone into europe, we have really been in this brownfield/greenfield discussion because of these other interfaces that are already fully standardized, but being priced, if you like, into a contractual negotiation makes it very hard for operators to say i’m going to mix and match vendors in the same physical area, effectively. You know, so that becomes one of the challenges. And they looked at greenfield rural areas or dense urban areas, because you have got to build a big area to sort of put these pieces together, because of this x interface issue. Mavenir is a 50-year-old company. It has been pioneering and disrupting the mobile marketplace with essentially virtualized systems from day one. There was the vision on ims type products. We compete and we carry a very large amount of traffic. Virtualization is not new to the industry. But it is coming new to the r.a.n., and that is where mavenir is now focused, taking that vision of a disrupted virtualized network into the r.a.n. That is what we have been deploying probably on every continent in the world, actually. You’ve seen the announcements of vodafone in the u.k. We have a live network there now. Excited to be a supplier for dish in terms of the rollout in the u.s., which will be the largest open r.a.n. Network in the world at the time. Been looking forward to that success. But in reality, it is no different to how we deploy in europe, to how we deploy in the u.s., to how we deploy in japan or australia. It is the same principles. It is how you make this thing easier, cheaper, faster in terms of deploying those elements. And what we have seen by using open interfaces and disaggregated hardware — you can do that very quickly. The supply chain really shorten to the extent i can buy computers off the shelf, i can shift my software. Provided i’ve got the radio piece sorted, we can build these networks very quickly. Talking about the radio, we have been doing some things for the european market in terms of getting radio vendors that are essentially open interface radio. At last count, there are like 19 radio vendors coming to market with radio products, and that is exciting. Open r.a.n. Has suddenly gone from being essentially two or four to 19. So you can see people are really buying into the vision of i’ve got a market opportunity and i am going to invest. We have also been working with facebook tip on the even start radio project for europe, bringing along low-cost radios which are just going into market trials now for european frequency bands. You bring that back to the u.s., the radio side is very, very different. It is tougher to get u.s. Band radios for the u.s. Market. I think we have even partnered with our competitors al tiostar to generate a portfolio of radios we are going to make available to the open r.a.n. Community in the united states. In summary, it is sort of all about execution, supply chain, and having the wealth of vendors with you to help integrate these products. From an experience perspective, it is the same in europe to the u.s. To australia to rectortown — rakutan. Cell site integration, software integration, and an acquisition of cell sites is one of the toughest pieces in here. The technology — i think we understand it and can make it happen, but it is all about the end-to-end process of system integration.
Chairman Pai:
terrific. Before i move to the next question for steve, apologies to those of you who hear kids in the background. You might say i have my own native open radio access in the house, so you might hear them from time to time. Steve, your company has also done deployments here and abroad, south america, africa, even here in rural america. Can you discuss what you see as the role of oh ram in developing environments, as well as rural areas? In particular, is there a leapfrogging effect where an area that might be on two g or 3g or even know g — no g’s is able to leapfrog to the most agile state of network architecture?
Steve:
thank you for posting this. It is a closure to be here, and thank you for that question. I happen to have a lot of direct experience with this. I’m sitting at the end of my driveway in rural new england with my 500 kilobit dsl connection. I’m well aware of the rural connectivity problem. The way to think of rural — beryl and 5g are the same problem. It is connecting ever-smaller numbers of users with ever-increasing cost to deploy and operate equipment. In that sense, they are identical, ok? So if you are going to serve urban areas with smaller and smaller populations, because you have denser cells, it is the same as rural, where you are serving fewer people, ok? And absolutely yes, open r.a.n. Holds the key to accelerating a global 5g rollout. That key is through innovation. I want to bring up one slide that evan has, to make a point that the key leverage point in driving out cost, it turns out, is cost effective energy efficiency for the cellular network. Not for the equipment in the cellular network. You heard mathew talk about massive mimo being essential for 5g. You heard john talking about the challenge of radio. Well, it turns out the radio is where the rubber meets the road, because it influences all these other costs. When we are doing deployments, if you look here at the cap x and the equipment, spectrum certainly appears to be very big. Well, massive mimo is what gives you leverage on spectrum. You can get a five x capacity, special utilization from that. You look here. The things like the site itself, the steel on the site — you need more steel if you have heavier, less energy efficient equipment. You need bigger copper conductors that create more wind load. You need bigger batteries, bigger backup generators, bigger cranes to deploy. The biggest leverage point is cost effective, energy efficient active antennas, right? The way you choose to build your sites gets influenced by the cost of the. If you go to the next slide, another way to see this is all the things influenced by cost energy efficiency — it is virtually everything in deploying a site. Where open r.a.n. Comes — when you have closed interfaces, the vertically integrated vendors are not particularly motivated to push the envelope, right? The reason nokia and ericsson have struggled to compete with huwei — hawei — huawei globally is energy efficiency. They are just now starting to catch up. If nokia and ericsson struggled to build cost-effective, energy efficient radios — i have spoken to eastern european operators that inch marked an existing vendor’s equipment with huawei’s and found that the energy efficiency savings would pay out within five years. If the big guys struggled to do that, how can small companies have a shot, ok? What is interesting — what is happening is because open r.a.n. Has opened the interface, the number of innovators out there that see the opportunity to come together in what i would call a virtual apollo project — how do we double the energy efficiency of this equipment, which fundamentally changes the cost equation? Whether it is getting energy to the site, the weight, you name it — that is what i am seeing going on. That is where the real promise from oran is going to deliver a material difference in rural connectivity, developing markets, and where we leapfrog to 5g. I was on the phone with another developing country, folks planning for 5g, a country with one of the most expensive electricity rates in the world. What i said to them is unless there is more energy efficiency in this equipment, there is no 5g for you other than a marketing version where we just change the number on your handset and deploy 4 x 4 mimo on the existing spectrum. 5g is active antennas. Mathew said it well. And you don’t get that without energy efficiency, especially in markets with expensive electricals. That is the response ability, to have the apollo project to change radio. That is the tail that wags the dog of wireless networks. Thank you.
Chairman Pai:
that is fascinating and we will be on the case to make sure you and your neighbors have added broadband to be able to do everything that you need to do. We are on the case. Let me turn to lori — laurie at at&t. You have experience with existing appointments and thinking about future deployments. Let me ask you about what you see the current deployments saying in terms of some of the scalability and options and barriers to adoption for oran. With access to fiber for backhaul be a limiting factor for oran deployments? How do you see that entire ecosystem evolving?
Laurie:
i’m honored to be here. Such an esteemed panel. First of all, at&t is very proud we are one of the founding members of the oran alliance. We have been helping to drive the architecture, the specification of open interfaces, working with both the operator and the vendor community. The challenge really for an operator like at&t — really with any network architecture, it is ensuring the reliability, integrity, and performance for our customers, so this is not unique to oran. As we introduce oran into our network, the key consideration will be whether we continue to achieve the same level of performance at scale that we have with our current 170 million mobile subscribers and connections. So we are actively working in this direction. Oran is still developing specifications at this time, and some are further along. Others are still in the early stages of development. But having the specs alone does not guarantee interoperability or performance. So we really see that integration is the biggest challenge ahead. You really don’t find the issues with the specs until you actually try to integrate the equipment of two different vendors together. I think the oran infrastructure has recognized that from the very start. The specs include interface specs, but also testing and integration specs. There is a test and innovation — integration focus group that provides an overall framework. It also supports plug fest, which provides facilities and mechanisms that can really drive the interworking between the vendors. One interface we see that is further along in development is the open fronthaul specification between the radio unit and the base vn. There has been a term of this amount of work done about the open interface spec as well as the test and integration specs. We have been contributing to that. We have also done some significant — some limited trials of our near real-time controller on our millimeter wave network in new york. We have demonstrated that we can have a ric working in a field environment with external applications. But there is considerably more work to dutifully validate those interfaces and to implement them at scale. So we are likely to see a more gradual introduction of open ra into our existing network, in particular depending on the development around the fronthaul interface, which is something at&t has been working towards.
Chairman Pai:
terrific. Appreciate those comments. Speaking of vendors, let me turn to marcus. There have been some who have argued that the development and deployment of oran-related technology might harm incumbent enders like nokia and others. Nonetheless, nokia is on the board of the oran alliance, has a software-related line of business, and you are participating in this forum. What do you see as the future of oran and nokia’s place in that open environment?
Marcus:
yes, enthusiastically participating. I have not been dragged here against my will. Let me provide context. Obviously, nokia and bell labs organization — you could probably argue the most innovative tech industry pioneer for the last maybe 100 years, certainly in the ict space of information computing and telecom. That is a rich history of innovation that we are always looking to leverage. A fundamental part of that innovation mindset since what was called the consent decree, if you remember, was actually making the intellectual property open and available for everyone to utilize. I think we have maintained that for the last 50 years, that separation. We fundamentally believe that openness is a key part of innovation. It was always open inside bell labs, but increasingly over the last 50 years, it is open to the larger academic community, to the larger industrial community. I would say openness equals maximum innovation. We would almost make that an identity and an equality. That leads us to lead organizations like x r.a.n. And c r.a.n., which were the precursors of oran. It is just a natural part of our culture, i would say. It is the bell labs culture. The parent company has always supported that culture. Our products typically have been more open than other vendors’ products, to be honest. The x interface — we opened our interface before anyone else, which is the one that allows ace stations to talk to each other. We pioneered a specification so we could get the fronthaul interface open. Of course, we have worked on network device interfaces which we have standardized, and the core interface which is open as well we have really always been open and we have found it a little bit odd that we were classified as an incumbent laggard who would not be open, because we have always considered ourselves in open company. To your question about why would we be open not just philosophically, but commercially — i think it is the fact that we see new value creation potential that is huge. I think stephen bye did a nice job of highlighting a couple of elements, but i will add more. Certainly, more radio variance able to attach to cloud-native baseband. That is huge for us, because if i can deploy a baseband anywhere, anytime, i can dynamically scale to a network. Why would i want to do that if networks have not had to dynamically scale so far? That is where we see enterprise and industrial opportunities, hugely value creating. The next industrial revolution is about enterprise, local area, wide area, requiring what are called private wireless networks, but what it means is sliced networks as well as individual networks dedicated to them, based on commercial cellular specifications as well as unlicensed specifications. For that to work, they want to deploy those networks live and dynamically. Cloud nativity is critical, because they can deploy it on any input structure. On top of that, they can deploy applications natively alongside that infrastructure as the ai application that not only optimizes the network by doing intelligent ai-based control of the network, allowing dynamic optimization of beamforming and signal-to-noise ratios for every user and application, but i can use the same processor, when it is not being used for radio optimization, for ai applications and facial — ignition or voice recognition, use it for aar mvr — ar and vr> — vr. At the edge are the same things you want to do for high-performance applications, and that is part of the new value creation paradigm. A couple of points — one is that this integration point is key. Stephen said it might be easy, l aurie said it might be hard. I’m going to say it is between. I think for baselevel performance, it should be relatively straightforward to integrate all those components and get baselevel performance, but the more sophisticated the performance you need under all circumstances of latency, reliability, capacity — obviously, the bar goes up, and i think it is as yet not entirely clear how we are going to verify the very highest levels of performance without doing more deep interoperability. If you could bring up my slide, we have been trying to work on that problem, because we are trying to look at an architecture that allows for any level of performance to be made available. It is an architecture we have called modular open network infrastructure, and it has oran all over it, so that is the first point i would make. Look. That is not just because i was invited to this panel. I don’t change that when i go to other panels. That is fundamentally part of it. If you had any doubt about the importance of those interfaces to us, you see it there. But the other thing we are trying to do is work on the fundamental set of modules you could procure in the open ecosystem. It is chipsets, but it is actually chipsets packaged into servers or cards that you could then package together, and those modules have different performance levels. If you see on the bullets i have on the slide, you could have a lower performance, mid-performance, and high-performance one. No matter what, they would self identify, plug and play, and allow essentially an automatic configuration and operation of that network at any level of performance. That requires a little bit more work in our industry i think we got baselevel done pretty well through oran and the interoperability forum, but i think we need to do a bit more work to complete this for all levels of interoperability, with all sophisticated radios, with all levels of industrial requirements, and make it really plug and play across all those dimensions. But otherwise, we are well on the way there, is my message, innovating the community to get maximum innovation through openness, but there is a bit more to be done to get to the highest levels of performance in our view, which is why we are working on that extension of the oran architecture toward a more modular, open, variable performance from low to high architecture. So thanks for having me and it is a great panel so far.
