Mike Cronin, CEO of Node-H, has been involved with small cells since their inception, actively participating in the Small Cell Forum and visible at many conferences. His company supply the complete software stack for residential and enterprise small cells, both 3G and LTE, now embedded in over a million units worldwide. He comments on market direction, differentiates indoor from outdoor and covers topics including LAA, VoWiFi, TD-LTE, SON and Virtualisation.
How do you think operators view Small Cells today?
I heard a common view expressed by several operators speaking at the recent Small Cells Asia event. Broadly:
- Everybody now accepts that Small Cells work. They may not be perfect but are more than adequate, and provide a much better and more consistent service than Wi-Fi
- Over 80% of traffic originates indoors but buildings are becoming more RF isolated from outside, especially for frequencies above 2GHz. This makes outside-in coverage more difficult and customers are becoming dissatisfied about quality of service
- Interference between in-building systems and outdoor macrocells is becoming less relevant
What is needed is to saturate indoor coverage without insisting on the very highest specification. We don’t need the most sophisticated SON LTE features to save a few low cost radio nodes, and don’t need full macrocell performance. It may make better economic sense to install ten standard small cells compared to five or six high specification picocells.
Urban small cells are more complex to integrate than in-building ones, requiring closer interaction with the macrocells. This is easier with LTE than for 3G, but does involve more technical work than for in-building.
Will LTE small cells be adopted more quickly than 3G?
It’s easier for operators to adopt LTE than for 3G because the complexity of the Small Cell gateway is considerably lower. 3G gateways have a lot of functionality and gateway vendors have control over which small cell vendors they support. This has meant that operators would first have to evaluate and select a 3G gateway vendor, then choose from the limited set of small cells it supports – possibly only from the gateway vendor themselves.
LTE small cells use a simpler gateway architecture and can be bypassed completely for small scale trials. This allows operators to evaluate small cell vendors independently and in parallel to the gateway, accelerating the evaluation and procurement process.
We’ve just finished an ETSI plugfest focussed on LTE, working remotely with a number of different gateway suppliers and core networks. It’s all gone very well, providing some good experience for participants and demonstrating an enthusiasm to ensure successful interoperability across the board.
Which new features are in demand?
Within an Enterprise, Carrier Aggregation is a natural evolution. People always want the latest and best capabilities. Practically speaking, it’s easiest to achieve when combining a couple of 20MHz FDD bands. This is quite feasible to incorporate on a dual-band LTE small cell and is already available in product today.
Mixing modes (TD and FDD) is more involved. That, along with putting together a hotch-potch of bandwidths will lead to some quite complex scenarios both for small cell and devices, and testing will be a challenge.
What’s your view on LAA (LTE in unlicensed spectrum)?
I think this is definitely the way forward and a natural thing to do. It’s not a huge change and we already have smartphones starting to support it – I’d expect the majority of LTE handsets would support it within five years.
It’s more relevant for Enterprise and in-building use where many users are sharing the same small cell. In a residential environment, it’s likely that only one or two users are ever concurrently active on the same small cell and would already get a fast, low latency service compared to Wi-Fi.
LAA can be deployed quickly in the US and some other countries, but Europe will require a Listen Before Talk feature to be incorporated through standards first. I’ve heard some Wi-Fi people being quite concerned they will be dominated by LTE but don’t believe that would be the case. LTE would give better performance and is particularly suited to higher traffic, more populated public environments.
Will Voice over Wi-Fi (VoWi-Fi) wipe out the market for residential small cells?
There will be some VoWiFi provided by operators looking to save the costs of small cells, but there are just so many little “gotchas” to be wary of. Firstly, VoLTE must be in place, something that has taken the largest operators such as AT&T and Verizon years longer than expected and they still don’t have end-to-end VoLTE calling between them. Not all networks will deploy VoLTE, many will rely on fallback to 3G instead.
Another issue with Wi-Fi is that the fundamental access mechanism is so rudimentary that services can easily be interrupted by other users, especially in the home where there may be high-rate streaming or casting.
This is why I think we may still see some new residential femtocells rolled out because it is a secure, mature and predictable alternative. There must be a strong business driver behind it, such as where spectrum is very limited. Wireline broadband providers with a small allocation of spectrum, such as Free France or TalkTalk UK, are taking advantage of this business model to disrupt the market.
How will independent small cells work with SON (Self Optimised Networks)
We consider it our role to provide a complete solution, not just 90%, and so this naturally includes a SON capability. This is a distributed SON (D-SON) that is more than sufficient for in-building use. There is always a certain amount of Centralised SON (C-SON) which defines the limits within which we can operate through a set of parameters. We also use the X.2 interface to share information between small cells, so most aspects of the real-time radio environment are understood by all.
Fundamentally, each operator has a different view of SON and how the small cells should interact with the macrocell layer. Some may want to minimise impact on the macrocells; others may want to maximise capacity offload. Parameters can be set to achieve these different strategies.
In the future, we could see a more dynamic interaction with Centralised SON. We’ve had discussions but not seen a strong pull for it yet. Operators want to know that it could be done in the future but it isn’t seen as necessary at this stage.
And finally, what are your views on virtualisation?
The Small Cell Forum has been working hard to define a set of architecture options, with different breakpoints that split functionality between the small cells and a central controller.
One might think the easiest technical option would be to use the FAPI interface but we consider this too brittle and unreliable. Missing a single frame tick in the LTE world causes serious disruption, so you’d need dedicated Cat7 cabling which is incompatible with current wiring and doubles the cost.
Instead we’ve demonstrated a solution based on the PDCP split which can work with existing Cat5 cabling and tolerate up to 30ms latency. We demonstrated this at Small Cells America using a Broadcom radio head and Intel for central processing. The advantage is that instead of running X.2 between small cells, it can share this information directly on the same centralised controller and optimise the use of shared spectrum more efficiently.
Where a building only needs a few small cells for an office, say two or three, I think it would be hard to justify the added cost of a central controller. Larger buildings with eight or more become viable, depending on the overall solution pricing vs cost of installation.
Node-H is a sponsor of ThinkSmallCell