The Telco Edge Cloud, ORAN & 5G
[00:00]
Abe Nejad: New edge compute solutions will help realize application centric networks and realize the path to near term edge use cases. On this executive broadcast leaders in the 5G, ORAN and telco edge cloud sectors will illustrate proven edge infrastructures and near and long-term use cases. Starting this panel is Juan Carlos Garcia, Senior Vice President of technology and ecosystems at Telefonica. Next is Cristina Rodriguez. she's Vice President and GM of the data center group and the wireless access network division, that at Intel. Next is Shamik Mishra, he's Vice President and global industry chief architect at Altran. And next in line is Saad Sheikh, he's chief architect of networking cloud orchestration, edge and 5G of the Saudi Telecom Company. And finally, we have Roy Chua, he's founder of AvidThink.
And everyone, welcome to the program.
Thank you.
Abe Nejad: Thanks for being with us. Cristina, I'm going to start with you, let's sort of set the stage for how the growth of 5G wireless technologies are necessitating approaches that include edge computing architectures. What is the real correlation between edge, cloud and ORAN developments leading to converged network designs and robust 55 networks?
Cristina Rodriguez: Great, great question Abe, to start with. With 5G we're going to see more compute push out to the edge, to the locations where are more suitable, more capable to deal with the latency, the bandwidth, the security, the privacy demands of the application. With that, now the radio access network becomes a key element in the entire network infrastructure for the operators to deploy and take advantage of that compute at the edge. The edge has the same type of requirements of flexibility, scalability, agility, that the rest of the network has. The requirements that drove the transformation of the network is starting at the core in the first place. So, all these together, you take Open RAN initiative, which is precisely driving the same concept, hardware and software disaggregation, open standards, open interfaces, general purpose compute platform, and the operators are looking at that and say, well, you know, this is going to bring me benefits that they are embracing that, and those doing that will be able to reap the benefit of edge computing, delivering new services at a lower cost. As you know, Intel has been at the forefront of this transformation, this industry transformation, driving the shift to a general purpose server base kind of network architecture from the core, but also steadily to the edge and to the access.
Shamik Mishra: It's the nature that the new network technologies like 5G [inaudible 03:26] 5G as a core and the new architectures that are based on SD-WAN based programmable transport, and the fact that all of these solutions are being built to run on software based and cloud native architectures actually give us a really good opportunity to look at edge cloud differently. In the sense that, can we have a common architecture that supports both applications meant to run on the edge? For example, consumer applications like gaming, industrial applications like industry photo augmented reality applications, and in the same architecture, can we also run the ORAN, the 5G as a core SD-WAN based programmable transport. The answer is most probably it is yes. The reason is that the cloud native architectures are quite a proven architecture for the application ecosystem. It is still being proven for the ORAN and the 5G core network elements, but it's not far away. The fact that these kinds of architectures brings in a lot of efficiencies and agility to the network deployment. And there could be a common model of operations and management for both application developers, as well as for network vendors to bring these out bring their equipment to the to the Telco edge cloud makes a lot of sense economically, because of the simple reason that the operator can have a common architecture to support both internal as well as external applications. The second important thing is that the standardization of the runtime, for example, the hardware and the software, the fact that cloud native has made it possible already, it means that the IT infrastructure can actually be scaled out to the to modernize it, or I would say to transform to carrier grade clouds. And that would mean that we could take the edge as a smaller version of the central offices, it would also be the mobile network nodes. It could also be an application host, it can also run on the customer premises. And that's the growth impetus that the operator needed in 5G, so they basically have to consume 5G and make money out of it. And these new architectures would actually enable that.
Roy Chua: So, I think as the other speakers have talked about, I think one of the things around 5G is the use of enterprise and cloud type technologies and architectures and what it allows us and more of an opening of the architecture. And so, what we're seeing is an architectural convergence between edge and cloud and 5G. Now, what that comes with historically is virtualization disaggregation. You know, all the cool things that we see on cloud architectures. Now ORAN is yet one more of that. So, I think historically, those of us in the Telco space have looked at the SDN movement, software defined networking, NFV, network functions virtualization, that's been a journey to basically disaggregate and virtualize all aspects of the Telco network. The RAN, the radio access network is I think one of the last bastions of that and so ORAN or open RAN is that disaggregation and virtualization of that same thing, the same way we did SDN and NFV along the way for all the other parts of telco infrastructure, you know, ORAN is to continue expansion of that cloud type virtualize elements that does that. And edge in my view is the platform that makes it all possible, right. Edge has distributed computing at all locations that powers effectively the new next gen telco network.
