Viavi Solutions Inc.

Hello, everyone, and welcome to this latest webinar from IHS Market titled, Serving the Mobile Transport with NG PON Technologies. Mobile operators need to address the strict performance requirements of a demanding mobile transport network. So today, our panel will explore the various challenges and detail how next-generation PON technologies can be leveraged to overcome them. Our webinar is co-presented by IHS Market and our partners, Nokia and VIAVI Solutions. My name is Alan Tatara, Senior Event Manager for the IHS Market Technology Group webinar team, and I want to thank everyone for joining us. So before we begin, I want to highlight some features available for you that are available on the webinar. So the console that you're looking at, it is completely customizable. So this means you can open, close, move, or resize any of the windows that you have open on your screen and arrange the console any way you like. At the bottom of your console, you see a number of application widgets which contain additional features available for you. I do want to talk about the resource list widget, and this is where you're going to find additional material about our topic, including the downloadable slide deck from today's session, as well as other valuable information, including a special report authored by analyst Mariana Angelou. So make sure you check these out during the webinar. You're also going to notice a Twitter widget, and this means you'll be able to tweet directly from the console. And today, we're using the hashtag mobile transport. We will also have a live Q&A session directly after our presentation, so please remember to submit your questions or comments at any time by using that Q&A box that you see on the left side of your screen. Also, this webinar is being recorded, and the on-demand version will be sent back to you within about 24 hours. And if you encounter any technical issues, just click on that question mark widget and you will get the answers that you need. So now let me introduce our panel. First leading our discussion is Dr. Mariana Angelou. Mariana is a principal research analyst in the service provider technology segment at IHS Market. We are also joined by Ronald Heron. Ron is a lead technology strategist on the fixed networks CTO team at Nokia. And rounding out our panel, we have Dr. Reza Vaez-Gayemi. Reza is Senior Manager of Product Line Management at VIAVI Solutions. So welcome to our panel. Pleasure having you with us today. So let's get started. Mariana, I will pass the controls over to you.

Brilliant. Thank you, Alan. And thank you all for joining. So over the next hour, we will explore some of the market trends driving mobile transport networks. And we will also discuss how broader networks can support the mobile transport and effectively enable new practices and operating models. A focal point to our session today is, of course, 5G, and Ron and Reza will walk us through the transformation of the underlying transport required to really support the 5G traffic. In the second part of our session, we will introduce some of the benefits of leveraging existing fiber access assets as a viable alternative for mobile transport, and we will discuss also what is the role that passive optical networks can play in the 5G context. We will then delve into the access technologies that are suitable to serve the so-called actual network and touch upon the innovations that are required to really meet the 5G constraints and serve a range of different services over the fixed access network. At the end of the session, as Alan mentioned, we will have some time for Q&A. So if you have any questions as we progress, please enter those in the Q&A interface, and we will address them at the end of the presentation. So with that, let's get started. So first off, what I wanted to do with this slide was basically provide some context around mobile transport that arguably represents a fundamental component of mobile networks. Indeed, in our annual survey on the evolution of 4G to 5G, we asked service providers in which part of their network they see the biggest amount of development required to make 5G happen. As expected, of course, 52% believe the radio remains the most challenging component, but it's followed by transport with 24%, and management at 18%. So in addition to introducing the technologically challenging radio component, for sure, the advent of 5G will drive transformation also in the underlying transport network to serve the range of new applications that are enabled by 10 times higher wireless speeds and very low latency, and ultra-reliable connectivity that everyone is really talking about. But with the introduction of 5G in 2019, IHS market anticipates growth, of course, in backhaul connections through to 2023, but with more fiber closer to the radio endpoint. So indeed, fiber makes up an increasing share of the mobile backhaul installed connections, and it grows from 44% in 2017 to 56% in 2023, actually taking share from the more traditional microwave and corporate technologies. As I said, fiber access technologies will assume a prominent role in supporting the mobile backhaul and it's going to be driven by the availability of that infrastructure and enabled by innovations, some of which we will address later on in our session. But because mobile transport is no longer synonymous only with backhaul, especially in light of the emerging 5G mid-haul and front-haul segments, next-generation fixed access technologies are actually required to cope with the constraints of that evolving mobile transport. So before handing over to Ron and Reza to walk us through some of these challenges of the 5G ex-fall, I wanted to leave you with a view of the next-generation PON activity around the globe. So recently, 10G-class fiber access technologies have begun to ship in volume, and in the long term, Growth will be driven, of course, by demand for broadband connectivity, but also to support the mobile transport. Existing 10G-capable PON technologies are on the rise globally, although we currently, of course, identify the long-term opportunity mostly in Asia Pacific and China in particular. However, we have seen spending increasing in EMEA and South and Central America, on Next Generation Pond in 2018, and we have also seen in the North American market some very positive indications, also with some 10G-capable fibers to the home rollout, either in progress or in planning phase. So, in the context of 5G, and in the race for 5G commercialization, the mobile operators around the world with access to that to a strong fiber to the home footprint will be able to tap into that existing infrastructure and potentially lower the cost of the mobile transport network. So and with that last comment, I will hand it over to Ron to basically talk to us about some of the challenges of the 5G transport. Ron?