Chairman Pai:
absolutely. Thanks for those marks. I will turn to marie in a moment, but let me capture the essence of what marcus just said with a question to mathew. How has jio’s deployment accommodated those low bandwidth or high-performance types of use cases we just heard about?
Mathew:
i think it is a very good question. Before i answer that specific question, chairman pai, i think what lori mentioned — laurie mentioned and what marcus mentioned in terms of interoperability, i think it is an extremely important component, because the reality is we need to make all of these interfaces work correctly, and work not just interoperating. We know when we get two routers to work together, which is supposed to be the likely network — it is a lot of effort that goes into it, whether it is a cisco or a juniper router working in silicon valley, getting them to talk to each other is not a simple task. Why i say that, to the credit of ms. Laurie –when people say we have got oran deployed, that me — i’m an optimist, but i am also a realist. I know what it takes when people say they are oran deployed. I think what we have before us, if i use the term bf r.a.n. — Brute force r.a.n. I say that, and marcus knows this very well. Brute forcing nokia to integrate with another vendor with no open interfaces, it is called the big hammer that is coming from the vendor, in terms of money. That hammer works. We have to evolve from what i call as the ef r.a.n. — Bf r.a.n. To the real evolved oran, so we can get the capability and the maturity the united states inherently has. The reason why i say, chairman pai — i know you have been across the globe a big advocate not just from a regulatory standpoint, but with operators. If you look at cloud, operating systems, ios, chipsets, semiconductors, which are all fundamental to oran, all these ai-based systems and chipsets, these are all fundamental building blocks to oran. We have the ingredients, but we need to come together on a common framework and standard. In terms of narrowband versus broadband versus very high band and low latency, in our network, we have really not got massive machine type communications yet. The reason is the specifications for 5g is not yet there. So when people talk about lobe and connectivity for iot, it is predominantly an nb iot connection. The type of connectivity in india today is predominantly from 500 kilobits to anywhere to about 11 megabits. And our intent is, through oh ram, through trusted partner ecosystems, to get that 10 megabits to 100 megabits. We sincerely believe with the right partnership we can get there.
Chairman Pai:
terrific. Thank you for that. Let me turn to marie with apologies for making you wait so long. I want to ask you, from your vantage point, how do you see the introduction and deployment of oran affecting the relationship between the core and the edge of the networks? If i might pose a second question to thank you for your patience, in particular, how do you see this affecting the internet of things ecosystem that already is flowering in the current environment?
Marie:
thank you, chairman pai. I’m very excited to be part of this great panel, and i am very excited to see the ultimate utilization of the network with the ran utilization, especially since you know hp has been one of the founding members, and we have been pushing for this virtualization for years now. It has been very effective in the core and now it is getting in the r.a.n. With oran, the r.a.n. Becomes a set of virtual network functions like the core virtual network functions. Everything else, whether it is on boarding, deployment, orchestration — this interface to every note, the slicing, the security is homogenous across the r.a.n. And the core, not just for 5g, but for even enterprise virtual cpe, and so on, so we have this vision that was built for 5g to have a common core, which is becoming real. Really, i think there is virtualization across the whole network, and it is not just virtual, it is this software defined network, stitching together all these virtual functions in the r.a.n. With sdn in a distant data center. We really get this homogenous way of doing things across the different parts of the network. Hp, we have been working on these topics for years now, doing that for 4g, for virtual lte, virtual cpe, and now we are doing it for 5g, evolving not just infrastructure, but more and more cloud native, container-based. It is very much ivory cloud, moving to hyper scale. Whether it is called network functions or r.a.n., now it is quite similar, same processes. In terms of the instances — you were talking about the edge. It does not matter so much whether you deploy in the core or in the edge for no latency or certain use cases. It is just a matter of placement. So just a matter of getting the instance where you want to deploy it, or where the ai tells you to deploy it. This is just placement. The back end is just orchestrating that. Once you have that — this is one of the beauties i see with the virtualization of the rem, and this kind of distribution we see now of the r.a.n. More operators are considering deploying at the edge of the network. This is another aspect. We are chairing at simec and have been advocating these use cases for years now. It has been challenging and so on, but now that the r.a.n. Is a make use case, now we are getting these edge locations where we came not only deploy the virtual ram — r.a.n., but also deploy core elements, some applications. I come back to your questions about iot — to enable these use cases, whether it is video servers with high-bandwidth or low latency use cases like the autonomous cars — we need this edge location. We have built oran. Now we are getting edge location, so there is a business case and an opportunity for operators to do that. So all about virtual ram — r.a.n. And oran, also the no black books, the dev ops — we can trace all the changes in the network whether it is core or edge or r.a.n. Authentication or things that are now common across the infrastructure for end to end security. It is even reinforced with some capabilities we have with a secure route of trust with embedded crypto. As mentioned, the energy saving, which is becoming so important, is now much easier with the virtual ram — r.a.n. Across the core and the edge and the r.a.n. By scaling down, you know, less traffic. Overall, i would say this virtualization of the r.a.n. Is really a great opportunity. I think it brings lots of flexibility. Everything becomes pure software, so it is much more affordable as well. It also welcomes new entrants. And all the work you have done, chairman pai, for that, i think is very important, to really empower the whole broader community and ecosystem to bring new use cases, and to be able to be part of this eager 5g as mentionsed also, interoperability will be key, just like it’s been key, we have launched this program to help, to sort out vendors, and in 5g, with all these microservices and interoperability is definitely key, which is one of the reasons why we have launched this 5g lab in fort collins. It’s a facility which is open to any vendor and operators and startups and so on, and it’s based on a secure data center, and it’s really an incentive to perform 5g and testing and interface validation and things like that.
Chairman Pai:
thank you for the description on how you see all of those things. We appreciate it. And again, sorry for waiting — sorry for making you wait too long. And back to laurie. Obviously at&t has a vart of different technologies embedded in its network, fiber, d.s.l., wireless. How do you see oran being integrated into that? Your colleagues would care to chaim in as well, since you have experience working with operators like that, feel free to do so as well. I’d just be curious how you see playing into your existing network architecture.
Laurie:
right, at&t is an integrated, brad band connectivity provider. So we provide it via fiber, d.s.l., and wireless. We strive to have an integrated network with seamless connectivity. Oran is really the architecture supporting our ag wireless service. We view that, it allows us to open up the interfaces between various andonenents radio access network and really shift to a more modular network with different components and software source from different suppliers. So we think 5g is really the perfect time to start thinking about openness, to help foster innovation in the ecosystem. 5g brings new capabilities. We talked about new spectrum, the massive, the higher speed, broadband connections, and also talked about the edge and all the new enterprise services at the edge. So with these open interfaces, suppliers can really take advantage of the capabilities like the a.i., like machine learning, and can bring those to help customize the network, to enhance the network. Via close automation. So we’re really excited about all of the opportunities that they have opened up.
Chairman Pai:
terrific. Steve and/or john, if you have anything to contribute on that that front, suspects you worked with some of the operators facing that chal learning i’d be curious to hear.
John:
let me jump in. I’m just waiting for the opportunity. I think that’s the best way to describe it. When at&t starts the process, we’re willing and able, and i throw that out to marcus about this testing. We believe it’s key to the success of this marketplace, and i was glad to hear mathieu talk about the b.f.m., so i’m going to utilize that going forward. So marcus, we are ready and waiting for the nokia radio. There’s a lot of discussion about open ram. But proof is there.
Marcus:
i would say absolutely. That’s in the absolute spirit of my slide and our intent. I also respond on the question of broadband generally, because obviously one of the things nokia does is offer d.s.l. And fiber and access and mobile access, one of our unique capabilities. Increasingly we’re looking at using the roan intelligent control as an access to intelligent control that would allow swoun fiber, d.s.l. Or cable connection to have the same optimization function as on a wireless connection, therefore, you wouldn’t know which connection. Perhaps you would know, but you wouldn’t care how you were connected because you were always optimized. So we’re looking to take the same concept and apply them, which i’m sure will please everyone. We’re going above and beyond.
Steve:
there’s a cardinal rule of the cell network, which thou shall not regress the capability at a cell site. And with 25 years, 30 years of the road map, there’s a lot to build, and so we’re working with large global frirets get that right. In many ways, the way the early data, intel data centers were deployed, they were actually in rural locations. I was a product manager for the first line of intel computers, was actually the at&t and c.r. Tera data back in 1993-1994. My customers were rural wal-marts, retailers in the philippines, etc., because they were willing to take risk, because it mattered so much to them, ok? And we’re seeing the same thing on a global basis. I’m sure there’s exceptions. We have some interesting green field networks, some we’ve spoke been today. But globally it’s more about brown field, and you can’t regress the network. And when you have user populations, i mean, there are countries where less than 25% of the phones can do a call today. So they have to rely on older hand sets. And those are the folks under the most pressure, more so than — it can be 50-x some of these other countries, but they need the same equipment for 5g.
Chairman Pai:
exactly. You know, stephen, if i could bring you back into it, one of the questions that is often raised is time frame. Dish, of course, doesn’t just want to build a wireless competitor, it wants to build a wireless competitor as quickly as possible. I’d be curious to see how the incorporation of some of the oran technologies that the company has announced of late, as i said, including today, might affect in a positive way the time frame you’re looking to deploy on.
Stephen:
Very good question. Whether it’s with nokia or others, even matrix, you know, they’re just the vendors we’ve announced and partners we’re working with today. And i come back to a comment i made earlier. We’re in a unique position where we don’t have any legacy. We’re not having to go back to integrate into 3g and 4g. Even when we talk about the radio, radio is only one component. I’ve been around this industry a long time doing different technology migrations, and one of the challenges you have is the oss layer as well. It’s not just the radio piece. So for us, we’re not encumbered by that, and we have a great platform in which to sort of build our network, and it’s an enviable position, frankly, and we see the technology at the state where it is today, it is mature enough for us to build and to build it at scale. It’s fantastic where we are, and it’s a credit to the whole industry to get us to this point. But dish is in a unique position to take advantage of that. When i look at it, it’s really the platform today, sort of 5g going forward, building on the components that we have, oran being one piece of that, but the native infrastructure, the development of o.s.s. And b.s.s. Systems today is way ahead of where it was 5 to 10 years ago. We’re in a position to really actually accelerate the deployment of this technology, and i think it will help the whole ecosystem with what we’re doing. I think it will bring the whole system along and allow other people to take advantage of that, whether it be in the developing world or other developed countries. And just to add some color to this, i think it is really challenging for an incumbent operator to actually embrace any new technology. You saw that with 2g to 3g, and now you’re seeing it with 4-g to 5g. It’s really hard t. Takes a lot of courage to take on that migration when you’ve got so many customers that are dependent on legacy technology and legacy systems. Unfortunately, or fortunately, we’re in a great position and a very unique position to accelerate this buildout and create a platform is that really is going to unlock innovation for u.s. Companies.
Marcus:
if i may, one of the issues the legacy vendors have is the number of features we have to carry forward. To stephen’s point. That’s actually what takes a lot of r&d, carrying features forward. So if there’s a perception of less innovation, i would argue we’re equal innovation, but people talk about buildings, but a lot is carrying features forward. And so that’s where some of that complexity and that spin comes from. Of course, the disadvantage, the inability of a new entrant to tchear makes perfect sense. But we have to. So we end up with this conundrum of we have to do the new thing as well as carry the old thing forward, and that’s, of course, makes some sense. And then the new entrance can start from a new perspective, because the disadvantage of not being able to carry the legacy forward, so i think it’s not really that’s sort of incumbents are bad and new entrants are good, we’re coming at it from different spaces and trying to converge upon a common ark terktumplete i think oran is helping us do that, to stephen’s point.
