The Future of Cellular Networks - Densification, optimisation, virtualisation

The Future Network spoke to Ray Butler, Vice President of Wireless Network Engineering at  CommScope about how he sees cellular networks evolving around the world, with a particular focus on the Asian market ahead of our Asia Meetup where Ray will be joining a panel discussion.

The Future Network: Please introduce yourself and your company.

I am responsible for wireless technical sales leadership in outdoor RF products at CommScope. The division was formed three years ago and is a customer facing, architecture orientated team with a global remit. We engage with the technical and engineering staff, but also the CTOs of MNOs, as well as executives at towercos and OEMs, so we really try to engage on the technical side with the decision makers.

At CommScope we are very engaged on the passive side of the telecoms infrastructure sector, with the tower companies, but we also engage on the active side, and we have supplied a number of active components over the years, so we are somewhat unique in the industry in our breadth of connectivity solutions and for being a part of the three major network trends: densification, optimisation, virtualisation.

The Future Network: In terms of densification, the MNOs are not the only ones who we need to consider - could you talk a little bit about your work with towercos and infracos to date?

We work with a number of towercos, and we see the model shifting. Years ago we would sell to and work directly with the mobile network operators (MNOs) and with the tier one Original Equipment Manufacturers (OEMs) and we still continue to do so, but in the last couple of years, with the migration to small cells, we see the towercos playing a much more significant role in the decision making process, some are even doing the RF designs for the MNOs, and overall their place in the ecosystem has been elevated.

The Future Network: In which of the geographies around the world are you seeing the most interest in the three key themes you identified earlier of densification, optimisation, virtualisation?

On the path to 5G, I fully expect that Asia and North America will lead the way. In North America, I see a lot of interest and significant progress in the last 12-18 months in the number of small cells deployed and the maturity of the products that are being requested. In Asia, in terms of deployments, the numbers are even larger. The population densities and the willingness there to densify the networks is world leading, especially with the Olympics coming to South Korea and Tokyo there is a real focus on having 5G ready. Consumer expectations in the Asian market are higher too, and there are much higher throughputs than in other areas.

The Future Network: Is there anything you can share about your experiences specifically in the Asian market? 

Our Asian remit includes Australia and we see some real innovators in that region, we also see the need driving innovation and customer expectations. I was in Singapore recently and when you do a quick speed test on your phone you’re getting in excess of 100 megabits per second, typically, which is fast when you compare it to other areas where it’s more like 20-40 megabits per second, or even less. That really jumped out at me, that the consumer expectation is those types of throughput rates. Some of the countries in Asia have densified using small cells, and some have densified using ‘six sector’ approach, or even higher than six - we call it HOS - Higher Orders of Sectorisation. This basically means that if you’ve got six, or seven, or eight sectors on a site it means that you’ve got that many E-node B’s on the site, and you’re pumping that much more data into the covered area, so it’s really about the density, it’s about the gigabits per second per square kilometer. This densification overlaps heavily with optimisation because in order to get the network up and running effectively it has to be finely tuned and heavily optimised.

The Future Network: In the Asian market, do you have any observations on the technology of choice for network densification? Is it mostly small cells, or are you seeing DAS too?

We see both, across both Asia and North America, but what is interesting is that in North America, at least half the cases of ‘small cells’ are actually full powered Remote Radio Units (RRUs) just mounted close to the street at maybe 15 metres rather than 100 metres, or what we might call small cells at eight - ten metres and the high power RRUs at fifteen metres, and then the five watt, lower powered ones at the eight - ten metre range. Part of this is that the cost structure of deploying small cells is still at a point whereby the operators are going to want to put full capability on the site when they do deploy it. We see good success with the outdoor DAS because of its multi-operator capability, and clearly this is the focus of the towercos - they want to leverage the location, the backhaul, the power - essentially to make the most of everything they brought to the site, this is best for the community and in the long run, best for the operators too.

The Future Network: We have noticed a shift from carrier funded, to enterprise funded network densification recently. Have you seen a shift in the business model and how these network densification solutions are being deployed and paid for?