Yes, okay, thank you very much. So 5G is coming, and we all know, we all hear about it. It is enabling new promising services that were not possible before with LTE. We hear of three classes, the enhanced mobile broadband, which is like LTE on steroids, focusing on higher bandwidth everywhere. Ultra reliability, ultra reliable low latency communications, which is an enabler of industry 4.0, or mission critical communications, robotics, autonomous vehicles, not necessarily high bandwidth, but very reliable and very low latency. And then the third one being massive machine-to-machine communications, aiming to provide massive connectivity to billions of devices with high availability and low latency. You can see these three categories have special requirements that were different from the previous generation that we need to take into consideration for the access network. Some of those are shown there. The number of things is at least an order of 10, maybe 100 times what we used to deal with with people, number of endpoints. The speed could be an order of 100 times faster. The latency requirements for the services themselves can be much lower than they ever were before. And we're requiring service-level agreements with quality of service for lifeline-type applications. And there's the emergence of this requirement for network slicing, which are virtual networks, logical networks within the end-to-end network. So we'll talk a little bit more about that. So we'll delve into some of these ways that the broadband access networks can address these. But looking a little bit more at the nature of 5G, let's look at the spectrum that is used. 5G supports more bands and more bounce, a lot more. Now, historically, the LTE and previous generations used a spectrum in the low band on the left side of the screen. where, you know, there's good propagation, good reach, which allows for macro cells, but it's a very congested spectrum. Everybody wants to use that spectrum, so there's only small blocks of it available. So the bandwidth was generally small. We're talking 20 megahertz channels with, you know, maybe less than a gigabit type services. But 5G is moving... into the mid and the high band. Most countries in the world have set aside spectrum in both of those areas. The mid band from 3 to 6 gigahertz still has a relatively good propagation compared to the 28 or 24 plus, but still less than previous. So this will be driving the potential for smaller cells. And then even more so as we go to the 24 plus, or 28 or 38 gigahertz per second, gigahertz, the propagation characteristics are much shorter. Therefore, we'll be having even smaller cells. The good news is that these are new spectrum with lots of available bandwidth, large chunks. So basically, the consequence of these two evolutions are that there will be a lot more cells as we go to smaller cells, and there'll be a whole lot more bandwidth to backhaul. So I'm going to pass it over to Reza to continue on this thought.

Thanks, Ron. So talking about bandwidth, now that we know about the requirements in terms of bandwidth, latency, the number of endpoints in the IoT world, what does it mean for the transport network? More specifically, for the backhaul and front-haul network. So what we know, obviously, front-back hole is the piece of network that connects the core with the radio, with the baseline unit. Sometimes this part of the network is extended through the use of small cells, and that's the piece in the middle. And then finally, we've got the front hole, which is this piece that has been very dominant over the past five-plus years.

very light, efficient technology.

