A smorgasboard of the latest mobile technology. This week's LTE World Summit provided a wide range of content which aimed to satisfy almost everybody's appetite. While I focussed on the LTE small cell aspects, it's clear that the purpose of the show is changing. The big RAN vendors were noticeably less conspicuous, with Ericsson and NSN not having stands while Huawei and Alcatel-Lucent had fairly modest ones. The wider eco-system of small/medium tool vendors, system integrators and associated businesses were more visible. I don't believe there were any standalone small cell vendors exhibiting, although SpiderCloud were quite active.
With five parallel tracks, several co-located summits (in effect, additional topic tracks), two sets of presentations in the exhibition hall, the choice became potentially bewildering. The fragmentation provided narrower focus for some of these streams, at the risk of competing for attention. Despite a total of somewhere between 1,500 and 2,000 attendees, some sessions were really quite sparsely attended.
There's absolutely no doubt that the industry has fully adopted LTE and is gearing up to do more. Sometimes the only factor holding back a launch is the regulator making spectrum available. KPN explained that they had already prepared their network for this by installing Software Downloadable Radios (SDR) in all their new SingleRAN macrocells, enabling them to launch commercial LTE service just 6 weeks after the spectrum auctions completed.
A status report of LTE deployment worldwide
Alain Maloberti, VP Orange Labs, provided a very clear status report of LTE deployments across Europe. Broadly, it's delivering on its promise of faster speeds and extra capacity, with a wide range of smartphone devices available from the outset. The performance of LTE at 2.6GHz beats that of 3G even at 2.1GHz, but it will take until mid 2015 before the network capacity fills up enough to truly make it cheaper per bit than 3G. Voice is delivered using CSFB (switching back to 3G) and works well, although devices often don't return to LTE at the end of the call. VoLTE can't be rolled out until there is 98% coverage, when it will provide a platform for a wider range of advanced services (Skype style video calls, video conferencing etc.)
Some outstanding issues include:
- Voice calls need a fast return back to LTE after a CSFB call and to introduce SRVCC for VoLTE (SRVCC uses 3G data for voice calls when outside 4G coverage). Today, after making a 3G voice call, the device can remain on 3G rather than quickly returning to the fast data speeds of LTE. Some device vendors have made non-standard fixes for this, and a standardised solution is currently being tested in the labs.
- The 800MHz spectrum released by the switch to digital TV still has a small problem with interference to TV receivers – surprisingly, even the latest TV's are still being manufactured without being compatible/immune from LTE in the 800MHz band. An EU regulation is being developed to force TV equipment manufacturers to fix this.
- Roaming internationally on LTE isn't technically possible yet because there isn't a smartphone that supports all the major frequency bands worldwide. A device is needed that has 800/1800/2600 (Europe), 700/1900 (North America) and 2100 (Asia).
- Self-Organising Networks (SON) needs to evolve further with features to optimise network quality.
Korea continues to be the country to watch for the latest in LTE technology. Both KT and SK Telecom are striving to keep ahead of each other.
- SK Telecom announced the world's first LTE-Advanced network, now live with Carrier Aggregation combining both 850 and 1800MHz bands to achieve 150Mbps speeds.
- KT (Korea Telecom) will have 18K small cells deployed and will launch a residential LTE femtocell and also 150Mbps carrier aggregation during Q3 2013. Their longer term WARP+ program is targeting a 500x capacity increase through small cells, cell virtualisation and other advanced techniques.
The choice for voice on LTE
Operators can choose between CSFB (Circuit Switched Fall Back), where voice calls are handled by the existing 3G network or VoLTE (Voice over LTE) where voice calls use the new IMS core network.
It was said that CSFB is pretty much essential, because it will be needed by any incoming roamers/visitors and to handle any areas where LTE doesn't reach. One problem can be where LTE service is better than 3G, so that there are places where CSFB would fail due to lack of 3G coverage.
VoLTE still has some deployment issues, but is live in Korea (4.5 million users) and the US (MetroPCS). We can expect more launches very soon.
In practice both will be needed and we can expect most LTE networks to install both. Some may choose to deploy only CSFB (at least for the short/medium term).
