We spoke with Renuka Bhalerao, Senior Product Line Manager at Radisys, for an update on TD-LTE, the move towards small cell virtualisation and a general market update. She tells us which LTE features are proving most popular, explains how SON is becoming mainstream and reveals an unexpected new market opportunity.
What's the current state of the TD-LTE market for Small Cells?
China remains the largest market for TD-LTE and China Mobile's nanocell specification is being satisfied by multiple competing vendors. Around 30-35% are dual mode LTE and also support either 2G GSM or 3G TD-SCDMA, so they can fallback to non-LTE and support non-LTE smartphones during the transition towards VoLTE. With FDD-LTE spectrum now becoming available in China as well, we are seeing vendors begin to expand support for that too.
There are a number of silicon platforms in the mix, with some of our customers selecting multiple SoCs (Systems on a Chip). They may do this purely for dual sourcing, to ensure continuity of supply or price point. Another primary reason to select multiple vendors is for support of 3G TD-CDMA.
Japan and India are the two other major TD-LTE markets today, and we are seeing quite a lot of activity in both countries. Japan has both outdoor/urban and residential product requirements for SoftBank and others. We are seeing a trend towards some form of virtualised RAN in a future version of eNodeBs (LTE base stations). Long term requirements are for very high capacities – using Carrier Aggregation across 3x 20MHz which could achieve up to 1Gbps.
Other markets also have TD-LTE spectrum available, but they have been slowing down the timescale for deployment to focus efforts elsewhere in the short term. Europe has plenty of TD-LTE spectrum, but has been deploying FDD-LTE first with a view to adding TD-LTE later, perhaps for LTE broadcast or other innovative intentions.
What is Radisys doing on RAN virtualisation?
We are excited to see the Small Cell Forum and ETSI aligning on NFV and Radisys is actively participating in standards development efforts. We have been discussing many of the various options, seeing many different proposals from different vendors, and we can envisage how these would be brought into the network. Radisys is gearing towards a software-focussed network.
We've been partnering within the ecosystem to bring in a completely virtualised base station solution - including the physical layer (PHY) – not just a partial virtualisation of the RAN. We believe this will be a very disruptive technology in the long term. There is an ongoing effort to demonstrate how easy it could be for small capacity additions to base stations, in both small cell and macro environments. These capacity improvements can be added very energy efficiently and with no latency limitations.
However, we recognise that not every mobile operator will want to rip and replace their existing network, although the standard 4.5G architecture could be a complete cloud-based implementation. So you will see both extremes – the PHY can be externally located in a datacentre, using a general purpose processor (x86 or similar) and not at the radio head. That would require high performance fronthaul connections between the datacentre and radio head, but we believe we can overcome latency restrictions.
Others may prefer to leave much of the processing at the radio node, as in a traditional small cell architecture. There are multiple variations of how to do virtualisation in between these two extremes. We recognise there is hesitation to move these RAN functions into the cloud and intend to support multiple architecture alternatives through flexible functional partitioning.
Which of the LTE features are most in demand?
The basics of LTE are now proven and deployed. LTE's initial features were centred around data throughput, hand-offs, SON-based self-configurations, VoLTE and CSFB.
LTE-Advanced is really a bag of features and not every feature will be fully deployed. Operators will pick and choose features based on their practical advantages and relevance in local markets. However, we see Carrier Aggregation very much at the front of the features that operators will deploy first. Once VoLTE (Voice over LTE) is more widely deployed, it will enable plenty of spectrum refarming. In turn, this enables Carrier Aggregation across multiple bands to achieve much higher peak data speeds.
eICIC (enhanced Inter-Cell Interference Co-ordination) is perhaps the next most important feature because it deals with the interference from denser networks. CoMP (Co-ordinated Multi-Path) is also catching up. Until last year many saw it as relevant but complicated. Now by using Cloud technology and virtualisation, CoMP is seen as more feasible and gaining popularity.
I also see strong interest in combining licenced and unlicensed spectrum - both LAA and LTE-U are key features. Several good enhancements in SON (Self Organising Networks) are also in demand.
Could you expand on how SON is evolving?
It's clear that SON will be a hybrid solution, combining both distributed (D-SON) and centralised (C-SON) elements. The D-SON will be part of the base stations (eNodeBs) and gives real-time control over radio performance. It's an important part of the small cell vendor's products. C-SON brings complementary benefits, and good interworking between both is very useful.
To reinforce our commitment to that approach, we've partnered with AirHop to complement our own D-SON capabilities. This provides benefits at three levels.
First is self-configuration, which accelerates the installation and reduces commissioning times during deployment. Then interference mitigation ensures that small cells don't adversely impact the existing macrocell capacity and coverage footprint. Thereafter, system optimisation, using a variety of SON techniques, really brings full value to what the network can achieve.
Do you see any other market segments opening up to small cells?
DAS (Distributed Antenna Systems) equipment is often supplied through high power base stations with much of the RF energy discarded and wasted – they would much rather be using high capacity/low RF power small cells. This is very costly, not only for initial CAPEX but ongoing OPEX because of the high power consumption and excess heat generated which needs air-conditioning.
To address that, we do see a trend where DAS vendors are developing or resourcing small cells to use with their own systems. They have already started to seriously look at it and put into deployment. We may not be seeing this everywhere yet, but it is certainly being looked at and is coming.
And finally, will Voice over Wi-Fi wipe out the need for small cells?
We think that cellular base stations are here to stay. We understand what Wi-Fi offers and believe that both Wi-Fi and LTE can co-exist. There will be some use cases where Wi-Fi makes the most sense and other cases where the QoS demands the use of LTE. We think VoWiFi will make its way into the network as a low cost opportunity for voice, but that doesn't mean small cells will cease to exist.
There is definitely a more methodical approach when deploying VoLTE end to end because it provides:
- Better coverage
- CSFB and SRVCC for provision of seamless handoff to 3G and other networks – voice calls can be handed off to available operator controlled network (vs less reliable Wi-Fi)
- End-to-end QoS handling (includes special QCIs, scheduling methods on the RAN side, special QoS for plain voice, video or HD video).
All of the above will make a noticeable difference for the end user.