The advantage of that is it's light, synchronous. The problem that we have with SIPRI as we move to 5G is the fact that SIPRI technology is really dependent on the number of antennas and signal bandwidth. So the more antenna ports we have and the larger the bandwidth and proportionate the signal bandwidth and the bandwidth for the SIPRI, transport increases. So in this example, if you see on the table, if you have a 64 antenna ports and 100 megahertz of signal bandwidth, then the necessary optical bandwidth on your SIP report is over 300 gigabit per second. That's just one example. There are many factors that play here in a row, but it shows us that CPRI is not going to scale well for the 5G world, especially when we cut to massive MIMO deployments. So what is the solution? The solution is to take another look at the baseband unit, the functional distributions between the baseband unit and the remote radio head. What you see on the lower hand of the screen is the option 8, which is that what is currently known as a SIPRI. Everything to the left-hand side of that is in the baseband unit, and the remote radio head takes the humble role on the right-hand side of the RF functions. Again, it makes it very simple for the radio, but the signal bandwidth is very high. To solve this problem, we need to put some of these functions into the remote radio head. So the current consent is if you want that to put the lo-fi function into the radio, and this is what is then known also as ECP predominantly. Some of the elements on the left-hand side can then also put even farther away. And that will be known as central unit. And then the piece between is known as distributed unit. So if you look at the new functional distribution, over time, we have on the low hand side, the low phi and RF in the radio. And then the middle piece, the distributed unit and the far end. distributed unit closer to the radio, and obviously the radio is on the rooftops and the towers. Just for the reference, I put some of the other terminologies that are used, not only ECP, it may also be known as FX, or NGFI1, which is the terminology used by IEEE 1914.1, and then NGFI2 is the Mithal piece, which may also be known as F1. And finally, the backhaul will also be known as ES1 interface. Now, what does it mean in terms of transport requirements? The transport requirements are very, very stringent for the frontal. That is where then we have similar requirements as with the SIPRI. E-SIPRI and SIPRI have the same, basically, latency requirements, around 100 microseconds, roughly the same latency requirements, 100 microseconds, and a frame loss ratio of 10 minus 7. This is primarily for user plane data. Those are very latency sensitive. There is also control and management plane. Those are more forgiving. Of course, when we design the network, we've got to design it with the services that are the most demanding. That's the user plane. So very demanding here in terms of what we need to solve in latency and frame loss ratio. And finally, the other requirements in synchronization are The synchronization requirements are primarily driven by the time alignment error, which comes from three GPP-associated organizations. When we translate it to the transport networks, it's known as time error, relative and absolute time error. And those, as you see, are very different. If you look at the bottom left-hand side of the screen, that's where we're coming from, the category C, 1,109 seconds in terms of time error. As we move towards more demanding services, a category B that represents carrier aggregation among other things, then you see that the time error requirements become much more stringent. We are getting to 60 to 70 nanoseconds and even more demanding time error requirements.

Okay, thanks, Reza. So at this point, we will delve into the options now operators have at their disposal to address some of these challenges for the Xcode network. So Ron, let's start with you.

Very good. Let me move that forward. The 5G deployment requires an ecosystem of technologies. So these include, as shown there, the radio, the user devices, the core, the cloud, and a variety of transport technologies. We call these AnyHall technologies, which are used for transporting these and include optical AnyHall, microwave IP AnyHall, and broadband AnyHall. So as mobile upgrades are accelerating and going deeper, it makes sense to leverage this deep fiber deployment. And these synergies were already being deployed or exploited by companies by operators for 3G and 4G LTE deployments, as you can see by the quotes from Tier 1 operators. But it is even more important to exploit the fiber now as the 5G deployment goes even deeper and requiring even higher bandwidth. So let's take a look at that deep fiber deployment and how we can leverage it. What you have on the screen is a neighborhood, maybe a couple hundred, four or five hundred homes. And the little green dots are homes. So you can see there's a lot of homes, and in there there's some businesses shown in red, and we've laid on there the large cells, which are the big blue dots, those would be the macro cells, and then smaller cells, which are shown in the little black dots. It's quite clear that if fiber is present to serve the homes, then there's at more than adequate coverage to serve the small cells at a point. So cost-wise, we need to take advantage of that. In fact, there are fiber access, you can say, ticks the box for key requirements of cost, scalability, and performance. So our estimate on the cost front is that by leveraging the fiber that's already in the ground, you know, that is spare fibers that are sitting available in the existing cables, you know, 50% cost savings can be, you know, allowed. And so this allows us to put the investment somewhere else instead of, you know, just digging. Dealability due to the large deployment of fiber access deployment we can quickly scale up the 5G rollout. And then, of course, performance can be easily handled with fiber. The high capacity, the low latency, the automation, and the network slicing, which are being put in place for FTTH networks could also enable the 5G transport. So, you know, continuing on this thought, we did an actual deployment analysis for a particular operator and found that for the large cells, just simple avoided civil works were in the order of $60,000 per cell. And for small cells, in the order of $25,000 per cell. Of course, it was more small cells, so the savings were even greater in total. So this is something that CFOs will certainly be wanting to exploit simply, The infrastructure, let's not do civil works twice. But that's only the starting point. And certainly there's more that can be done by leveraging access technologies. There are a few considerations that need to be taken into account. And one of them is, where are the small cells homing to? Now, what we have there is an OLT on the left-hand side going out into the access network to serve homes, let's say, that would be on the right. Now, if those small cells are located out there, where do those small cells go back to? Do they go back to a macro cell, like spokes from a macro cell, or do they go back to a more centralized location? And certainly there's benefits to going back to a common higher level centralized location, even having the macro cells go back to that higher centralized location. And this is going to tend to occur more with operators that are connected, access and mobile operators that have some sort of looking for synergies. And so, you know, this is just a consideration. And then the other consideration is how much separation do we want to have between the mobile network and the access network? All things are possible. We can have a carbon cable with segregation of fibers. We can have a common fiber with segregated wavelengths, and we can have common wavelengths with segregated time slots. So these are all possibilities. And kind of building on that a little bit, we could call this spatial, spectral, and temporal separation. And whether it's PON or whether it's point-to-point, these notions apply. So we can have spatial separation, a dedicated fiber for the mobile. We can have a dedicated wavelength on the same fiber dedicated for mobile. So there's no interaction of the data streams in any way. And then the third one would be temporal. That's where we actually use a common wavelength, a common TDM pawn or whatever kind of pawn for serving both residences, businesses, and mobile. And we keep the separation through things like slicing. which are things that are being proposed and just simply dedicated bandwidth for each of the applications. So these are all possible and the choice made by operators will probably depend on how integrated the mobile and the broadband access organizations are. And just looking at how these can be applied to a PON network in particular, again on the left you can have a dedicated PON It's not because it's PON that it has residences on it. It can have a dedicated PON serving only mobile sites, small cells. You can have a shared PON where you have a wavelength for mobile and a wavelength for residences. Or you can have a converged PON with all sharing the same TDM stream of data.