Chairman Pai:
terrific. One of the things — this might seem like an off the wahle question, but it occurs to me given the title is "lessons from the field," i think of my own visits to the field, and i’ve heard everywhere from the delta in mississippi to a dusty village outside of new delhi, india, one of the challenges that operators have found is the workforce, that there aren’t as many trained crews that are able or willing to build these towers, to maintain them, to attach antennas and the like. I’d just be curious for anybody on the panel, how do you see oran, the deployment and operation as affecting the need for a physical workforce? Obviously it’s going to drive economies of scale. Does it also ameliorate the gaps we snow terms of a workforce that is able to do the work in the field?
Steve: we need to pick up the active antenna, bring it back up.
Chairman Pai:John:
i think it opens up the door for job creation, not only in the field, but also, you know, in the operators themselves. I think we’ve seen that the skilling of operators over time as they shed the load to incumbents, but in reality, it’s bringing back the opportunity for operators to compete on a feature by feature, product by product basis, you know, if we do the job right, mobile networks of the future will look like the internet. And to the extent that the operators can develop features themselves, try things out, shouldn’t be needing all these big supplies to produce some of these competitive arguments. And i think you’re seeing, certainly seeing in other parts of the world now, recruiting engineers to take on the oran-type principles as i call it, or open ran principles of open interfaces and interoperability, but also the technical people that actually can get their arms around it. And also the other side of it, once you get these open interfaces in there, some of the features that are there in 5g, you know, you can get into true networks. I’ve got one radio, but i’ve got three operators that are using it. So they’ve been involved in some of these neutral host type trials in the u.k., for instance, where you’ve got one bay station, potentially three or four operators wanting to use it. So i think in some respects it may even lessen the capital needs in the field and the operators become more true software vendors or operators if you like. But that’s sort of the vision all the way down the road. But, you know, i think once we get the supply chains working, you know, some of that’s going to naturally happen.
Chairman Pai:
that actually — marie-paule, in terms of technology, there’s a complex and so on. I mean, now we see, we built 5g call. So this technology, anybody can write microservice now, it’s soize. Of course, you know, then you need to use the vendor and so on. But technology-wise, we have moved away from proprietary protocols, this is really i.t. And internet technologies. And there’s lots of open source as well. So i think it really opens up to be new entrants, and out there in the world, actually.
Stephen:
i think there’s two ends to this. One, we still need people to climb towers and to install equipment and maintain equipments of these remote sites, but i think with the capability that oran, but cloud native and virtualization, in addition to sort of oran enable is the ability to open up an environment to allow for more innovation, and it’s not geographically constrained. Somebody has a great idea, you know, can really develop some capabilities that we can deploy on the network in a way that we weren’t able to in the past. And i think, you know, at one end of the spectrum, we can unlock that and allow people to develop software that really enhances the network in a way that was doyle integrate. But on the other hand, we still need people that are actually physically connecting cables, that are physically having to maintain that equipment, you know, because somebody’s got to replace that when something breaks. So we see the need at both ends of the spectrum. Clearly going to take advantage of the build we’re doing to help stimulate that growth.
Mathew:
just to add to what was said, that is, when we think about people on the field to your question, we look at people on the field no longer as domain experts. We look at people on the field as task masters, meaning the systems in the back end is able to tell the individual what task to be done and how it is to be done. Because you don’t need a very big expert on 55, because we can’t have pht’s across the field, so that’s one big thing that we’ve done. The other thing i would add is oran, open ran is only one small component. There is the infrastructure that steve papa talked about, but also the i.p. Layer, transport layer, systems layer. There’s a lot of work to be done for us to become the true scale i.t.-centric entity that we need to evolve to.
Marcus: Well said, mathew, yeah.
Chairman Pai:
we have about five minutes left. So i don’t want to moan on the part of replies all of the time with questions posed to each of you, but is there anything we haven’t had the chance to discuss yet that a member of the panel would like to amplify on, or a new issue that she or he might care to discuss?
John:
i want to follow up on this open versus open discussion. Earlier today, i think marcus alluded to a little bit of it, that there’s a mention that standards are open. Well, in reality, the devil’s in the detail. Clearly we’ve got to make sure that open specifications, full product specifications, not frame works that can allow for it to creep back into some of the networks we’ve got. I think that’s the important issue going forward. I also think — and this sounds like operators may need an excuse, but in reality, some of the operators are sitting on the fence in terms of where open ran is really going to go, and if they make a decision to go one way, they lose support in the other. In a sense, they’re looking for somebody to point to and say, well, it wasn’t my decision, it was chairman pai’s decision.
[Laughter]
Chairman Pai: i’m used to that.
John
i think some of that is necessary and motivating operators to move forward with the concept of open, disaggregated networks.
Stephen:
one comment i’d add, chairman pai, is to applaud the f.c.c. And your actions over the last few years to bring more spectrum to the market, to encourage the discussion around o ran and the adoption of capability as we go forward. So i think and you your fellow commissioners have done a terrific job of advancing the agenda for the u.s. Economy. I think i speak on behalf of the other panelists on that as well. Well done.
Chairman Pai:
thanks investment it’s been a slog, of course, on both the license and had unlicensed bands, but our hope is that we will, to the maximum extent possible, remove spectrum supply as a constraint on some of the network innovation that companies like all of yours have been table pioneer. So we hope that the future is bright in that regard.
Matthew:
one component i’m sure you have touched on is not just the supply chain diversity or the supply chain security, but i would also say the supply chain superiority. What i mean by that is, how do we ensure they get access to the best technology? They are not because of their size at a disadvantage, right? How do wen sure that? How do we ensure that the raw materials we have access to, why i say raw materials, raw materials will compute, raw materials for storage. Raw materials for semiconductors, right? Even for the companies like qualcomm, nvidia, now broadcom, each one of those folks needs their appropriate raw materials, because even for oran, that supply is superiority of access to the raw material, i think it’s very important. Having the smaller splares be enabled to be successful is also, i think, equally important.
Chairman Pai:
terrific. Marie-paule: agree with that. This is one of the reasons why we have welcomed different vendors to put some equipment in the lab and then welcome some of these other players to be able to test with this environment. I will come all of you to actually contact me and drawing this initiative, and also, i think i would like to emphasize the fact that 5g is great, deploying and so on, lots of radio, and we have this emphasize on the oran and ran. But the brain of the network is the call. So we need the 5g duel deploy to really have the beauty of the capabilities, including the slicing and the security and so on.
Marcus:
the only thing i would add and you have done expertly, vendors are not bad. Incouple bent vendors that embrace open should be encouraged as part ecosystem, because a massive scale 5g network will require a large amount of incumbent vendor deployment, as well as new entrant deployment, so the hybrid is not one or the other, it’s both that’s actually a very good message and a positive message that vendors like us like to hear, as opposed to being painted as the legacy laggard, which i don’t think is quite so, so i think that’s been very valuable.
John: Open Ran is all inclusive, on so it doesn’t favor one to the other.
Chairman Pai:
To those who are viewing, I promise you, i did not set up the resolution of this panel on such a positive, uplifting note, but i will express my gratitude to marcus and john for enabling us to get there. I want to thank our panelists for taking time out of their busy days to share expertise. Obviously there’s much more that we can discuss, but i think we’ve gotten a tremendous amount of feedback on some of the positive developments in the field. So in that regard, i want to thank each of our panelists, john, laurie, marcus, steven, steve, marie-paule, mathew, and anybody else i might have left off inadvertently for convening today, and look forward to continuing the conversation. This is obviously not the end of all the lessons that we need to learn. There’s much more to come, and i look forward to being a participant with all of you as we get into this new generation of wireless infrastructure together. Thanks to everybody, and we will now take a very short break until the con veeng of our technical panel, which will be led by our august chief technology officer, monisha ghosh. Thanks again. We’ll see you soon in the virtual and hopefully physical worlds sometime soon.
Afternoon Technical Session
Monisha:
Good afternoon, everybody. We are at the final panel of the session. We have saved the best for last. Before we kick off the final session, i would like to give commissioner rosenworcel an opportunity to say something about the event and her statement on open ran. Commissioner.
Commissioner Rosenworcel:
Thank you, and good afternoon. It is a treat to join you to kick off the final panel of the day, a technical deep dive into open radioactive’s networks. Now, i might be the last one of my colleagues to address you today, but the record will reflect that i was the first commissioner at the agency to speak extensively about the power of open ran, and that was more than a year ago. And it’s for good reason. Open ran has extraordinary potential for our economy and national security. That combination is something to seize, especially right now, in the early days of 5g deployment. Our 5g future is about connecting everything. It’s about radically higher speeds and lower latencies, opening up possibilities for wireless that we can’t even fully imagine today. If we do this right, it could render our smart phones the least interesting thing about the future of wireless technology. Think of it instead as an input in everything we do, bringing new efficiency and effectiveness to every sector of our economy. Doing this right also means putting security first, and it means recognizing that this new connectivity brings with it a whole lot of vulnerabilities. Today, the fcc has focused its efforts in this area on limiting the deployment of insecure wireless technology by restricting the use of network equipment from chinese companies. We have encouraged our allies around the globe to do the same, but we shouldn’t be lulled into a false sense of security by flashy decisions about hardware and administration headlines about huawei. The 5g security challenges much bigger than simply dealing with a few chinese companies. Restrictions on while way and — huawei and cte are a small fix. No country can isolate itself completely when we are connected worldwide. Plus, our national ambitions are far too great to be defined only in relation to a single country. We must focus now on our competitiveness, on strengthening our alliances around the world, and on reasserting our values by building a new market for 5g equipment. That, i believe, is how we will restore american leadership and secure 5g. This is ultimately what open ran and this whole conversation is all about. It’s about making the market for 5g equipment both more competitive and more secure. Last year, i explained that if we can diversify the equipment in this part of the networks, we can increase security and push the market or equipment to where the united states is strongest in software and semiconductors. That means we can increase vendor diversity and, in the process, increase competition and resiliency. Plus, we will all benefit from software centric innovation, faster upgrade cycles than with traditional hardware. I’ve since testified about this idea before four different congressional committees. Four. I can say with authority mimed him — authority momentum is building. Not everyone is convinced. Earlier this year, the attorney general called open ran just high in the sky. One company has made it into the — just pie in the sky. One company has made it into the history books as the first to launch open ran. So, i i believe if we want to lead in the development of open ran, we need to take action to help build this momentum and the movement from the ground up. Gathering today and learning from one another is a great start, but there’s more we need to do. We need investment in research and development, both from the government and the private sector. There’s a bipartisan bill in congress right now called the usa communications act that would have the fcc provide $750 million to accelerate the development of open ran in the united states. This is a good idea. Second we need to launch and open ran — launch an open ran test band. That’s something we can do at the fcc. Even better, we could build this onto our ongoing work with the national science foundation to authorize citywide test bands for new york and salt lake city. We need to build scale economies for open ran technologies. This happens when we coordinate with other agencies here at home and increase our dissipation in international standard-setting organizations that are doing work — and increase participation in international standard-setting organizations that are doing work abroad. We should consider incentives for replacing equipment with open ran architectures. We need to build incentives for producing the next generation chips that open ran technologies will require. And that includes research into the use of gallium nitride chips, because early research shows it could be more efficient than traditional silicon chips for 5g technologies. That’s a lot of work, but look at the expertise we have gathered virtually today. I want to thank you for being here. I want to thank you for listening. And most importantly, i want to thank you for being part of creating a more secure and competitive 5G future.
Monisha:
thank you very much, commissioner rosenworcel, for kicking off this final panel — commissioner rosenworcel: i appreciate that. I appreciate your time.