In-building, clearly venue and enterprise funded systems are the trend and the path to the future, as long as MNOs can align with them. In North America we often think of small cell networks as ‘the 5G network’, and its just being used for 4G until 5G is ready. So I think part of the business case is the path to 5G, and getting ready and beginning to densify in preparation for 5G and that pole that has 4G / LTE electronics on it today will have mmWave and current band electronics on it in the future. It’s about getting PB&S (Power, Backhaul and Site acquisition) and the difficult (and expensive) part of small cell deployments comes down to PB&S: can I get power to the site, can I get backhaul, and can I get the site - for example a pole and all the necessary approvals. Once the electronics are there, the easy part is to remove or supplement 4G / LTE with 5G. Right now, the sites being deployed are for the large part in preparation for 5G, and I think where possible they will be leveraged for both fixed and mobile communications. Indeed, in many cases, in some of the more expensive areas, being able to leverage both fixed and mobile communications will be a key part of the business case.

The Future Network: We’ve seen a lot of activity on the convergence of key solution providers and deployers in recent months, with ATC and CCI acquiring fibre companies, and Corning’s acquisition of SpiderCloud, for example, how do you fit in with this trend?

We see the activity that is going on at the moment, and for your subscribers that aren’t aware, a couple of years ago CommScope acquired TE Connectivity. This discretely leverages us on the fibre side of the market; we have significant reach and depth in our solutions and on our customer side for these fibre based products. We definitely see fibre as a key part in the convergence of fixed and wireless networks, and we have a strong presence in this market. We also acquired a company called Airvana, and we now have the OneCell product, so we are very well positioned to be a significant player on the connectivity side of convergence.

The Future Network: Are you involved in the operation of any of these networks?

No, we provide equipment and solutions, on the DAS side, we provide services on the design and installation side, but we don’t operate the networks or operate as a neutral host. The one exception to this is in the CBRS 3.5 GHz unlicensed spectrum in the US. There is a dynamic spectrum management solution called SAS (Spectrum Allocation System) and we are a provider of dynamic spectrum management - this is currently being developed and we are testing with other SAS providers.

The Future Network: What do you see as the main demands on, and shortcomings of the existing infrastructure, in terms of being able to meet the demands of new technologies like augmented reality and the higher data throughputs they require?

It certainly is a very exciting time at the moment for our industry. We are involved in this shift from hardware to software based infrastructure and we will provide the fibre, the cabinetry, the connection, and all of the RF connectivity to enable virtualisation. The thing that comes along with virtualisation that becomes very interesting is mobile edge computing and the need with some of the use cases to drive down latency. So we expect to see  data centres at the edge where you have base band units, and then combined, so centralised first and then virtualised. We’re already seeing this happen, as small cells are deployed in North America. Generally those base band units are being centralised (to the small cells), so they’ll either go in an existing macro site structure, or in many cases they’ll go into cabinets which are centrally located at the edge, and this move to the edge will significantly affect the towercos, operators, and the OEMs. It will also affect the cost structure - with the 5G architecture that is being adopted, it looks like of the eight splits, split two and split seven are the splits of choice, and what that means is that where today you have a base band and a remote radio unit, with 5G you’ll have three units; a Common Unit (CU), a Distributed Unit (DU) and a Radio Unit (RU). We can see DUs located near the RUs, or we see CUs being much closer to the core and they are likely to be the first to be virtualised because they have fewer real time function. It’s exciting to see how the architecture is taking shape and translate that to look at how it will affect the network and the tower companies.

The Future Network: It sounds like an awful lot of capex needs to be deployed in order to achieve this...

Initially the capex deployed will be on servers and hardware accelerators which will help make the real time functions possible, and that’s basically what we’re seeing happen now. I think a lot of things that are in dedicated units today will be centralised and boxes become applications. We see the architecture of our multi-operator small cell solution, the OneCell as the true virtualised RF, so you can really shift capacity from one radio to another through the use of this and similar small cell architecture and I think we will see this adopted for 5G to help make it a truly flexible network with the ability to really manage capacity as needed.

The Future Network: How have you seen the operator uptake of your multi-operator solution?

We have gone through all the interoperability testing with a European operator, and we have deployed a pilot in a large UK stadium, and it performed exceptionally well. We are seeing traction and interest and we are getting early approvals; the architecture is so revolutionary it’s bound to gain traction with other operators - it just takes time to get these things approved.

The Future Network: What are your next steps in Asia? 

In the Asian market we are pushing both the small cell antennas and the Higher Orders of Sectorisation in terms of approaches to densification, and depending on the operator there is a preference for one of these approaches over the other. From a convergence of wireline and wireless, on the fibre side we are definitely seeing interest, especially in the more developed parts of Asia. In Australia we see definite convergence of CRAN and wave division multiplexing for small cells, so to me it’s the same story there as it is in North America and Europe.