Okay, so thanks, Ron.

So before we talk about the applications in more detail, let me remind you that if you would like to submit questions, you can use the Q&A interface, and we will get to those at the end of the presentation. So at this point, we will deep dive into the specifics of using a point-to-multipoint architecture to serve the 5G transport. And in what follows, Ron and Reza will help us understand what exactly that means in practice. So, Ron, over to you.

Okay. So, we'll take a moment to look at the broad families of fiber technologies that can be used in the access and also for mobile transport. Shown on the left is the PON 0.2 multipoint architecture. I've I presume most people are familiar, but in case not, it's simply the notion that you have a single laser at the left side feeding a fiber, which goes to an optical splitter, which splits the signal into multiple drop-side fibers. It could be 16, 32, multiples of two, basically. And that smaller signal, that weaker signal, then goes to all of the homes, businesses, and mobile sites on the right side. And then on the upstream, the data streams are synchronized so that only one of those N units is speaking at the same time to go up in the upstream direction. What I'm talking about there is if you're using the same wavelength TDM pond. So single wavelength with a splitter and multiple terminations on the right side is point-to-multipoint. Now, on that kind of architecture, you can have different technologies. I've shown them on the bottom. GPON was the first at 2.5 gig with one wavelength and then evolving to 10 gig XGS and into the future 25, 50 gig PONs and even higher perhaps in the long term on a single wavelength. All of these single wavelengths actually can coexist because they're different wavelengths on the same PON so you could in theory have all of these things coexisting together at the same time. The recent innovations in the NGPON2 were to do what we call TWDM PON, which is to stack multiple 10-gig TDM PONs on different wavelengths and allow the ONTs at the customer end to tune to the correct wavelength and get that 10-gig signal. Now, this could be extended to higher bit rates as well. And the third kind of family of PON technologies would be WDM PON, where you have one wavelength for each end unit for each. and this would typically be used for business or for mobile. So one cell gets one wavelength, and all the wavelengths are stacked on top of each other on the same pond tree. On the right side, we have point-to-point fiber, and this is used in some cases for residential, but not dominantly, but is used more for mobile macrocell backhaul today. So you see an evolution of the bit rates over time. I don't want to take too much time, but for future reading, we could delve down on this slide a little bit more, where it talks about the GPON, the rates, XGS, the TWDM, which was adding multiple wavelengths of 10 gig, and then into the future, where we're looking at 25, 50 gig, and point-to-point WDM PON. Reza has kind of introduced this notion of separation of functionalities. in the mobile side. So on the top, what you saw in the animation was all the functionality starting on the left side. That's where they were in previous generations, in LTE, for example. And as 5G came along, there was this notion of taking some of those, you know, complexity in the radio sites and the radio units and centralizing it and creating an interface between these at these different locations between different groupings of functionalities. So bringing them back towards the right, which is the cloud or the central location. So we've created in 5G the simple RU, a DU, which contains some of the functionalities, and the interface between those is what we call a front-haul or eCIPRI, as Reza had mentioned. That has extremely high bandwidth and very tight latency requirements in the order of 200 microseconds. Between the DU and the CU is what has been called mid-haul or F1 interface. And the bandwidth and the latency requirements for that are quite similar to the historic bandwidth backhaul transport on the right side. So backhaul and F1 mid-haul have similar characteristics. And you can see the technologies that might be used to support those on the bottom. From a bandwidth point of view, you need less bandwidth for the interface on the right, the lower speed ponds. In the middle, it's a little bit more. And on the left, it's much more bandwidth. And we're going to also have to deal with this latency issue. And we'll come back to that question later on. So... We took a look with a Tier 1 operator at how they might leverage their fiber-to-the-home network for a large event to provide mobile 5G capabilities inside the field, inside the arena, and also in the surrounding areas using a common technology. What is shown here is actually the PON technology is is dedicated for the mobile, but it's using the very same PON technology that is sitting beside to serve residences. And we're using here XGS, which is only 10 gig per second, and we're looking at dimensioning for F1 mid-hall interface. So there's two categories. There's the sub-6 gigahertz, which was used to serve outside of the stadium, and we used millimeter wave to serve inside of the stadium with shorter reaches and obviously different spectrums and different levels of MIMOs leading to different bandwidth requirements at a peak level and on an average use basis. And based on that, we were able to show that, okay, you could support nine remote units, radio units, or seven radio units just simply with a PON network at a 10 gig. If you were able to implement some statistical multiplexing, which you can because this data transport is statistical in nature, and we do it all the time in the residential access, the broadband access, we can achieve many more cells, so 55 or 42. But even 9 or 7 is a reasonable number. Anything above 4 becomes... something that's usable or interesting for PON. So you can get the cost savings of reduced fiber and the lower costs of PON technologies. So on this note, I'll pass it back to Reza to give us some more insight.

