Pam MALLETTE: Welcome back to 15 Minutes with 5G – a bi-weekly discussion with the industries big brains about the latest developments in 5G and what they mean for consumers, businesses, and society at large.
Joining us today is Peter Linder, who is responsible for business development for Digital Services in North America.
Welcome back to 15 minutes with 5G Peter, it’s great to have you.
Peter LINDER: Thank you very much, Pam, a pleasure being back on the air.
MALLETTE: When we talk about 5G in this podcast, we often talk about things that are a bit far off in the future, but today we’re really going to talk about something that is near-term, talking about the core network, where operators can make changes now to improve their service, while also preparing the network for 5G. So, let’s start with the basics. What is the core network and what does it mean for the core to be ‘5G ready’?
LINDER: The core network – you can say that the mobile network – consists of two main parts: the radio network which is closest to the users, and the core network which is the back-end. The core network aggregates all the traffic from the different radio nodes and makes sure that all mobile traffic can communicate with the internet.
In the context of 5G, we typically talk about a core network that can be designed in two different ways. It can be designed as an evolution from what we have today, and a more dedicated core for 5G. The main difference between the two is that the evolution from what we have today, which we often refer to as EPC, evolved EPC or EPC plus, is something that can support 2G, 3G, and 4G today where you add capabilities to connect to 5G radios, and where the next generation core is something that’s optimized for 4G and 5G.
MALLETTE: Great. Can you talk about decisions operators need to make about upgrading the core networks, specifically what are their options to improve performance in 4G that might also help them get ready for 5G?
LINDER: I think there are a couple of different things that are happening – different steps that are taken for the core network. One thing we look at is how you can increase the peak rates for the maximum speed, that could be taken through the core and delivered to the users. The other thing is the distribution of the control and user plane. Today you have the control plane and the user plane in the core network located at the same site and physical location. With the separation of the control and the user plane, you can bring the user plane closer to the subscribers.
The primary benefit with that is that you will reduce the latency for the services – the round trip from the radio through the core network. These things can be introduced now for 4G, and they are important steps on the way to 5G. So, as we look at this, we typically talk about a number of different things that are evolving, the first one being how we connect to the radios where we have options: where you go straight from the core network via 4G radio to a next-generation 5G radio; or whether you go straight from the core network to the 5G radio, and connecting to the 4G radio through that, a number of different options are defined. The second point we often talk about is the typology. In current mobile networks, the core (the EPC) is located in central locations. As we move forward to 5G, we can talk about the distributed cloud or mobile edge sitting closer to the subscribers, hosting some of the functionalities. The third thing that we often talk about is how software is going to be deployed in these models. Are we going to deploy network functions on parts level? Are we going to break it down to even smaller component on micro-service level? Or we are going to aggregate it to a higher level on virtual network solutions or network slices, where network slices are more or less addressing the whole set of functionalities required to deliver a given service or use case.
MALLETTE: It sounds like there are at least three options to start to improve performance now, that will also help operators get ready for 5G.
MALLETTE: Let’s talk about strategy. Should operators move fast to make changes to the core network to get something to market quickly or should they spend maybe a little bit more time and take a longer-term view of the architecture? What are some of the tradeoffs?
LINDER: As you said the first one is time to market – do you want to get out early, do you want to learn from 5G and evolve in small steps? A strategy advocated pretty much when you think that 5G has a lot of potential but also has a lot of uncertainties, and the best way to understand certain aspects, is to go out and do active things with it rather than trying to think it through, what the future might look like.
The first tradeoff is clearly between time to market. Right now, a lot of the emphasis I will say is on the radio understanding – what the radio technology is and what can be done, and also looking at with these new capabilities, how can we use the network in different ways? Most of the discussion is around network slices and what is the optimal architecture to deliver the kind of performances that we are looking at in the future. But it’s not really a choice between either/or; it’s more a choice of which steps do you take on the journey. It will evolve as we learn what flies and what doesn’t fly, and when we know more than what can be thought out in advance.