Abe Nejad: And Saad, the correlation between ORAN, edge and cloud in your perspective?
Saad Sheikh: Yeah. What we believe you see, as telecom operators, we have started this journey, I think, now almost a decade. And we have learned some good and some not very good experiences, real challenges during the NFV commercialization. And we see over time that the cloud model has changed from the central cloud to the regional, and now to the content or the edge distribution. So, naturally, we think edge as the natural extension of the cloud. And we really need to harmonize and make use of all other investments as well as the most important experiences, right, because we have seen that either from VR people, or NFV as [inaudible 07:59] to some area where we now see the commercialization of the technology, right. And if we see the pure vision of the edge itself, I think it's very important to see that for the edge application itself, considering a neutral view of a developer or a third party, how they are hosting today, right? Maybe on public clouds on the cloud itself. So obviously, from the attraction of the developer or the community, it's vital that it's developed on the cloud native principles and harmonized the cloud or the vision of the cloud itself, right? When we see about ORAN, I think theoretically ORAN is a different driver, which is the total can be salvaged. But, we all know, the most important as we are stretching the over cloud to the last mile to the 5G side, the most important thing is to have an efficiency on the RAN itself, because really, the cost of the RAN boxes are not too much as of today. So, how to achieve this RAN automation or intelligence, that is something obviously that requires to follow the crowd principles. I give you one example, right, for example, the latest drive to the [inaudible 09:21] right and the orchestration, this all is actually aligned with the cloud again. And the further optimizations where we want to use the use cases like the AI and the machine, right, that is actually based on the open [inaudible 09:37] and the open data platforms that they can only see from the cloud. So, the [inaudible 09:45] how we see that they have this digital transformation. There are many streams, right? And I think around a decade ago, many operators, they drive many journeys in parallel, like there is a VN [inaudible 09:58] on the hardware itself and the cloud and orchestration and [inaudible 10:03] it has to converge, at least on the app [inaudible 10:08] because we see the most important thing for technology commercialization is to find a common deployment model, you know, because we cannot touch the [inaudible 10:21] different times for different technology. So, the harmonization and the agreeing on the commercial deployment model, I think that is where I think edge, ORAN and the cloud totally harmonizes and deliver value.
Abe Nejad: Juan Carlos, I wanted to ask you, what are the real Telco challenges in addressing the edge opportunity from a technical, organizational, and commercial challenges?
Juan Carlos Garcia: Well, that's another great point, Abe, and with all the issues that you have mentioned, are actually challenges for telcos. So, the first one technical, because network experts need to embrace the cloud philosophy and practices to manage the new technologies. And that is not a straightforward, second organizational, because [inaudible 11:05] have been traditionally independent, and now are becoming deeply coupled with network teams becoming new customers of the AD departments. For instance, the radio team requesting a cloud environment for their distributed unit and central [inaudible 11:19] units and the [inaudible 11:20] becoming the new architects of our networks. Third, the commercial aspect, because new services can be launched faster and will have shorter and more dynamic life cycles, moving from a few generally services for millions of customers as of today, to multiple vertical specific ones, addressing thousands of customers each. That means in different ways of marketing and selling, especially as sales force, and new channels as well, including a great deal of self serve, and platform oriented approaches. And finally, the management and orchestration, multiple clouds, from on premise, to the edge cloud to the public cloud is an issue that still is not totally solved today. But, this is becoming more relevant when new cloud flavors like the different types of edge clouds are added to the multi cloud environment. This is both a challenge but also an opportunity for service providers like Telefonica, that can help their customers in navigating through this complex portfolio of options and support them in designing and supplying the most optimal combination. Well, that's my view. And I hope it answers your question, Abe.
Abe Nejad: Absolutely.
Shamik, I will be coming to you about the objectives and goals for the Telco edge cloud. Before I do that, Roy, I wanted to ask you about who are the potential collaborators in the telco edge cloud ecosystem right now?