All right. Thanks, Ron. So now a few more slides on the examples for how to use the PON technology in the XR networks. It's just some examples that could be many more combinations and variations of use cases. But when we look at the use cases, you know, there are two major categories. One is are we going to have a separate PON network for the mobility application for wireless, or is that going to be an overlay that combines? mobility with some legacy like residential business services. That's one major decision that has to be made. The other one is are we going to use the pond for the back hall, front hall, mid hall, or maybe a combination of that, maybe a combination of front hall and back hall, whatever. Of course, the choice depends on the geography, fiber ownership models, the number of radio units, all of that. But these are some of the different options that can be looked at when designing or when considering PON networks. Now, as Ron pointed out, there are different variations of PON technologies. I'm going to just pick two, TDM, which was known as temporal in Ron's PowerPoints, So TDM uses obviously the time division multiplication technology to multiplex multiple streams coming from the radio to the distributed unit. So the problem we have is as the number of these radios increase, we can have excessive latency on the upstream. And we already saw what are the latency requirements on the front haul. So that can become a problem. How can we solve it? There are different ways of looking into that. If we are, for example, using PON for a conversion network between mobility and residential services, then one way of solving this is to prioritize the rate of traffic higher than the residential services, for example. Higher priority gets more attention, more bandwidth, and lower latency. of solving that problem is to use something that's known as dynamic bandwidth allocation where we are looking at the upstream traffic that the buffer status and determining based on that what should be the allocation of the time slots for the upstream traffic. I think the most attractive one would be if now we can take advantage of the knowledge that the DUs have or the BBUs have and allocate the bandwidth based on that. is also known as cooperative dynamic bandwidth allocation, in which case the DU and OLTs talk together, and basically the DU tells the OLT, this is what I need for this stream of traffic coming from whatever UE, and the OLT allocates the respective bandwidth for that type of traffic. Now the other one that Ron was talking about was WDM or TWDM where we are using multiple wavelengths and different things to look at, really lots of different use cases and design alternatives that we can have. But just to look at some of the very basic things we should look at when we design these networks, obviously the very first thing is the optical path loss. is enough optical power available when we consider all the connections, the connectors, the max, demax devices that are between certainly the fiber lens. Latency obviously, as we said over and over again today, very important. For that, look at not only the fiber transmission, the speed of light over that fiber, but also at any WDM processing elements that are on there, what kind of latency they burn on the way. Theoretically, we can have basically even different wavelengths used for different types of traffic, but it is a theoretical one. I haven't seen it, if we want to consider that. And finally, bidirectional optics is also very important as operators look to simplify the design of these networks. The advantage that bidirectional brings is now that you have one fiber used for uplink and downlink, you can really make it much simpler to manage the networks. When you have dozens of radios out there and you have all these fibers coming down, bidirectional optics can really help us manage the wavelengths and the fiber more efficiently. In summary, really with the combination of PON technologies, we have much, better technologies now that are coming and making them an attractive choice for ex-hull networks.