Monisha:
thank you very much. Panelists, could you please turn your cameras and microphones on? Ok. So, this has been a long day and i hope most of you on the panel as well as in the audience had a chance to listen in to the morning and early afternoon session. There were a lot of very interesting and i would say at times maybe even provocative questions raised in terms of the preface — promise of open ran. What i would like to do with this panel, with the experts we have today, is take a deep dive into a few of those technical issues that we raised and try to really tease apart what is available today, what is possible today, what the technical challenges are for open ran to truly become a next-generation architecture. I’ve always thought that maybe open ran is the way to get rid of all g. We are stuck in this generational mode of thinking, one g following another. Now when we are in this mode where we can continuously evolve with an open radio access network architecture, should that be the part we are taking? That’s just one among all of the other technical issues we want to raise and discuss in this panel. The way i’d like to do this is a little bit different from the previous panels. We have 10 extremely accomplished experts in their field. I won’t go through their bios. Their bios are all listed on the fcc website. What i will do is go down the list in alphabetical order, give each panelist a couple minutes to just express where they’re coming from into the open ran discussion, and then we will take it from there, into question and answer and discussion phase. Without much ado, mihai banu. You’re muted.
Mihai:
thank you, monisha. Good afternoon. Thank you for attending this panel. I want to focus on the challenges implementing all ran for the sub 6 ghz bands. Massive mine was used for forming which was a key feature for most of the 5g radios. The widespread use of 3d beams is a prerequisite for delivering the highest capacity of cellular networks. This is particularly important for bands like 1.5 to 4 ghz. These are the most valuable bands, with good propagation. As we know, low bands suffer from insufficient bandwidth. Another important thing we need to remember is that the large portion of the spectrum is —
[indiscernible]
so, let’s talk a little bit about the oran challenges. It requires a large number of dissertation stages, such as 32 or 64 — of d igitization stages, such as 32 or 64. On one hand, the ran system is significantly more complex and expressive. On the other hand, we want to split this complicated system into portions any vendor can implement to connect flawlessly with the other proportions provided — portions provided by other vendors. There are so many disputes on this topic, with vendor biases. It could take longer to become the standard than we hoped. We believe we have a solution. In our opinion, the issue is not
[indiscernible]
a reasonable number of digital teams. But it’s unnecessarily complicated with a very large number of digital streams, upper porch — a brute force approach. As illustrated in the slide, can you please show the first slide, one slide i have? I don’t know. I cannot see it. Is it shared?
Monisha:
it’s coming.
Mihai:
first, let us focus on the hardware coherency. For any active array, the only way to make a predictable 3d beam without — dsp is if they are almost identical. In this case, the average phase element is five or less after all the hardware calibrations, and it has to be from all — for all hardware calibrations. The conventional massive mimo does not use hardware consistency — coherency.
[Indiscernible]
processing are obvious. The system on the left digitizes the whole array, but only as many times as the lakers we expect — layers we expect. It is created simultaneously by the aperture. Digital 3d beam forming between the digitized streams can be done as in the conventional system, but with far less
[indiscernible]
it works as well in a —
[indiscernible]
Under all operating conditions to create any set of agile 3d beams by hardware, in addition to digital beamforming. The bottom line is that massive mimo uses far fewer streams than conventional. Therefore, it can be brought to practice right now. Thank you very much.
Monisha:
thank you, mihai, for jumping right into the technical details of what makes massive mimo work.
Mihai:
It is supposed to be a deep dive, right?
Monisha:
Absolutely. Thank you so much for kicking it off so brilliantly. Charles, i’d like to go to you next for your opening statement.
Charles:
Sure, thank you. My name is charles clancy. I’m an executive with the miter corporation. We have a unique vantage point across the entire federal government of a lot of exciting things happening in 5g with adoption in different agencies, interest in standards and policy and rulemaking in different agencies. I think some of the key things that are exciting, particularly from a technical perspective, is the use of artificial intelligence and machine learning in 5g. If you look at the role that machine learning can play in how we optimize resources within our rf environment, it opens up a whole new set of capabilities to deal with a wide range of deployment scenarios. In a typical, i don’t know, u.s. Urban environment, you might be worried about congestion in the rf environment. Lots of cell towers and they all optimize around each other, potentially some limited interference, but not significant. In that scenario, the radio intelligent control that is part of oran provides an interesting set of opportunities for doing all of that optimization. If you think about more hostile environments, particularly as the department of defense considers using oran in some of its systems, you get an additional set of capabilities when you work with more of a contested rf environment, not just congested. The ability to manage around sources of interference or jamming comes a really exciting component that i think the machine learning aspects of oran can provide. One area that’s a little understudied so far is security, so that’s my personal area of interest and expertise. We haven’t really thought through the threat models for oran yet. Do we need to be encrypting the fronthaul link between the cu and du? Is that’s necessary — that necessary? What are the risks of intercepting these over a fiber-optic cable versus over the air? Suffice to say, the more you are going virtual, the more things are being provided by third-party apps, we need to make sure we think through that end-to-end in terms of security. Third, i’d say we need to ground what we are doing in use cases. Oran in the abstract can be very complex. It can mean a lot of different things. If you start talking about specific use cases, whether it’s radio resource orchestration for drones or it’s dealing with contested safety environments, we can firm up what we need out of oran now versus what we could get out of oran long-term. Thank you.
Monisha:
Thanks, charles. Those are topics we are going to dive into in greater detail as we go along. Next up, cheryl davis. Could you please just tell us a little bit about yourself and where you are coming into it?
Cheryl:
sure. I’m very honored for the opportunity to be with you all today. I’m part of oracle. We are a leader in telecommunications solutions with a global footprint, supporting communication service providers as well as our customers. Why are we here talking about oran? As we are offering our 5g technology that’s built on our cloud, open ran is very important to us. Just a few points on why. We provide our 5g service and software so that our enterprise customers have 5g for their int needs. We have to be able to ensure that service works with whatever our customers select for connectivity, therefore we have to be able to operate with vendors and landscapes. Moving to a ran that’s open, interoperable, and standards-based allows us to integrate. As parts of the ran could also fit as cloud deployments, we will be following these opportunities and developments. As we look ahead, we view a 5g solution that is open, interoperable, and standards-based will unleash the next wave of american innovation, as we’ve heard today. Moving from this monolithic model to one in which systems and vendors can work together to deliver 5g solutions is important for supporting multi-supplier networks, and that’s a core principle insecurity. Because our customers are deploying these open standards-based 5g solutions, they have the opportunity to innovate future innovations, whether it be in 5g and beyond, in a timely and cost-effective manner. And we are eager to be part of this and to bring some of the security innovations we’ve been able to develop in the cloud into the 5g system. Thank you again for having me here. I look forward to our conversation.
Monisha:
thank you. Ashitosh, you are up next. As– ashutosh, you are up next.
Ashutosh:
thank you for inviting me. We work with many customers, the department of homeland security, u.s. Army, navy, to see how this technology can take care of many of the points they have. The 5g pillars we have today, to support a variety of applications, like low latency application, iot type application. How they can make the network more reliable and flexible and adaptive. All these things can happen because of many of the 5g
[indiscernible]
it was mentioned in the earlier panel. Things like mobilized cloud, orchestration, network slicing, and security, among others. As the previous panelists mentioned, many of these have been deployed in 5g core. Now we are trying to bring that to radio access network as part of the open architecture and provide the same flexibility to support this kind of network, make the network more adaptive. These characteristics have lots of benefits we can take advantage of. At the same time, in addition to the actual benefits, they also provide a focus on security. If you have that , how can you scale up and down the network and still provide resiliency? How can you take care of network slicing to provide the — and isolate mission-critical from non-mission-critical? You can mitigate within a reasonable period of time. These are some security benefits we can take advantage of. If we apply these two cloud ran — to cloud ran, you can make use of this to support billions of devices, support latency, massive content, and make the network — building ddos monitoring functions. How you can correlate the traffic. You can detect it early enough to mitigate the tasks. This can be done by — there are also challenges we need to think about. The commissioner said before, it’s not supply chain security. We have to think about what are the original security challenges we have to take care of while we deploy this. If you think about virtual ization, if we are deploying slicing, slice to slice attacks, slicing manager attacks. If you’re putting the virtual network functions with functions on cloud ran — you are adding more virtual functions. You have this automation. It is a real benefit. If you have the orchestration, you are depending on the apis. You need to think about orchestration security. These are the things you have to really think about. There are opportunities by moving into oran that are security challenges that we need to think about. As charles said, we need an in-depth analysis of end-to-end components. Open it up, taxonomy, come up with security best practices, and build use cases. How can you support our government, different agencies use cases, and take advantage of 5g enablers to make it secure?
Monisha:
thank you very much, ashutosh. Somebody made the statement that it’s easier to find cockroaches with the light turned on. Yeah, but they find new hiding places, and you have to go after them. Yes, by opening up the network, we do make some things easier, but we are probably changing the threat models, and we have to go over — after the new ways the network can be compromised. Lori is representing a carrier. She’s from verizon.
Lori:
Hi, monisha, thank you for the invite today. Verizon is excited to participate and we are extremely interested in advancing open ran and have already been working with existing and new vendors to help with that effort. The potential for small technologies is a great opportunity. We intend to set up challengers to be successful in this space. Tasks like taking cloud to the edge, decoupling the ran, creating a flexible architecture, and bringing more vendors to the ran ecosystem are tasks that are being advanced rapidly by verizon. We know many of those tasks will be tackled in the oran alliance. These will be critical to promoting the technology and we helped bring 5g to fruition and we’re a player in oran as well as an early adopter. As one of my colleagues says, network architecture, disaggregation — it’s a journey. We are kind of at the first step of that journey, which is the ability to mix and match software with an open ru, and we are excited, because there are future stocks. We hope that will include additional disaggregation and open interfaces beyond baseband and ru. There’s strong evidence that oran is driving a diverse u.s.-based ecosystem. The challenge that we see at verizon is scale and maturity. We have a mature network here at verizon and it’s not a greenfield network. We support oran entirely and know it is the future. We will be adopting this critical architecture and in a timeframe that successfully allows the network to mature gracefully, but at the same time protecting our customers. That’s really the key for mature networks, like many of the tier one operators, making sure we get these needed technologies, like oran, quickly and as painlessly as possible for our customers. I wanted to tahnk y– thank you again for the invite. I’m looking forward to the discussion.
Monisha:
Thank you, lori. That echoes what we heard in the previous panel about how entrenched providers are more beholden to being backward-compatible or supporting existing users, whereas new entrants can be much more agile and jumpstart and move quickly over to a new technology, and that’s a challenge that the industry has to be able to come to terms with. Manu?