Roy Chua: So, I think it's we've heard , you know, Saad, Cristina, Juan Carlos is that there's this convergence. And when you open up the technologies on the mobile enterprise side, you open up the ecosystem as well. What we're seeing on the outside, if you look at the network edge, right, the telco network edge, you have the telcos certainly, you also have the hyperscale cloud providers that are coming in, as well as co-location providers, right, the Equinix’s and Digital Realty’s of the world. They're looking at different ways of collaborating on that side, right. On the radio edge, you have certainly the carriers, the mobile network operators, the MNOs, but you also have the tower companies, some of whom are interested and have designs on finding ways to add compute or capabilities to those locations, right, at the cell towers. And then moving on to the enterprise edge. Now that ecosystem is a lot more dynamic, right? You have the telcos coming in, because historically, they've had some kind of customer premise equipment or CPE there. You also have the hyperscale clouds, right, whether it's Google or Azure, or whether it's AWS, they all have designs on that place. And some of them even have offerings in the enterprise edge already. But, beyond those players in the telcos, you also have traditional ICT vendors that are interested in coming in, and the software platform vendors, so your HPEs, your IBM, your Red Hat, your VMware, they all have an interest in that space. And certainly for the system integrators they look at that as a great opportunity. But, it comes down to you know, who owns the customer relationship, what unique values do they bring, right, who owns the real estate, who actually has expertise and economies of scale to leverage and benefit from all that. And that's still in the process of being worked out as we as we see. But, a lot of partnerships today. You see, telcos, hyperscalers, software vendors, ICT vendors, announcements all over the place, and I think it remains to be to be seen how it's going to play out.
Abe Nejad: Shamik, to you, what should be really the objectives and the goals for a cooperation towards a global telco edge cloud?
Shamik Mishra: Right. So, many operators today, whether they are global service providers or for that matter, even enterprises, will demand a global footprint. And when we say global footprint, it means that they will tap the application developer ecosystem today. And the application developer ecosystem is currently with the public clouds. It's a fact. And that public cloud interfaces or that public cloud ecosystems would have to be brought in some way to the operators. So, there is a need to have a single interface to deploy and manage workloads. And as I said in the earlier, earlier part of the of the discussion, that the workloads could be network functions, workloads could be applications, they are all going to be cloud native, most likely. And they have to be managed across multiple geographies. So, you cannot possibly expect an application developer to make 10, 20 different versions of the same application just because the operators and the enterprises have different ways of managing workload. It's just simply an impossible task for an application developer to maintain that level of, you know, diversity in a single application. So, the first thing that the global telco edge cloud should actually try and achieve is to figure out whether they can build this common interface for application developers and then provide a larger scheme in terms of their reach. The second important aspect about the global telco edge cloud is that the operators themselves need to have a differentiator. For them the telco edge cloud is a beachfront property, because of the fact that it is very close to the user, the end user. So, the operators will have to figure out what are their differentiators which a developer can consume, which the developer cannot consume, otherwise, they can only consume it within the operator network. And one of the critical use case for that is the network exposure functions that 3GBP has been standardizing and edge compute will actually enable that consumption model with application developers. Those could be specific use cases like location, the quality of service, bandwidth, mobility aspects of which the operators are perhaps best in in terms of position to provide that interface or that requirements to the developers, notifications, quality of service [inaudible 17:26] it could be done much easier in a much faster way if the operator does that. Roaming is another very important domain, I mean, when edge computing use cases, particularly on the consumer segment will become normal, and everybody's playing this a game or using an edge compute service. And then people start traveling from one location to another, they would expect the same level of application enablement across operators. So, roaming and how that how roaming actually impacts the operator’s architecture is also very, very important. This is also important for connected cars. For example, if they move from one geography to another, then the ability for the application to migrate from one operator to another is also critical. And only operators can do that if they come together and build this global telco edge cloud. So, there is no doubt that operators will reach that global footprint with agreements and or in some form to ensure the uniform service and facilitate the federation of their edge requirements. The combined work of the GSM and operator platform group, as well as the technical reference framework that we are building as part of the operator platform group and the business requirements that the telco edge cloud group in GSMAs are building, together we will be able to provide the commercial business and organizational models to make that happen. This is perhaps is required because otherwise, I don't see how a global telco edge cloud can really succeed if the operators don't start to collaborate within themselves.