MALLETTE: What are some of the choices the operators face when choosing that core strategy for 5G?
LINDER: Choices they are looking at is very much, what is the starting point, where we are right now and what can be done with that starting point? The second thing is when we build the 5G network, do we build it as an overlay? Do we build 5G as an evolution from the 4G network? The previous network generation has been built as an overlay all the time. When we had 2G, we built a 3G overlay. When we had 3G, we built a 4G overlay. But now when we are getting towards 5G, it’s perhaps more extensions of the current network.
I talked about the typology question: where will we place functionality in the future? With certain things moving up in the network, other things moving down and some changes in what we are having today. The third one is perhaps an interesting choice which is what we think about network slices – what are the type of network slices which we think would be interesting and deployed in the future?
MALLETTE: What are some of the anticipated ways the network could be sliced?
LINDER: I think today you can say there is a starting point is very much where we have one slice fits all. So, all the data traffic that is piping through mobile network will get prioritized in the same way. So, to some extent, you can describe it as one very large slice encompassing all. Or it’s actually no slicing done at all – another way you may want to look at it.
The first obvious choice would be to segment on basic traffic that are widely different. Think about four slices: one for enhanced mobile broadband, one for fixed wireless access, another one for massive machine type communication and the fourth one for critical machine type communication. If you just divide it in those four slices, you will have very heavy traffic load from fixed wireless in one slice. You will have the large volumes on enhanced mobile broadband in one slice. And you have to make sure that two new types of categories, massive machine type of communication, and critical machine type of communication, get their attributes that they require to be taken care of.
The next one you can consider if it gets too crowded in one of these four top-level categories, you start segmenting per industry instead. So that the national safety and public security are dealt with one slice, automotive with another slice and health care with the third slice. Because there you can argue that different types of industries have different specific needs: healthcare obviously being one about security, automotive one about low latency, etc. So there where we tailor it for each and every industry.
But even if you look at an industry, there might be an interest to look at the slices one level below that and going all the way down to use case. So, if you look at automotive, we have different types of traffic that are going to go to a car in the future. There is going to be traffic for entertainment in the car, and traffic here is not much road traffic, it’s like traffic in the air. There is entertainment to the car; we also have signals related to anti-car collision and autonomous driving. And we have signals related to data sent from the car to the manufacturer or to the dealer to see if they are prepared to provide support. And each of those ones are different in nature. So, you might want to go to take network slicing down to the level of use cases but where the same type of companies use the same use case, where different car manufacturers and car entertainers can share the slice. Well, the far last lowest level is perhaps when you make a slice for a provider and a type of use case and get it down to that.
I don’t think we’ve started to see any trends clearly yet exactly how far it will go, and it’s a big question if you are on the top level you have one slice, if at the bottom level, you have perhaps hundreds or thousands of slices, not a hundred thousand, but hundreds or thousands of slices. So most likely, you are going to see a gradual evolution here of what seems to work out in real life situations.
MALLETTE: It does seem that in 5G the anticipated use cases are going to have quite an effect on many of the decisions that we make, not only in terms of how to slice the network but also maybe how to upgrade the core and you know what strategy we use to do so?
LINDER: Perhaps the most interesting part in that one, in a lot of the use cases, we don’t know exactly what this network is going to be used for, and we don’t know the really interesting use cases, the amount of the meter. So it’s little bit like getting ready, and being flexible and adapt to this going forward. It’s little bit like guys who invented the PC. They didn’t know exactly all kinds of programs we are going to load on it down the road. Same thing for smartphones – no one can imagine what kind of apps there were. It’s a little bit about creating a flexible and powerful platform that is designed for the anticipation of a lot of things we don’t know yet about in the future.
MALLETTE: Well, we have been telling our listeners that 5G is coming ready or not, choose ready? It’s clear that getting the core network ready for 5G is really a great place to start. So, thanks Peter for spending time with us today and thank you all for listening.
If you’d like more information, please visit ericsson.com/5g-northamerica and if you like what you have heard, please subscribe to the podcast.