Manu:
Hi, good afternoon. Thanks for the invitation. To connect the dots, i’m here representing the opening comments that commissioner rosenworcel made for the open ran test mentioned in new york city and salt lake city, representing the platforms of that public-private partnership testbed between the national science foundation and 35 members of industry. These platforms have been built since 2018, so we were at the forefront of building open disaggregated end-to-end test bed thefts. These test beds are available for testing today. I would encourage you to think of these environments a little bit differently than a lot of the community labs and system integration labs that you may come across. These testbeds have been built both vertically as well as horizontally. Vertically in the sense of having an end to end network architecture with all of the disparate elements from the radio access network back into the edge or the core, but also horizontally. Within each module, there is the ability to stop different pieces of the puzzle in new ways — to stack different pieces of the puzzle in new ways. The testbeds are commercial. It allows early prototyping, proof of concepts. You can build from the start, keeping these principles in mind as these tests are built. Monisha, if you wouldn’t mind putting up the slide. That would be helpful. Thank you. These are the current portfolio of the prophet — public-private partnership. You will hear from two of the three later on. The fourth one, which we heard a lot of in today’s conversation, was around rural broadband. We will hear some of the comments made by the fcc commissioners around the thinking that’s going around the program, trying to think about how do greenfield and brownfield deployments in a rural context come. We have a test platform that’s going to be selected, and it’s going to begin its deployment and construction in 2021. It will be interesting to see how that journey maps the journey that’s been next slide please. In terms of the open research problem, we see this mostly, the biggest one would be allowing innovation of visibility for our community which is the academic community to look under the hood and to give us more exposure into the parameters based more fully so new can be developed and can be deployed. In addition to that, we need to think about how these can be deployed all of these different as indicated and allow for innovation in an open, a truly open system. A couple points that we discuss later come to mind around artificial intelligence and its role especially around the radio intelligent controller. There are more touch points to gather data. As more data comes in, that sort of strengthens the story that we heard earlier today around measurements and the role of the ran, what we are exploring and the general system are the ability to develop data sets for training, training, the ability to compare off line and online learning as well as optimizing the three different latency balances that are built into the system, the real time, the near real time and more real time. We need open test beds for innovation and this needs to be to access the start of the system and the vendors but another major pillar which is the academic community. We still need to — we think about these as a bridge that connects openran research goals into commercial or precommercial trials, especially when the early innovation concepts are being tested. On this piece again, you have three different, front, mid and back pieces and on top of that to connect all of them, you have the intelligence that ties, everyone familiar with the apps, the apps are developed in this open ran system that allow for deep vertical integration off technologies and solutions in the form of applications which are nothing more than a bunch of containers running on top of a stack and be chosen by the operator to meet the objective that they want, back to you, monisha, thank you.
Monisha:
They will, we will do a deep dive into a number of the issues that you raised later on. Let’s move on to a new start up in the space and he’ll tell us about how open ran plays into that strategy.
James:
Thanks, monisha. As you mentioned, networks is a starter, we focused on radio access that works for middle range spectrum and really counting on open ran for our success at this stage. So as you know, the f.c.c. Has licensed five gigahertz and this is seven times for licensed spectrum than what has been available for multiservices in the 600 to 2,600 mega hertz band. It’s supported by the global, as part of the global 5g standard, a number of semiconductor vendors and hand set vendors who says they have c.p.e.s, yet hardly any carrier in the u.s. Right now is planning to build a nationwide network. Verizon is ahead of anybody else, but still they’re doing a number of large cities, some deployments, but nobody so far is saying that they have been offered grade service to 70% or 80% of the u.s. Population. And the reason is simple, a network requires 15 to 20 times more cell sites than a network in the lower frequency bands. It’s not 15 to 20 times less than the cost of building a low band or mid band subside. So this is a huge problem and i think unless we solve it, the spectrum will not get utilized widely. I think that open ran connection can help in solving this problem. Now speakers this morning, others talked about two aspects of open ran. First is the one ran functions as software off the shelf hardware and the second topic that almost everybody had is how open ran brings open ran and interruptible interfaces to the network. Actually, i don’t think that it really helps millimeter wave because if what you’re going to do is, if you’re going to run millimeter wave on the same off the shelf hardware and you’re going to run mid band or low band 5g, he is not going to make millimeter wife 5g more expensive on a basis, if you’re basically end up costing about the same. So if you really want to drive down the cost of 5g deployment, i do believe that we need innovation in hardware design and i think in the u.s. We are pretty good at hardware innovation. So where open ran really helps deployment in mill peter wave is through interfaces. I think this is these open interfaces as well as all of the methods and procedures and integration and interoperability that can allow them to enter the market. All of its strengths as a global standard is still, it’s very complex, it’s a standard, they come out way new release eight to 18 months, they want to encapsulate the features. Interoperability and integration of the network. With open ran and new entragets can focus on what they can do best, new entrants don’t have to try to build the entire ran from the ground up and keep evolving it. And i think this is the method through which we can expand the ran eco system, as a small entrant, new entrant, we can work with others in the market and put together a solution that makes sense for service providers in the u.s. That’s something we’re committed to doing.
Monisha:
Thank you, for that. Yes, that’s one of the topics has always been on my mind is when we talk about in the context of today, it’s usually in the mid bands, going moving up to millimeter wave does open up, i think, a whole new world of challenges in terms of — in terms of all of the other control aspects that need to be fully explored. Let’s move on to james, james is a well known figure in 5g and recently moved to a new company and i’m sure he is carrying on his work on 50g as well, james.
James:
Thanks, Monisha. I appreciate the opportunity to be able to speak on this panel with all of you. First and foremost, i wanted to talk about two different topics, both representing my former company, my current company, both research and test and measurement. I think one of the exciting aspects of open ran is this concept of open interfaces. By having these open interfaces combined with open source software, you can see where this eco system from a research perspective can be created. I mean, one of the challenges that we see in terms of the research community is if you have to basically reinvent the wheel every time you develop a prototype, it gets very, very expensive. And being able to utilize open systems and open interfaces enables researchers that may have one particular aspect of the network that they would like to optimize and experiment with, they don’t have to be an expert on all of the other pieces. And i also think that oran because of itself, the open interface, more commercial companies will be able to participate in these test beds, in prior generations, with the proprietary equipment, researchers had difficulties tapping into the equipment that was provided and a lot of times it was difficult to use. Open source software, i think that could really create a m more robust prototype as we involve the networks. These open interfaces, they create opportunities for testing. Each of these interfaces must be tested. And i think one of the things that people have asked me, they said, well, if you have to test all the interfaces, isn’t that going to increase the cost of the system and in my experience i think no. And, in fact, as you build the c.u., d.u., the o.r.u. These systems can actually be tested independently. Without access to a c.u. Or d.u. Or a combination thereof. That facilitates testing earlier on in the design process. You think about actually doing performance testing on an r.u. Without the base band being present. You can do that. That’s what oran facilitates in this context. In closing, i think that oran is going to change a lot of things. It’s going to change hopefully the way that we research, the way that we test and hopefully the way we deploy next generation networks.
Monisha:
Thank you, James. Ivan?
Ivan:
Thank you very much for the invitation to the panel. Manu already talked about power. I’ll try to stay away from it a little bit. There are three parts, maybe the kcost in new york city, but some are talking about what we gathered. I wanted to touch on related topic of how the industry uses these platforms and to sort of touch on the experience we had with the app operation that are hosting these parts for the last couple of years and now moving towards the so-called establishment and part of deployment and we’ll talk later on about are actually open tests and integration center are being established right now. A couple of experiences there, really the fact that we are facing something that we haven’t faced before which the complexity of the systems. So the complexity becomes a crucial aspect. When i say that, there are a couple of issues. So quite often we have faith and performance problems that these systems are delivering and when you measure it in the context of integration, it’s something you are seeing and quite often they require hands on and onsite presence with people are teaming them up, and the individual testing is not enough. You saw a need to put into context of deployment of big system and not just the ability between the visual components, sort of putting it all together end to end and quite often there are a lot of surprises when we do it for the first time. We see the difference in several industries if you call it that, we have on the one hand traditional vendors who are building the reliable telco equipment that is being deployed and on the other hand, you have the agile software developer culture that is finally jumping in. That creates a huge problem, so we need to sort of merge into society somehow, the two groups and that’s an interesting thing. We need to make it widely available for the community to develop new and better things, beyond what we are traditionally doing with the telco equipment, we need to have these well defined and absolutely open a.p.i.s on top of everything, and of course, have facilities going through with it. And hope as is the case, we will suddenly see things we never thought about, being developed. I hope to discuss more later.
Monisha:
I think you bring up an interesting point about two different communities, the software community and the traditional networking engineering community. We heard about workforce development, but there is a new breed of engineers who have to manage together, the software programming with the network and academia has to gear up if oran is here to stay. Finally we have the p.i. Of one of the power platforms and i will let him talk a little bit about the power platform in utah. You want your slides up at this point, right?
That would be great. I’m a professor in the school of computing at university of utah. I am in one of the industry funded platforms that has already been mentioned. So just to sort of set the context a little bit . The softwarization and the opening up of the radio access network is just the logical next step in the open software defined networking evolution that has been going through networking more broadly. I think an important point, though, when we talk about the radio access network, it is more complex than any of the other networks where these open technologies or open approaches have been applied. And sort of has a position statement, i guess, and i’m echoing what several others have already said, i think there, in the case of open ran, a tremendous need for practical experiment expiration validation in a realistic environment and being able to explore use cases, explore the complexity of the interfaces. Explore the security thereof, explore the new attack vectors that are enabled as it were in an open ran environment. That’s kind of the context of the power platform that we are developing, it’s really meant to enable this type of research and so the picture here, i’m not going to walk through this in great detail, but this is sort of a comic version, if you will, of the platform we are developing here in salt lake city. We have a variety of end points. We have a variety of radio nodes in the ran, both at street level and roof tops. We have h-commute, we have a fiber front hole and back hole. In some sense, we have a ready made wireless lab that is being built in the real world to enable this kind of research. Next slide, please. A fairly unique thing about our platform and, again, this ties with sort of the general trend as a whole, this is software defined end to end. So all of those little building blocks in there, you can think of as running on general purpose hardware and so the functionality is defined by the software that you put on there. So that gives a huge amount of flexibility in terms of experimentation and research. That is going to show on the top left. On the top right, we also have what we call a bring your own device model where you can kind of in this whole architecture come and plug in your own devices at the various layers if the platform we provide is not what you want. That’s it in a nutshell the power platform. So the figure at the bottom is to make it more real in the context of open ran and specifically some of the things that we are doing. So if you take this generally purpose hardware, right, so cloud compute form software defined radio and you combine that with software and oftentimes the software can be open source software. Then you can get this open source open ran environment. This is something that we’re specifically working on by taking the open source components as coming out of the o ran alliance and we’re kind of combining that with other open source stacks out there in the form of open air interface and l.t.e. To get to this end to end open source environment. And then just to echo some of the things that one of the earlier speakers said, a key thing that we try to enable in our platform is that, you know, you don’t have to be an expert to use this. And so for anyone who tried to stand up any of these software stacks, you’ll know it’s super flex to do that. So we have, we follow this kind of lego building blocks approach throughout our architecture and so we try to provide building blocks where you can sort of by the click of a few mouse buttons in oran environment or l.t.s. Environment have a end in which to build. I want to emphasize that power is a platform that is open to the community at large, both academic and industry for research and development. Thank you.
Monisha:
Thank you. As someone who came to the f.c.c., i can’t reiterate having these open test beds. What we have heard about innovation and about new companies and new entrants, all of this is going to happen from people who are maybe in school today and for them to get their hands on these test beds where they can actually play around with these kind of networks is what is going to seed future innovations. So, thank you all, for your opening remarks. What i would like to do now is go through a list of questions and ask you to engage in some discussions about it. The first one is about use cases. We heard that very often, not only among from you guys, but also in the previous panels, so what are some of the use cases that you think are uniquely suited to open ran? Are there certain splits and configurations for example that you would prefer to use in open areas and hot spots, are there performance tradeoffs which make sense in such situations and not others? We heard the nun that only 10% of networks will have oran, will be oran in 2025. Now, should that number be higher or should that be more targeted to a specific application that will benefit the most? So i would like to start off with asking charles clancy about your thoughts on this topic?
Charles:
Sure, so I think there are two types of classes of use cases to consider. One is potentially the deployment scenario use cases where your motivations are moving more economically driven, how much it’s going to cost to deploy a particular type of technology and a particular metro area, for example, you may get cost or economic sort of efficiencies out of oran as a technology. The other way to look at it is from an application perspective, one of the applications that are going to be the killer app for 5g that could be uniquely enabled by o-ran and not a monday religiousic ran architecture. I’ll get to the radio controller, within oran is one of the key things that can enable applications better and faster perhaps than you get in a monolithic ran architecture. Particular use case, i’m uniquely excited about is u.a.s. You think about commercial drone operations, you’re doing optimization of everything, got to do route optimization of the drone in order to get it where it needs to go, achieve air traffic control functions, air safety functions, flight operations functions, but at the same time a drone has a very unique propagation environment, right, it’s at a higher altitude. It’s able to see many g.o.b.s simultaneously and cause uplink interference to cell sites simultaneously if it isn’t managed. Joint optimization of drone flight paths that achieve the metrics jointly with radio resources and being able to manage your space effectively with mid band spectrum in order to assure it’s allocated to a drone from one cell site aren’t interfering with latency sensitive time slots that might be happening on another tour nearby. So i think those are some of the examples that benefit. There are others, autonomous vehicles, not the same environment. I would encourage us to think about the application line that sits on top that can be uniquely enabled in addition to the deployment scenarios that can be done more efficiently with oran.