Juan Carlos Garcia: Let me just add an observation so some can make comments. So, they saw a strong demand for this collaboration between operators from the multinational customers, from the B2B segment. We are serving them I would say pretty well in our local markets, but they are requesting also to have a reach beyond our footprint. So, we are [inaudible 19:30] than our neighbors. This would require collaboration with a lot of operators to provide solutions like [inaudible 19:36] that we have delivering now for private networks, and it's manufacturing another sector. And what usually our customers are not operating in a single market. They are really having operations worldwide. Yeah, and that's an important point to consider.
Abe Nejad: Cristina, I want to get back to you. So, where does the network differentiation reside for operators in the telco edge cloud and what real business or technology initiatives does this serve?
Cristina Rodriguez: No, very good question, right? Well, this is a new area, right, and we are going to learn a lot in the next few months and the next year, and we have to take all those learnings and this is an industry effort, right? So, we have to take those learnings and be agile in applying all that shared knowledge.
Saad Sheikh: I think for this we actually want to play differently, I think first view is very simple, right? Because over key differentiation, right, it is somehow defined by the physical containment of what the Telco operators are owning. And, obviously, the VC, the [inaudible 20:56] distribution network on the RAN [inaudible 20:57] on the site itself, right? This is where already the telecom operator is right. So, it is definitely one of the things that users are already using and building anything additional data leverage on this, it means that we can really breakthrough, right? But, [inaudible 21:13] to the other players in the edge, because if you see an overall situation, the public cloud, they are already offering the [inaudible 21:22] I think almost three, four years, right. So, we really need to see that for example in [inaudible 21:29] there are some other variables, which I think is very vital to offer a carrier grade add service, for example, we see the identities, for example, you see the public cloud are ubiquitous, right, but at the end of the day, when it comes in terms of ownership or contractual payments, right, they somehow are not managing in a way, and maybe by the geographical boundaries, or by the other constraints, security and data governance is again, one of the things where I think if over exhibition, we define it, this is definitely one of the things. Then the quality, I think that's very, very important. I do agree that today, the telecom operators, they have not gained some success on the edge [inaudible 22:21] Even before these edge applications, you know, when we talk about a distributed cloud, this issue is still not well addressed. And we think that in the coming year, there needs to be a special effort to see how we can manage a, let's say, a massive scalable edge network, how to orchestrate it, how to manage it, right? In a pure cloud companies, you know, using the advanced [inaudible 22:49] or cluster management, they are doing it but it's not carrier grade, right. So, this is, again, that quality, it comes to SLAs because for a customer I think if you need to see differentiation, you must think about the essence. For example, even if you own OPG and DCMA initiatives, we have been especially emphasizing to define at least those sort of metrics, which we know that only telecom operators can provide. One of the main reason is not just connectivity, is end to end orchestration, because you just see the overall picture of this, right, we have the edge, and we have some central cloud and orchestration and the maybe underlay network, the MPLS internet. So, to meet this end to end offer, I think this is where we need to really play. But, this part is about the physical containment as I see. I think the other part is a wide open question, right? This is where we really want to play differently compared to over 3G, 4G timeline, right. For example, how we can really become the enabler of this service. Juan is actually with us and I actually raised this even within the community as well that we need to touch the [inaudible 24:05] but the SASE part because we see the right now or enterprise customer, their requirement is how we can bundle and give them end to end thing. And even they expect, you know, we have a good offering for the developer themselves. For example, what they require is a is a complete solution with the developer SDK, the testing tool stack and the others and the most important thing you see currently edge has a lot of constraints, right, edge site has a security constraint. It has network connectivity, it has delay issue, it has a lot of other things. I think to really play a part in this is very important that as telecom operators we play in the software value chain, because many of the issues [inaudible 24:51] the cloud or the I mean, the traditional [inaudible 24:55] cannot solve it but we'll see if we take an application and solve some of the issues through cloud native itself, for example, so the declarative API's and the open telemetry. And so, I think that's an important and especially I think, as we scale and even when we find a wholesale model, where the services are moving, it's very important to play on a I should say, upper layer on the open API and SASE. I think this is where I think we can really make differentiation, especially when we talk in terms of public cloud because I must admit they're too good on this. And this is where we really want to emphasize, we are trying to bring them in, in this journey. And I think this is here, the focus must be because otherwise just focusing on the physical containment, I think, when we go to enterprise and private edge data differentiation is not going to create a long term value. So, that's some key finding or direction so far.