Okay, so we're going to talk a little bit more about what's coming down the pipes, what's coming in the future. One area of development is in the area of future pond technologies. We already touched on existing and emerging and future pond technologies. I think what we can say here is that higher bit rates will be required, especially for front haul, and in some cases for very high-spectrum large band uses with the F1 mid-haul. An awful lot can actually be done with existing XGS 10-gig and NG-PON2 TWDM technologies. So we can address a very large segment of the scenarios with those existing 10 gig next generation pond technologies. But as we go further, as mentioned, especially for the front haul, we'd be looking at 25 or 50 gig TDM ponds. And we'd be looking at possibly WDM ponds with one wavelength per user. And even some work on higher speed TWDM ponds. Activities are happening both at IEEE and at ITU for these next-generation ponds. In ITU, we're calling it Higher Speed Pond, HSP, and also some WDM pond. So bandwidth is one thing, but it's not the only thing. As mentioned, latency is a significant issue that we need to take into account. Now, let's think about latency. Latency is required for the services, some services, not all services. Video, you know, watching broadband video, you know, latency is not an issue. Gaming, yeah, latency becomes an issue. But really, it's not that hard to meet those requirements. Where the real latency difficulty comes is for front haul, where the split is such that in functionalities, is such that you need instantaneously the data stream at the remote radio unit from the time it's sent. In fact, as I mentioned earlier, there's this 200 microsecond kind of target. It's not a hard target. It could be 250 or 200, depending on the radio, the mobile equipment. But this is the amount of time that you have to get the signal from the OLT side out to the radio unit. 200 microseconds is really only 20 kilometers at the speed of light just firing down to the other end. So we don't want our TDM pawn to be slowing things down from a latency point of view. And there are things that we're doing to make this as close to lightning speed as possible. And one of these is what Reza had mentioned, which is cooperative DBA. Previously, the DBA was provisioned on demand when there was a requirement from a user. The user says, I need more demand. It sends a message from the ONT back to the OLT. The OLT provides the bandwidth, and now it can use it. However, that time, we want to reduce that so that there's no latency, there's no delay in providing the more bandwidth. So the proposed approach with coordinated DBA is that the mobile network already knows that there are radios out in the field that are requesting for bandwidth. So let's not wait for that process to happen and then to make the request to the ONT. Let's have the mobile network proactively communicate with the OLT. So the OLT can put in place the bandwidth available for that special increase in bandwidth requirement. So this is one of the things that is being pursued. A demo has been made with this in 2017, and there's a trial with a customer as well. And the standards are in progress to address this. So this is kind of one of the elements that makes TDM better suited or improves its suitability for doing front-haul. A third area of innovation is this notion of slicing the network Software-defined networks have been something that we've been working on for several years now in the Access Network, which allows us to do many things, including logically slicing portions, resources of the Access Network for different functions. So although it's one physical network, you can create individual slices for different purposes. Now, in the world of mobile, this is also being pursued, and so it can be exploited. In the world of mobile, you may want to have a dedicated slice for a particular application, a particular service, or a particular operator. And this needs to be provided on an end-to-end basis. So the access can do its slices, but it needs to be part of a bigger orchestration from an end-to-end point of view, creating slices for these different services. So this is work that is being done as a vendor as well as in standards to make these things possible. I'll pass it back to Mariana.

Okay, thank you, Ron. Thank you, gentlemen. So before we wrap up, Nokia and VIAVI will take a moment and give us a very quick overview of their specific approaches in the context of 5G transport. starting again with Ron.