Monisha
That’s fascinating, when you’re thinking about drones, are you thinking of them as clients or base stations themselves?
Charles:
That’s a great point. There are two scenarios there. One is just your typical delivery drone that we imagine will be in our future one day delivering our food and our packages. And in those scenarios, the drone is really just a u.e. That is connected to an urban 5g infrastructure. There are sorts of interesting problems to find out how you’re doing the real time machine to machine communication and structure rather than just o ran, you do have, rather than the cell on wheels, you have the cell on wings, the airborne cows, another use case, you may be using millimeter wave back call for augmentation. There the challenge is if i have a drone that’s acting that is and a position, radio resource selection algorithms might have all of them go connect to that drone, it has the strongest signal strength which would quickly overwhelm it, congestion issues, the controller is a unique resource that you can do to do all of the resource planning in both scenarios.
Monisha:
Thanks, charles. Lori, do you have something to add in terms of use cases that you see are unique?
Lori:
Absolutely. As an operator, this is one of the most important topics and one of the most difficult ones for people to grab aahold of. I thought it would be good to highlight a couple of examples. We experience a need for a variety of radios for specific use cases and today’s day and age we’re often limited by one or two suppliers have to offer. That can slow down bringing new technology advancement to customers. It allows more players to fill the gaps to address deployment-specific needs. So some examples include, for instance, the inbuilding space. Customers want high g vendors. We need to bring a variety of options to the space as the customer needs are varied. One solution doesn’t necessarily fit all. So if you go into, you know, a customer’s warehouse and put a 5g radio there, that may not be what works in a hospital. So having more options and a variety of radio and base band vendors from which to choose, it helps us customize our solutions for our customers and ultimately provide the correct solution with the best characteristics for that customer. Another example is when different form factors are needed to satisfy zoning or physical space requirements. So again what one or two vendors can provide may not suffice. And flexibility can be the key to gaining approvals to do the physical installations that we need to bring.
i think also with new service offerings and agility in the space, it requires innovation within the ran and reduced barriers to entry. When we achieve open front hall, x-t, interoperability, which can lead to oran, we give ourselves a path to evaluate these new services quickly. That also allows us to introduce new services at scale. So again, one size doesn’t fit all, and the ability to mix and match solutions to our needs are really going to help to push technology advancements. And ultimately having the ability to disaggregate software from locked hardware is a necessity to see these use cases materialize.
So another question that has been on my mind is we hear a lot of talk about latency. Actually at this point i am rather confused. Does oran actually lower latency, or does it increase latency in terms of now because you have so much more software and general purpose hardware in the stack, which in general works to reduce latency. What about applications that require low latency? Does anybody have a thought as to whether oran is the right solution or whether oran would be useful?
I think from that point of view from the r.a.n. Side of things, oran should not increase the latency it could actually decrease it somehow because part of the functionality right now that exists in the b.b.u.’s in 4g is moved into the radio. Now the splits have to be designed correctly . You don’t create some other bottle neck. This is being reviewed and all of that in the oran work groups . The 7.2 split is efficient in that respect. It should be efficient. Now again as i mentioned earlier in my discussion, we have to be careful. What is it we are trying to split here? So that if we — again i am going to go back to the earlier point i was saying. If you are insisting on doing what people have been develop fog 4g blindly and say i am going to be stubborn and go with these millions and zillions of digital streams just because i want to be digital, blind like that, then you may get in trouble. In fact, people are getting in trouble right now. In terms of the amount of hardware that you need to put in the radio to do all these extra calculations and circuits and so on. On the other hand if you do what we do, which is get a little bit more wise about it, then you realize that you can do the same or even better. In fact, you can do better if you simplify things. I mean with all engineers, or most of us, we know that the solutions that usually work in the real world are the simple solutions, not the economy indicated solutions. Simplicity always wins. So that is the way we have to think. We have to rethink about — don’t take what exists out there and let’s just automatically try to put it in the oran. Let’s just rethink the system a little bit and be smart about it.
TBD
Ok. Are kind of ties into a later discussion topic. What we can briefly touch on here. If we are going to think about doing it intelligently, but it is also — you know, it sort of brings together the oran specification, which is now happening in the oran alliance, and you have the 3-g p.p. Organization which is developing what was traditionally the radio access architecture. So where does this come in? Where does this kind of thought process find a home, and which standardization? Does oran need to work much closer with 3 than g p.p. Or into they internalize the fact that futures have to be built, keeping in mind we are looking at open interfaces. Thoughts on how the standardization process would work out or where these very important issues get resolved.
TBD
I can take a stab at that. I think what is interesting about what is happening with oran is yes, it is a standardization effort, open standards, open specifications. But it is also an open source implementation. So that combination, it is not just standards, it is not just open source. It is sort of a standards based open source initiative, and i think that is fairly unique. The other part that is fairly unique is that there is a software ecosystem within that alliance, and i am not directly vofrled in oran, so this is sort of a little bit someone looking out from the outside looking at this process. It seems like a fairly well engineered software architecture. So all of those things, open source, open specifications, well thought-out architecture, that happens to be flexible and open in my mind puts us in a completely different space in terms of thinking about standardization. But if you have a framework that is well engineered, well thought-out, and so flexible that in fact you can add new things in that, then it really changes the way we think of standardization. As you were saying earlier, it might be that now we think of this framework as becoming the standard rather than the next g becoming the standard. I think there is real potential here for allowing us to think completely different about standardization and what that means.
TBD
Does anybody else want to jump in there? James, go ahead.
Yeah, i just wanted to say in general, for those of you who have been involved with 3 g p.p., that standardization body has been wildly successful. If you think about what that group has to do every quarter, all the different r.a.n. Working groups, and pulling together all the requirements. So i’m a big fan, given the way that some of the political biases have creeped into some of the other technical standardization bodies, i think 3g p.p. Has actually done a good job there. Then in terms of oran, i think oran and 3 gpp, at least at this particular point in time, they are actually complementary in my opinion. Release 15 and release 16 of 3 gpp really opened the door for this oran ecosystem to be created, and i think that the 3 gpp members embraced this sort of natural evolution. And to give you an excellent, looking your latency example, when in early 16 you have many slots and two step techniques to lower the access time. Those are more physical things. But if you are communicating with a device that is thousands of miles away, the round trip latency is going to be a function of physics. In an oran network, if it were to implement these techniques like mini slot and two-step rats, just as two examples off the top of the head, then if the service for that was at an edge device or a r.i.c. That was located in a close geographic proximity, then you can see that latency would get the benefit of 3gpp and the benefit of the open ran architecture. So i see them as two things. However, i will do the — you know, i can’t predict the future, but i do want to give props to 3gpp and what they have done. It has been a hard job, and i think they have done a good job.
TBD
You wanted to say something?
TBD
Yes. I think 3gpp has done a really good job so far. If you look at it, 3g is not defining the ran rights. It is defining what the hand set is doing, what the code is doing and what the ram is doing. And the interface between the hand set and the ram is an open interface. We are buying hand sets from one vendor and ram and code from other vendors for a very long time. So that is an open interface, and it is working well. I think so far the interface between the ram and the kornet works has also been open. It has worked very well i think across the world in many networks. Carriers are using on a market by market base, we have ran from one vendor and code from another vendor. I have been part of start-ups where we have been to ran products introduced to new carriers and connected over the interface or i.u. Interfaces that were defined by 3gpp. So that has worked well too. To an extent the challenge over time has been how to do interon itf testing. A lot of that work has been testing. Ideally testing is done by the large vendors, and the carriers have hipped in getting that inter operativity done. We had that done. What has happened in open rand right now, if you look at the hand sets, the core knit work and the ran. Now there is a network to see can we — if we actually are willing to break the ran into smaller sized problems, whether it is r.u.’s, d.u.’s, the rake, the centralized unit. If we can did he compose it, maybe it is easier for there to be more innovation. Otherwise we are dependent on just a few companies who have these things integrated. That is what is creating new interfaces which are not very well defined in 3gpp because that is not what it was initially trying to do. I think they are still doing — 3gpp is still doing work in defining the interface between c.u.’s and d.u.i.’s and other ran interfaces. But still that hasn’t hat not been traditionally the area that 3gpp has focused on. That is a gap, like the oran alines is trying to address at this stage. In the past there were other bodies like the small step forum, that confined the i.u. Interface. So other attempts have been made in the past to try to see how we take the ran, did he composing the smaller elements and allowing more innovation to happen. That is what you are seeing with the oran alliance. We have to be careful. A lot of times we are intermingling when we use this open ran umbrella, we mix up things like oran alliance, which is doing interfaces, other things that are using open source soft ware. The tell come software project. These are different things, but hopefully they all lead to more innovation and mo gress in the ran space. But i don’t think this is an thing to say this is a 3gpp standard setting process.
TBD
You have your hand up?
TBD
I was going to say that doctor i don’t want to say a problem, but one issue is the speed at which these are happening are quite different in terms of 3gpp and the oran alliance. The software community tends to do things at much faster turnaround times, even to a point where people don’t really think through all the interfacing, but come up with an implementation that works, which then creates a problem later on. But that is a separate issue. Where 3gpp tends to be way more thorough in terms of making sure whatever implement aches they come up with are well shot through. Everything said that, that doesn’t necessarily mean that this cannot be arm neitzeled in some way, and maybe we just sort of let the oran community run in this cycle, frequent updates and frequent changes to serve stabilization of the whole thing. It is not clear that all the standards need to be done on a 3gpp level.
TBD
All right. Let’s switch to a slightly different topic now. Did you have something to say?
TBD
Yes. I just call that actually my team that is following closely the oran activity and so on, they are often giving this gesture, and this may be relevant to the government here. The national labs were somehow to pick up on this call for sbroppability — interoperabi lity. That is an interesting idea. I wanted to share that concept with this panel. Again, it is more like way hear from my guys as a suggestion.
TBD
I think the power test buds is one way of doing it, and there are others. In the previous panel, they mentioned that they have a 5g open innovation lab. So there are ways to do this. That segues nicely into the next topic that i wanted to discuss in a little bit more depth, the use of a.i. And m.o. We heard a number of speakers among you as well as the previous panel as to how the open ran wnba enable artificial intelligence to better control networks. We also know at this stage that a.i. Is not a cure-all. It has its pit-falls, biases. Who do we make sure as a community — how do we know that the a.i. Doesn’t quigley — configure in a way where they are worse off than before. You have so many touch points in a network to decide who gets service and who doesn’t. In the past the f.c.c. Looked at people getting service based on the signal strength they are receiving. Now today, signal strength is not a good proxy for whether you are actually getting service. It is all of the intelligence behind the network that gives you the blocks you need. What are your thoughts on how we can make sure that as a.i. Rolls out in open ran that the checks and balances are in place to prevent these kind of things from happen? Manu, you want to take it?
Manu
Sure, thanks a lot. I think with swearble a.i. Or my model, it is only as good as the data that is available to it. Clean data set, and this is talking about april javenlt technologies. This is not a telecom and open ran issues. It has things in everything we are doing. These issues of dean data and bias — clean data and bias, this is much more than what we are looking at. To bring it into the realm of open ran, specifically the idea of developing the data sets for training and entrance. This comes back to off line learning and online learning. The consensus in the specifications, if i am not mistaken, they are looking at some type of offline training, sucking the data and running it into models they are building to see what comes out. Yes, your question of the bias that is used, it actually may go the other way. If you have seen a reasonable science experiment of how do you think you build momentum on a swing set? A robot was trained to do that, it actually started to pump it’s legs on the way back, which is a complete opposite of what the human in the loop would have done toe do a task. The objective is going to be defined by the operator, and i think of thinking in a.i. Terms, we have to meet that objective. In the open ran world, that is going to happen at the r.i.c. The r.i.c. Has been broken down around latency, they have disaggregated the r.i.c. Even in meeting some latency balance. Sub million second, 500 million seconds and above. Non-real time and real time rate would be something we see. Just to recap and summarize, clean data set and then essentially building models, and one day also thinking out into the future, we may not neat any manual algorithms, because the ran would be optimizing the different touch points itself.