Shamik Mishra: Yes, the Telco edge cloud is quite deeply integrated in the network, right? It's, well inside the operator network. And they, as I mentioned earlier, they need to share this space with the network functions and that give access to the, you know, the different access networks and the core data planes, and then provide secure connectivity between the customers who are say enterprises on their premises edge data center to a network edge residing in the operator network. And apart from that, as I mentioned earlier, the network as a service model, which the operator will need to define, in terms of what kind of API's they can provide, to the developer to the enterprise's to leverage, the network differentiation is extremely important. This could be, you know, this could be a tariff or a new revenue generator model for operators. It could also mean that the application developers would demand these kind of services where they get some kind of enriched data on the unpredictability on network. Say, for example, if I'm an application developer, and I would get to know that the network might fail at a certain point and I need to move my workload to somewhere else, that would be great. But, will that really happen today? Most probably not. But, the operators can do that by providing additional capabilities to the network as a service APIs. Similarly, bandwidth, location, there could be other network data analytics, which the developers can consume. So, eventually, for the operators, the telco edge cloud is not just a technology initiative, it's a business initiative, because they would end up monetizing these APIs and capabilities, which they can't do today, because of the over the top model.
Abe Nejad: Roy, is it a business or technology initiatives that the Telco edge cloud serves?
Roy Chua: Yeah, so I think, as Shamik indicated, right, they own that beachfront property, that last mile, that last kilometer, and they control that. So, I think, you know, from that perspective, certainly, they can own a quality of experience, they have a trusted relationship. It is a secure network. And so the Telco can bring that to the table. And I think, you know, from a network deficiency standpoint, I think it comes down to how do you leverage the network? How do you monetize the network? How do you make it available for the enterprises to consume in a way that makes sense, right, that aligns with the application and business goals. And so, I think the differentiation there is not just the quality of the network, and the real estate and the proximity that they have to enterprises, but also their ability to expose that differentiation by APIs to the potential partners, right, the hyperscale cloud providers and to the enterprise's, either directly or indirectly. And what in the end, it provides enterprises is a more dynamic network, one that's capable of blending better with applications, for instance, if you're doing unified communications, maybe you want low latency, you're trying to do bulk transfer in the night, you know, you can pay cheaper price for more bandwidth, right? So, there's all these opportunities that are available for the enterprises, and I think the Telcos do have a role to play in that ecosystem.
Abe Nejad: Juan Carlos, I want to get to some near term or long term use cases for the Telco edge, can you give us an overview? And also talk about the importance of the connectivity side of that as well on the Telco side.
Juan Carlos Garcia: Okay, well [inaudible 29:45] cloud is deeply integrated with the network, it is residing on the edge of the network. And it is because of this providing direct access to the access networks on the core data planes. And this means it may offer a secure connectivity for the customer between their premises and the edge data center. And this is the something extremely relevant, not just for the performance and the reliability of the connection, but also for the security, because the communication is never leaving the network and he's not going through the public internet. So, this brings an external [inaudible 30:19] of reliability and security to the network. But, coming to the use cases and of course, one of the ones at this moment is the application for [inaudible 30:35] networks and very recently we announced one of our agreements in Spain with a manufacturing company, Gestung, that is building parks for the automotive industry. What we were providing the 5G connectivity and also all the edge computing that they're required for the digital twin application. Okay, what they have been doing is just replicating the physical factory in a digital twin that they use to design, refine, monitor, and optimize the processes inside the factory. Well, this represents a strong a big amount of information going from sensors, cameras and different devices in the factory going to the edge to create this a digital twin, this replication in the digital world of the physical world. And this represents a lot of information that needs to go through the network very quickly and reliable, reliably in real time with very low latency and then massive amount of data storage and processing in the cloud very close to the to the customer. This is one of the applications, there are many others. In reality, we have remote control of robots and machinery end to end tracking of assets, full sensing and monitoring of processes. So, it's a full range of different applications that can be provided from the edge to optimize and digitalize the processes in the factory. There are other cases like the surveillance etc. But, in general, the Telco edge is going to help our business customer to further overload the customer premises in meeting the complexity of devices like gateways, sensors, cameras or communication CPs underneath for on premise servers. And this is very relevant for companies that want to digitalize their processes and optimize their production in the production and operations. From a private network will be [inaudible 32:48] first experience is on industrial edge use cases, while the telco edge and the network slicing are made available. Okay. That will be the probably the next steps towards to offer to business customers network slicing as a compliment to edge computing to build not just computing capacity, but also a full digitalized network for the customers, especially for the customer.