Okay, so what I want to do is zoom out to a much higher level. I alluded to the need for end-to-end solutions. And the access, the broadband access is an important element of this end-to-end solution. And we've bundled it together under something that we call AnyHall. And it includes transport radio and other transport technologies, but broadband access is going to be extremely important as these cells go deeper into the network. But there is so much more. There is the slicing that we just talked about and integrating that and orchestrating it from an end-to-end point of view. There is creating these connections. There's the automation. There's the service monitoring from an end-to-end point of view. And there's rapid delivery of of the service. With the right mix of transport technologies selected and the end-to-end network automation, we can enable operational excellence. Broadband Any Hall certainly has a valuable place in the 5G ecosystem. With the acceleration of FTTH and 5G, we see more opportunities for synergies between fixed and mobile. So Nokia is investing heavily in new features that will enable broadband AnyHall to follow the 5G evolution and to be part of a complete end-to-end service delivery package. So that's our view from Nokia.

Thanks, Ron. And I'll just do it quick from the VR perspective. Obviously, the verification of the SLAs that we talked about today is very important. It all starts with the optical layer, checking the connectors, the fiber, the wavelengths, all very essential to get the fiber technology right. And then, above that, the transport and synchronization, verifying the SLAs for time error, for the packet loss, bandwidth, all of that. And finally, all of that will only be good if we can verify the performance over the radio interface. verifying the spectrum, the beamforming functionality, and managing all the data in a well-suited network that can manage our data for good workforce efficiency.

Okay, so with that, and before we go on to Q&A, I will do a quick wrap-up of what we discussed today with a few key takeaways. So, we talked about 5G applications being the driving force for the strict latency and bandwidth requirements that really kind of shape the mobile transport network, but as well as, of course, the radio component. So, indeed, there's a fundamental shift in mobile network architecture that began in later releases of the 4G specifications and evolved to what we call today the actual network. We then, of course, introduced broadband networks as a viable transport solution to meet the stringent 5G requirements. We delved into the cost efficiencies that can be achieved with point-to-multipoint architecture and how, indeed, ponds can be planned and designed for different use cases, where, for instance, you have the ODN dedicated to 5G or share it with 5G. And then, especially when it comes to the latency-sensitive front-haul networks, concepts like service differentiation and cooperative dynamic bandwidth allocation need to be considered when leveraging TDM PONs. TWDM PONs can also be used to serve the 5G transport, but in every case, the design of the network will require careful consideration to really meet the latency and bandwidth constraints and then moving forward of course innovations on technology architecture and operational model are required to address the 5G transport challenge at scale We have the concept of network slicing, opening up opportunities for new operational models by means of what I like to call kind of a cross-layer intelligence that brings all their resources together. But also, in the near future, pawns with higher than 10G data rates will emerge to really enable scaling of the transport network with traffic growth. So with that last comment, I would like to go on to Q&A. I think we have about 10 minutes and questions came in indeed. A good range from high-level to technical, although I have to say A lot of them had one common theme, which was the front-haul segment, which should come as no surprise. So let's see. Okay. So maybe that's one for you, Reza. Someone from the audience is really asking how the other technologies compare as a front-haul solution. I mean, what is the comparison with, say, WDM, TSN, and the other more kind of traditional optical solutions when compared with PONS?

Absolutely. I see that. The PON technology kind of between WDM dark fiber on one side and the TSM technologies on the other side. The advantage over WDM is now we are making better use of the fiber of the wavelength if you're using WDM PON. Because in WDM PON, obviously the wavelengths are locked to every radio. And here we are actually making a better use of those fiber and wavelengths. TSN on the other side, timing-sensitive networks, those technologies are taking advantage of statistical multiplexing on the packet layer and can make even better use of the technology and the fiber. However, timing-sensitive networks require really new technologies in terms of packet switching that are, can deal with the latency requirements and the jitter requirements that we have in frontal network. So PON seems to be a good sweet spot between the two technologies.

I might add to that, if I may, Mariana, that one of the things that the PON world is looking at is how we might implement some of the notions of the TSN in a broadband access world. So time-sensitive network, we're trying to carry time-sensitive traffic, and if we can do this in a in a synonymous, in a synchronous way, how would I say, in a similar way to the other TSN networks, then we can actually possibly call the broadband access a TSN-certified element in the overall end-to-end network.

Indeed. Okay, so I have another question about Actually, a few questions on data rates. What are your views, gentlemen, on what will be the dominant data rate for the 5G network, of course, in the context of ponds? Is it going to be 10G? Is it going to be 25G? What are your views on that? Ron, maybe you want to take that?