TBD
If i can make quickly latch on to that, i think that the data sets as he said is important, i think in terms of verification, i think making data set available to the research community will be pretty important. And then just another maybe smaller comment. To the extent that machine learning and a.i. Is sort of involved in everything that we do, there is a tendency to think well, it is going to do these bad things. But i think manu again said that very clearly at the beginning that it is not as if there is one a.i. Program that is just going to run the network. At least let me say not in the foreseeable future. So there are objectives that are being applied to these models. And these are typically much, much smaller things that might be optimizing energy and performance at the same time. So things that are really hard, but it is not as if you are saying hey, here is a network, go run it. I do think it gets broken down into smaller pieces where at least for the moment there are very clear objectives that are being applied to the models that are being trained.
TBD
So if i may pick up on the question of data, and we have lori from versus doctor verizon here, they have the data in the g.o.b.’s. When you are talking about disaggregated ran, you have a lot more data. But presumably all that data will still be owned by the carrier running the network. So when you talk about now everything data available to the academic community or researchers how do the carriers look at making training data or test data available to the broader community to help make sense of what is going on in your open arc temperature? — Your open architecture.
TBD
I can take a stab at that being the operator representative on the panel. We encourage data sharing on test beds and anything with the academic community, we are largely engaged in it i think where we have to be careful is when we are sharing data on a live network. Verizon has been expressly against that. We are very protective of our customer data. We don’t sarah that. But we are very engaged in test beds, and we encourage that. You learn when you are on the live network more so than on any other types. But we have been able to do that in closed atmospheres so that it is not impactful to our customers, but we can provide that level of data, and the interaction of elements that is needed to provide those test beds.
TBD
Ivan you had your hand up?
Ivan
Yes. I guess there is an additional issue here as well, which is if we are monitoring and we have the ability of observing quite large pods of spectrum. So we can analyze a lot of data. Then the issue becomes who and how can they have access to that data and whether we can sanitize that enough to make it generally available to the whole research community? Or whether we have to as a community have some sort of a — i don’t know — depositor of data where you have to be vetted first before you are given access at a certain level. That is something that the deck democratic and non-academic community should invest. The ultimate goal would be to come up with something like famous jpeg thing where we can then at least use that as a baseline for evaluation of whatever we are trying to do put in place. Whether it is open ran or not.
TBD
Right.
TBD
Thanks.
TBD
I wanted to say a few things . We are also a part of the area. We are studying the pulisicability of the i.r.l. For cases against security being my topic. They have been looking at how it can take advantage of a.i. Due to predictive security. That happens based on the data set you have, based on the behavioral pattern. How can it predict if an attack is going to happen and if it can prevent that attack altogether. In the network, we have a security expert group. We are coming up with use cases and beyond that. Try to see how we can do some cone accept based on a.i. Maybe there is a test bed we can try that. I just wanted to allude to that fact.
Monisha
Thanks. Charles?
Charles
Sure. I wanted to i guess note that i am not particularly concerned about some bias in a.i. Resource allegation really happening in a realistic environment. If you think about a 5g network, it should be stratified into network slices that have i.d. Q.l.s. Are recombirmentes. You may have some belfort effort slices for general broadband mobile consumption. But you have s.l.a.’s, and q.o.s. Profiles defined for those slices already. Then it is up to the r.i.c., the radio intelligent controller, to figure out how to deploy resources in the ran in order to meet those s.l.a. Requirements. Really that is an optimization process. So i would suspect that near term, a lot of the a.i. That is going to be applied is really just sort of sophisticated optimization algorithms that are really no different than the way the radio resource management functioning work in today’s enov’s. Now on a disactivated genodeb, you are doing that across sites in a more efficient way. We may get down the road where there is some black box deep learning that is booing used, but i think we will have built up to that over time and we laugh a lot of sort of let sophisticated more q.l.s. Driven models that we are starting from and using to validate along the way. My point is we are not just going to hand over the entire radio access network to deep learning tomorrow as part of oran did he employment.
In one of the previous panels there was this notion about these network interfaces becoming like a.p.i.’s on your phone. While that sound like a great way to go, it also opens up, and this sort of segues into the next topic, which is security. When we are talking about opening up these internals of the radio access network, as easy a way, that is visible to programmers outside, as security experts what do you think of that? Do you think that that is enhancing the security of oran, or is in certain cases is that opening up new threat models that we haven’t really considered in the way we design a 5g networks? Cheryl, maybe you could chime in there a bit?
TBD
Sure. I wanted to take the opportunity toe share some of our perspectives on security from our demungses and cloud background. We look at 5g and offering it via cloud beyond some of the foundational security measures that are already background into fichinge. We look at the cloud waysed design as giving them the opportunity to shiffing this away from malicious actors. Taking a step back as a cloud provider, we can leverage the scale and concentrate all of our resources, our engineering and expertise in securing that cloud. What we see then is services that are then running on that cloud environment like 5g, and then they have the opportunity to leverage those securities. We look at 5g then inheriting the cloud scale reliability security and allowing those securities to be rapidly deployed with tailored security inches. Encription identity, access control, these are all things that can be brought to bear to secure and then tailored to meet whatever the application is certainly. But to the conversation we just had about a.i. And m.l., we see this as critical to the path forward in security not just for the core, but for securing r.a.n. What we have been able to do is help the a.i. Powered autonomous systems. We have the data as to how things are running in the cloud and apply that to security, to detect and respond to predicted threat at machine speed and taking that burden away from the work force. The people before who have been going through the work logs and giving that to mcdonald’s so that work force can be put toward more strategic tasks. As we are looking forward and the security landscape is changing and how we are accessing this sort of web-based system, our view is the cloud can be at daunting, but we view the cloud as a real opportunity, and the technologies that we have been discussing to really provide some enhanced security for an open 5g did he employment.
Monisha
Go ahead.
TBD
This is an important topic. When you look at the evolution of generation of network sites, the threat model keeps on changing. As i was standing in my official talk. So you really have to look at it end to end, look at the points of attack that can happen, and build that. Then look at the additional security pillars. I think some of them were mentioned. Orchestration, slicing its cloud. These are benefits. You can take advantage of that. If you are having a jamming attack, you can probably do some kind of a k-9 milk radio dallas scheduling. You can provide elasticity. You can really have monitoring interfaces, like data plan, or a control plan to see where is the potential source of attack, where it is coming from. If you have all the security controls properly in place, you should be able to take advantage of the oran features. At the same time, you protect your assets from attack the, either from the subscriber side or the internet side it should not be an after thought. As you map your obligation in specific parts of the network. If you really want to add security detection, you need to put that on the edge. Do that mapping, and then find out what kind of security concerns you have, what else you need. I just give you an example. We are working very closely with the department of homeland security right now, and they would like to take advantage of the 5g things we have now. They are watching closely. We are looking at the requirement, we are looking he can k.p.i.’s. We were talking about key security , not k.p.i. What is the risk? What is the probability of attack? The threat factor once you determine that, then you see the grandlater of attacks, whether you want to put more security control on one versus another. We call it had a threat model, and try to figure out the opportunities, additional challenges, and coming up with technique. And then use the test bed and prove the concept of different use cases. If things do not work, go back and try to change. I know that is how it started. This is how you have to run a security process.
Monisha
Do any of you,, can i point to any new security threats that you think might arise as a result of the disaggregate gation of the r.a.n.?
TBD
I think one — if we think of the key prim tiffs that 5g provides, it is the aspect of slicing. Slicing is about isolation, and now we are saying wait a second. We are going to have this thing that is going to have a cross slice view. And oh, by the way we are exposing a bunch of interfaces here, and we are going to run applications and top of that. I think for me — and it might be a small one, but that is an example of wait, this is something completely new. We have had these isolated slices. Now because of the benefits introduced an entity that has a view across the slices, and we are going to have applications running on top of that. I think i can generalize that to say i think the largest unknown here is because it is a much more complex environment, we don’t really know. And they said the same thing. We have to work through that, we have to work through use cases. We have to figure out the threat model. But we have an inherently complex system which we are now opening up by a bunch of open interfaces. So at the ve least twofe acknowledge that we are opening — we are making our attack surface much larger than it used to be, and we have to think carefully about what that means and what new attacks that might enable.
TBD
I would say that i think this gets to what the actual scale of the business models of 5g end up being. There are two seven osi theory. There is one scenario where a 5g network is running three network slices arcs wrob slice arcs communication slice and a machine communication slice. Each one has a default set of q.o.s. The requirements that go with it, and it is carrier cent rick in the way their monetized. You have three apps plugging into your r.i.c. That is all really contained in the carrier ecosystem. So the carrier can be responsible for the security and robustness of all. It is not lilly open in the sense that the iphone ecosystem is open.
TBD
Every one is going to have it’s own network slice, every vertical i.o.t. Has one. Networks is going to be running 200 network slices or a thousand slices. In this case, this is a much more complex problem. It is not clear to me the time line of how we get from three slices to hundreds of network slices. Well look and see how this changes over time and build up incrementally the trust in how we would vet these apps running on top of the r.i.c. I don’t see hundreds of them by third party vendors showing up initially. I think initially it is all going to be more carrier cent rick.
TBD
But as somebody said we didn’t know about uber when 4g was standardized. We don’t know what is waiting around the bend when 5g gets deployed. Changing topics to testing, there was a comment in an earlier panel that testing and integration of oran is not really that much different than what happens in networks today. We do that all the time. We put things together from different vendors and we have to do interoperability, my opinion is that scales it up a lot more. I know the oran alliance has testing sub groups, but who at the end of the day — is it the carrier that is responsible no making sure that all of these interfaces are tested, work well? Is there a third party that ago gates the network together who is responsible for making sure it all works? There is also a question about performance versus oran trade-offs. I have heard in other contacts that it is really hard to come to the same performance as our current network, and i have heard the opposite, too. That oran can help with speck officialsles. How do you put all these varying demand on the network in terms of testing together? Ivan, you go ahead.
TBD
I just want to draw on the experiences that we are currently having. As i said they are both otic, meaning we are hosting these talks. I think there are actually two issues here. One is having the environment where a player can play to the stuff or show the things are working. Some is who can certify what they are promising there? I am not saying that one place cannot do both things, but certainly there are different objectives, and therefore probably it shouldn’t be a single place where these two things are happening. Then the second aspect of the whole problem is the fact that we are talking about a number of non-trivial environments. Vertical things that we are not used to being viewed as a target for various exoachts in the system. Then this idea of having a single place where you are going to verify that everything is working according to plan is probably not feasible at all unless you are talking about something very sophisticated, meaning something that sort of extends beyond a single physical environment. Given that we are pushing performance further and further away, chances are that that is not necessarily the oran. That is something you touched on, where people are saying oh, no, in this particular case, function roos are way better in terms of performance than fully programmable software is. There is a range of things that we need to be testing and arguably a whole range of environments that need to be tested. The test beds need to be available for people to test things. A completely independent issue is who certifies you are v.n.f. Is based on 3gpp specific cation or oran specification be, and it is working well with somebody else’s core network. Actually the number of combinations we can act ply play with makes this a somewhat non-trivial integration problem.
Monisha
Do you want to go next?