Roy Chua: You heard Juan Carlos give a lot of examples already. So, I'll just emphasize a couple and add a few more. So, in the Telco edge, certainly there's all these cool use cases, you know, industry 4.0, either on premises or the Telco edge, because perhaps you know, you don't want your data center running in the factory where it's hard to maintain maybe dirty, or precision agriculture where you know, perhaps at farms, or in certain locations, you don't want a rack of servers you know sitting out there, you're next to you know, next to your cattle, right. And so you can provide edge capabilities in those Telco edge locations, for your certainly use cases along those lines as we've covered some of those. But, beyond that there are cool capabilities like for media entertainment, right, obviously, we talked about cloud gaming. But, there's also things like the ability to do remote production. So, you can have camera crews running around and everything's running at the edge. You don't have to have those big TV trucks anymore, which you may be familiar with. But, you could do that all remotely over the Telco edge. And then on the enterprise edge, a lot of use cases around industry 4.0, around healthcare, as well as other security, video surveillance related use cases. We're seeing a bunch of those across the board and beyond enterprises, smart cities, governments emergency response, those are all the use cases that we're seeing that are either leveraging the Telco or cloud edge as well as the enterprise edge.
Abe Nejad: How will these use cases that Juan Carlos mentioned, how would they transition to an expected operational transformation, if you will, for the telco edge cloud?
Shamik Mishra: Right, for the for the Telco edge cloud in principle this is an operational transformation. And if you look at it in multiple ways, we'll see why. First of all, the same Telco edge cloud architectures are going to support applications, which the operators haven't hosted before. It's a third-party applications. The cloud companies do that on a regular basis, the developers work with them, the operators will now be running these workloads on their on their network. So, that means they will have to transform the way they onboard application developers, the tool chains they provide, the CICD pipelines they provide, the developer environment they provide, the sandboxes they would provide. So, it's definitely an... it's a transformation in terms of architecture. Now, whether this is going to be a transformation for the network itself, yes. The reason is that there will be many Telco edge clouds, right, I mean, there cannot be just a few data centers anymore, there will be hundreds, possibly thousands for an operator. So, the scale of the Telco edge cloud will be massive, even for a single operator. So, for them to manage such large scale distributed cloud with so many application developers working on them, their own network functions like Open RAN, programmable SDN, 5G core all running on them, it would mean a massive transformation for operators in terms of how they manage this, manage all of this. Also, the fact that 5G and Open RAN would introduce multi-vendor ecosystems, that would that would mean that there will be multiple vendors in on the network alone. So, the operators most likely will have to invest in, in automation systems, possibly driven by artificial intelligence because it might be nearly impossible to maintain the network, provide predictability in the network, maintain carrier grade standards, maintain a developer confidence without being able to do things in a more predictable manner. And for that, I think artificial intelligence based operational transformation would become mainstream, and the operators will have to invest in it.
Abe Nejad: I wanted to thank everyone for their time. Juan Carlos, as always, you always give us your time. We really appreciate that. Saad, it's the first time we were doing this. And I appreciate your answers. Roy, sorry, we didn't get to as much as we'd like to. Shamik at Altran actually made all of this possible. So, I wanted to recognize that, wouldn't be having this session now if it wasn't for him. And then Christina of course, we'll connect real soon and we'll get you all set up. Okay?
Roy Chua: Thank you.
Cristina Rodriguez: That's great.
Shamik Mishra: Thank you.
Abe Nejad: All right, you guys. Thank you. To all of our viewers out there, thank you again to our speakers on this executive broadcast called the Telco edge cloud. For this broadcast on demand and all of our executive sessions, please log on to thenetworkmediagroup.com. So long.
[38:09]