Yeah, maybe I'll jump in quickly on that. So as usual, it depends. But broadband, sorry, I meant backhaul and midhaul have significantly less bandwidth than fronthaul. So those ones are going to be carried on, in some cases, on GPON and also on 10-gig PONs, whether it's XGS or NGPON2. This, you know, when I say it depends, it depends on two key factors. One is the level of MIMO, and the other thing is the number, the amount of spectrum, the size of the spectrum, radio spectrum block that you have. So it almost scales linearly with those two things. However, the beauty of backhaul and fronthaul is it, sorry, backhaul and the F1, which we're calling midhaul, is that it's very statistical in nature. It kind of goes up and down with the usage of the user. and therefore we can do a lot of statistical aggregation, and the possibility is to feed several of these radio units even with 10 gig. There will be incentive to go to higher rates, and that's why standards are looking at 25 and 50 gig for mid-hull. But the real driver for the higher rates is front-hull radio. And there's a question here, is it 10 gig or 25? You can do a lot with 10 gig with a point-to-point or with a wavelength dedicated for each node. But if we want to share multiple radio units, I think we need to be looking at 25 or 50 gig for a shared front hall on a single data stream. And then there's even in Asia and China, there's interest in 25 gig serving a single radio unit. So, you know, future radios are going to be requiring for front haul 25 gig directly to serve them. So it's a bit of a moving target, but that's the lay of the land today as we see it.

And then what about the suitability of the TDM pond versus maybe a WDM pond?

Yeah, TDM, again, plays really well for backhaul and midhaul, and we're finding solutions for fronthaul with, you know, with higher data rates and with latency-reduced approaches. But there's, admittedly, a little more of an uphill battle for TDM. It is, you know, definitely going to play a role with operators who want to look for synergies. I think it's, you know, We can expect the WDM is going to play a role for front-haul as well in certain parts of the world that are exploring this. Korea has been pushing in that direction for a while, and China is kind of also exploring that.

So it will be a toolkit, I guess. What's your view on that?

Sorry, was it for me, Mariana?

Yeah, yeah, no, I was just thinking that you were trying to weigh in.

No, I just agree. I think, again, it depends, as Ron was saying, but, yeah, WDM mostly in front haul and then TDM backwards here in the mid-haul backhaul. And I agree, 10 gig is right now the mainstream for front haul, but the cost of 25 gig optics are coming down, so we are expecting more takeout rates for 25 gig. And then on the mid-haul backhaul, 40 gig, 100 gig, and 50 gig is also coming. It will take probably some time until the costs are under control. But I agree basically with what Ron was saying before.

Mm-hmm. Okay, and I think we might have time for one last question. I would like to ask, so basically what sort of adjustments do you think ponds require to accommodate the 5D transport? Can we just use off-the-shelf products or do we need to make any adjustments to the standards, for instance?

Ron, you want to go first?

Okay, just like needed, yeah, to support 5G. Again, is it backhaul, midhaul, or fronthaul? If it's fronthaul, latency is one thing that we need to work on, and that is being worked on. And then also, you know, the higher bandwidth. For backhaul and midhaul, a lot of work has been done to enable things like timing and synchronization, some additional functionalities to be added into the device. And one of the things we would like to get to is pluggable SFPs that have all those functionalities in them so you can plug them directly into the radio. Definitely GPON has those timing and synchronizations. And then 10 gig is very close to having that as well with NG-PON1 and NG-PON2 or XGS and NG-PON2. So those would be things we would want to have that are in some cases there off the shelf and in some cases needed to be added.

Mm-hmm.

Yeah, really, the front hall, we talked about it, right? The cooperative DBA, that's going on in ITU and other respective organizations. And also, as Ron was saying, on the back hall and mid hall, continuous improvements on the optics, et cetera, right? But yeah, definitely things are going on in standard bodies and in vendor communities to improve it and adjust it for the 5G world.

All right, thank you both. So with that, I think we are at the top of the hour. There are a few more questions, of course, that we did not get the chance to get to, but I'm sure that the Nokia and the AVI teams will be happy to follow up offline. So with that, I will just hand it back to Alan.

Thank you, Mariana. I also want to thank you for leading our discussion today, and a special thank you to Ron and Reza for sharing your expertise with us. I'd also like to thank everyone for participating on our webinar and for submitting all your questions and comments. An archived version of this webinar will be made available shortly, so feel free to come back, view this session again, or even pass it along to your colleagues. Now you will see a short survey pop up at the conclusion of the webinar. We'd love to get your feedback, so please take a few moments to fill that out. And lastly, make sure you continue to follow us on Twitter and LinkedIn for information on future IHS Market Technology Group webinars. So again, thank you for joining us, and have a great rest of your day.