TBD
Sure. I was about to piggyback on what you said. I think it is the whole oran community’s responsibility. Not just the carrier, the vendor or manufacturer. Testing challenges, they can be a threat when it comes to open iran. When it comes to testing on wide scale, it is difficult to identify a peels of hardware operating on a speck twh there are so many vendors evaluated. That doesn’t mean it can’t be accomplished. It adds a different level of evaluation that needs to be done, and it is a layer of complexity that can be resolved by eliminate proprietary interfaces and standardizing with mature technology.
Monisha
Thanks. James?
James
I was just going to say — make two points and piggybacking on what ivan said earlier. All of the component in the system have to perform to 3 gpp specifications whether they are oran or not. That is a given. That ensures a minimum level of q.o.s. And compliance to f.c.c. Regulations and or standardization bodies. Oran doesn’t alleviate that. Then just as a general point, lori’s point earlier about the oran community having to police itself. I think the oran community does believe that we have to do that . At some point these interfaces, it is just going to be accepted that they work, and the testing will be light. But i don’t think in this time frame, the early time frame, that it is going to be a little bit of — it is going to be a challenge to ensure interoperability. Bhu we all have a vision, i believe, of getting there.
Monisha
Manu?
Manu
Sure. I think — the time is not near , not putting words in lori’s mouth, but anywhere near a production operator environment a test oran element is going to be simulated. The way the big four data center hyper scale networking operations work is real time de-bugging, fixing, feature addition happens in a live operational network. That time is something that needs to be maybe rethought. I do still agree there is some sort of separate thing. The oran aligns itself through a program or someone else that does shoulder the burden of doing a lot of the interoperability activities. Then we i believe as a community, and most likely the operators have to think about what is the way from somebody hoe is work negligence academia and then comes together to a start-up, comes to a forum such as oran, who validates what they have done, meets all the performance q.o.s. And metric and p.p.i.’s put forth. Then what is the on board looking like? Right now each operator has to self-define one and works well for their onlyives. But i do think there is a place for a testing entity either spread across — and maybe it can also be topical, but something where the operators have a front row seat into observing what is anding in their own production networks in order to see what might make that jump or make that shift.
Monisha
Go ahead.
TBD
I wanted to make two points here. I think the first is we talk about interoperability, but i think it is more about that. It is about continuous integration because features do change. We can get initial interoperability, but six month later there is a new future that a customer wants to bring in, and that may affect the hand set, the r.a.n., the code system, and you baverbingly have to figure out how it all get integrated before it can be rolled out in the network. I don’t have a solution to offer, but as an industry we have to think through how we are integrating and rolling out new capability the on a regular basis. I do think doing that does require a lot of technical expertise. Whoever is doing the system integration, whether it is a third party or the carrier, you do require a fair amount of technical expertise to pull that off. I think that is going to be a big challenge for smaller care ehlers. I think one of the comments made early on was that programs smaller carriers or rural carriers who have chinese equipment should look at using other equipment. I have worked with rural carriers in the past where the entire network would be one person, and he would have a c.c.e. Operation. He wrob like an i.t. Person. Now if we do to this and say make this oran equipment work. It would actually not work out. I almost feel that the open architectures we are talking about actually will lead to more consolidation in the industry. The example manu gave about the integration, what we have seen in the data center industries is you only have four big companies who can run — four or five large companies who can run data centers in that fashion. Everybody else is out. Either you use a cloud from one of these folks or you are not really in the business anymore. I think we have toe think about as we are moving in this direction what is the implication for smaller carriers in the market. There are a lot of people that have mentioned what does it mean for the big carriers he themselves as far as how they are transforming themselves to benefit from this technology direction. Those are just things to consider.
Monisha
Thanks. Did you have something to add?
TBD
I just wanted to make people aware that there are different kind of testing. We have specific issue testing. And we have a specific use application, and you want to make sure the k.p.i.’s are properly met or not. You probably need the function amount of his cloud. You are sending lot of bandwidth, and you may not need that. I get that mapping is important when you try to do some test can’t. We have been doing a lot of stuff in the power test phase, the use cases. We are developing some testing procedure as part of the testing initiative own the oran technology. There are different kind of of testing we are talking about here.
The question in my mind is with all this desegregation, we hope that networks do not become more difficult to maintain and up why grade — and upgrade. But i heard a couple of statements that when you add a new piece in, what ask that do to the network? What are your thoughts on does this desegregation actually make the network easier to maintain, or does it make it harder to maintain, or is it somewhere in the middle? How do upgrades get rolled out and how do you test them? Face it. As soon as you move to a software world, you are talking about constantly living in a better version. We are talking about software released without being fully tested. So there has to be a way to think about how you do this network upgrade and maintenance in real time in an ongoing fashion. Who wants to talk that?
TBD
I can give a shot at it.
Monisha
Ok.
I think that one of the things, and they aren’t the only company looking at this, if you look at the network, instead of doing end to end testing, there will be nodal testing on different suspects of the network. Then there are devices that you can hang off the network that also will be able to test any particular upgrades. So i think it is evolutions to existing test equipment that is out there. But as long as the test equipment has these interfaces, then they can test the individual elements that are being rolled out in a controlled. And then the test and measurement salukises can scale to end to end and even down to the confuse itself. That sort of mitigates the transition costs for these particular upgrades. You can look at individual pieces, test individual pieceles of the network, even if it is deployed, and then the solution scale as you you are allowed it in, any additional service or feature.
Monisha
Anyone else have any thoughts to add to that? Cheryl?
Cheryl
I was just going to add from a software perspective, going back to the comment about a.i. And m.l. I think there is a real opportunity that did he segregating and putting everything into software, once you get your a.i. And m.i., your autonomous systems up and running, the opportunity in the cloud is once a vulnerability is identified, you a identify when things are going well, and it is attached in real time. That is something that we think is absolutely critical when you are up against with the threats and the speed of the threat that are coming. It is an opportunity as this matures and it moves to desegregation to the port that need to be tested.
Monisha
We have just a few minutes left. What i would like you guys to do is just — i will go in reverse alphabetical order and ask you each for one research challenge that keeps you up at night? What do you think this community should be looking at?
TBD
I just want to stay i really like reversal fa bet cal. The — reverse alphabetical. For me, research challenges don’t keep me up at night. Those are the things that make me sleep at night. Just to try to answer your question a little bit more meaningfully, i really think the whole idea of open r.a.n. Is interesting from a research perspective. I do come back to the complexity of the system as a whole, the implication little of that specifically with — or in connection with security. I will go back to saying i think the way we need to get there is through experimental evaluation and use cases. No, i’m not staying up at night. I’m very excited about it.
Monisha
Great. Hope you get a really good night’s sleep tonight after this forum.
TBD
I would say complexity first.
Monisha
You can repeat.
TBD
I am going to go to the next one down the list. What we can learn from the development community. Can we actually come up with some sort of a resilient continuous deployment and avoid stuff like complete collapse of networks, which happens to all the big software vendors in this universe. I don’t know. Can we live for a few hours outside the system? I don’t think so. Nevertheless, maybe we can learn something.
Monisha
Thanks. James?
James
There is nothing that really — i can’t point to one item or one thing that keeps me up at night. But i do think there is a need for a little built of patience, because right now s.a. Networks are just now starting to be deployed. I think you’ll see the benefits of network slicing as well as open r.a.n., at least initially some of these early deployments are sort of enhancing the enhancement of a broadband case. But as we move forward in the short-term, i think that once these networks are deployed with these interfaces, i think that it will create significant economic opportunity. I just don’t think that we are there yet.
Monisha
Great.
Manu
I am part of a new company, and they are completely folk used on the r.a.n. Product, and folk used on the economics. I feel that radio access networks are access networks, and they are the most important thing — the most important can’t is the economics, the cost per byte that we can deliver service information it works, it will create a lot of opportunities. That is what keeps me throughout the day and late at night working on it.
Monisha
Manu, i think one of the last things you mentioned, the cost per bit is not something that the research community has focused on much. But the broadband platform will set that right.
Manu
quickly i would say one is incentivizing and attracting the academic community. If you look at these forums and aligns that are out there, i encourage you to look at the metropolitan roster and try to identify a single u.s. Ack kembaic institution there. You will be hard preparesed to find one. That is something to think about or to consider. The second is availability, and that might be tied to availability of software. The dearth of truery open source — of truly open source things for a student to absorb. Even for a single deep dive, some of of that hardware that we in the bubble system are trying to put effort into building up a repository where it is available unencumbered through some of the i.s.r. Issues that we haven’t discussed are there to bring to the fore. All that is tied to the work force. All these companies out there need to hire the students coming out of the the academic institutions. The pipeline is not being built as you are deploying oran, i want to make sure we emfasties that and build this pipeline. I didn’t mean to sound ominous. This is a great time.
Monisha
Lori?
Lori
i am lucky like my peers that this is not what keeps me up at night. One of the things to keep in mind is this isn’t a one size fits all network architecture and i think that the research will help bring us together from an iment be plementation and standards perspective. 5g is really interesting. 29 different from 3g and 4 g in that it is focusing on what they can be to be stiegally important. They could be different. More research we get out there, testing and standards, it will help bring us together for one common goal which is to escalate oran, to bring it to fruition and to make sure there is an importance that it plays in the ecosystem.
Monisha
Thanks.
Lori
I break it down to two things. One is what keeps me up at night. The second thing is what keeps me motivated. Testing and implement as, how to make it secure at the same time providing right services. Those are the important things that keep me up at night. But the things that keep me motivated are understanding that use cases of different verticals. 5g has lots of opportunities. We really need to understand each vertical and map it to the 5g stack. Last but not least i think the collaboration among researchers, regulators, operators, vendors, build up a proof of con accept, try a test bed, go back to standards and create a process. I think that is very important.
Monisha
Absolutely. Cheryl?
Cheryl
Again maybe not something that keeps me up at night, but looking at the opportunity space. We like to focus on the what and the application. The 4g example is great, but with 5g being not just an evolution of 4g, but a revolution. During the idea whatever it is, you want to do with these, particularly with private networks and then testing it out and not letting the policy rely strict — policy restridget this — restrict the ins and outs we need to do. Looking at what the department of defense is doing with 5g pilots, that is a great model. You decide what the use case is going to be and figuring out the gaps of where we need to go the. We will all be there moving right alongside them. Thank you.
Monisha
Thank you. Charles?
Charles
I will tie the scale of evolution that we will see here. What i mean by that is code that a p.h.d. Student is running on a test bed that has three users that runs just long enough to do the data collection before it crashes is very different than code that will run on 100,000 cell fours supporting 100 million users. It is going to be incremental. It is going to be exciting to see this roll-out in the coming years.
Monisha
Thanks. Last but not least. You are there.
TBD
I had the honor to start this, and now to end it. I guess the discussion that we had last on the integration challenges and however to make sure of the different parts provided by different vendors, that is a problem. Now it doesn’t keep me from sleeping at night. My wife says i always go to sleep first and then i senior. But that is it a serious problem that we need to think hard about. On the other than hand, i think james was the one who mentioned that the economic forces, once this becomes a real opportunity in the market, and a little we go, it will, then either the current company will evolve into providing these services. Or maybe putting my hat on as an entrepreneur here, maybe there will be some start-ups that will be the experts in that. That sounds like an interesting possibility that. Will pick up that and do it. But that is a gap that needs to be filled one way or the other. However, i do have the confidence that it will evolve balls of the economic — because of the economic forces here and the opportunity for making money, i suppose. Back to you.
Monisha
Thank you all for your insights and your expert opinions. This has been a great day. I don’t know about you guys, but i think i learned a lot from each of the panels and from all of you. We are at a really exciting time, just changing completely the way mobile architectures get deployed. We didn’t have time to touch on one of my pet topics, which is we always tend to think of cellular and wifi as those two systems who don’t talk to each other, and oran can be brought in to do better sharing of licensed, unlicensed and spectrum sharing in general. But those are topics that will continue to evolve in the academic as well as the industrial community. With that, thank you all again very much. Thanks to the audience that tuned in. That’s it then. I am happy to close this event at this point. Thank you all