Intel promoted a new smartphone LTE transceiver chip to compete with Qualcomm who dominate this market. Unusually, it ONLY does LTE (not any 2G or 3G), but does include all the major bands worldwide. For those who only want to use VoLTE (or just LTE data on their tablets), it could be a good solution. Even without much LTE roaming in place, it could allow local SIM cards to be used, and provides simplification of the supply chain for device manufacturers selling worldwide.
Radisys talked about a lot of action in the TD-LTE market. The main markets are China, India, Japan (Sprint) and US (Sprint/Clearwire). They've been working with their chip vendor partners towards a commercial small cell release by the end Q3 2013.
This should allow field trials to start during the second half of 2013, in the 2.3 and 2.6GHz bands, mainly for indoor use. Small cells would initially be 16 active concurrent user, migrating up to 32 and 64 subsequently.
A plea for indoor coverage
Before we get too carried away with ever higher speeds and performance, one Enterprise CIO reminded the audience that there are many corporate locations that don't have indoor coverage. This is getting worse due to modern construction methods, with energy efficient materials making buildings more like Faraday cages. Many of the sites he manages don't have full coverage in all areas of the buildings. I asked if he wanted coverage rather than capacity, which he clarified as both. It's not enough to simply bring basic coverage on its own - people want to make full use of their smartphones, so that means a sensible amount of capacity too. Repeaters are no longer an adequate solution for high speed data capacity and instead small cells would be needed.
Spidercloud promoted their E-RAN solution to address exactly this problem, highlighting their integral SON (Self Organisation Network) which reduced the deployment time and skillsets required.
Network Planning, Performance Management and SON
It seemed that quite a few of the exhibitors were offering software tools to plan, manage, optimise and/or automate radio networks. These included radio planning tool vendors, such as Infovista/Mentum, who outline an ordered strategy to work out where and when is best to deploy small cells. Not surprisingly, they point out that you really need to have your 3G/LTE macrocell network in tip-top condition first.
Performance management tools on display included Posix data warehouse, which archives every last protocol message in the network over the last 2 weeks, allowing diagnostic drilldown for any reported complaint to identify the actual cause. They can also present this information in various reports. Arieso (now part of JDSU) revealed some interesting statistics from their customer base - it seems the real traffic "pinch points" remain with relatively few customers consuming lots at relatively few places - video being by far the worst. Newcomer Tulinx demonstrated their easy-to-use GUI that didn't have a dropdown menu or a manual – they say it's as easy to use as your smartphone. Actix, who claim to be the largest vendor in this space by some margin (over 50 networks use their ActixOne platform), say the real value of network performance data comes with combining it with subscriber and revenue data. Teoco, with their background in margin protection and assets acquired from Schema and TTI Telecom, say they can crunch the "big data" to prioritise what and where to deploy the most financially attractive small cells.
One operator had setup a separate small cell optimisation team which is now disbanded. The whole network has to be optimised so that macro/micro/small cells are work harmoniously together, and that needs a team structure that's also aligned towards that goal.
Several operators have nationwide deployments of SON (Self-Organising Networks), but it's still fairly early days – various vendors spoke to me of a journey to be taken that involves both technical equipment, building trust that the system will work, and organisational/personnel issues as the role of radio planners/optimisers evolves. Nonetheless, there are some LTE SON systems already live, including that at SK Telecom which allow them to deploy large numbers of small cells daily, improve customer experience and network efficiency. Cisco's purchase of Intucell has raised the profile of SON over the last 6 months. Actix tell me they are already live with SON on a national network in Asia, refining their solution and battle-hardening it in the field, suitable for wide-scale commercial deployment next year. Neil Coleman, their Director of Global Marketing, sees SON as a journey rather than a product and it will take time for the industry to build trust in fully automated systems. One operator said "SON isn't enough on its own – you have to change your organisation to take full advantage of it".
Aricent have developed their own SON module for LTE small cells, but have also built a highly configurable rules engine suitable for what they term "Universal SON", which oversees multiple radio technologies on a central basis.
LTE requires a slightly different set of interworking to enable roaming when visiting foreign countries. This is still at an early stage. While delegates pondered the lack of a global smartphone that can operate on all the different LTE frequencies worldwide and how they could put new roaming agreements in place, most delegates turned off 3G roaming to save themselves from the high costs. I bought a local SIM card (1Gbyte for 20 Euros), while others used Wi-Fi where possible. Many delegates felt that LTE roaming was irrelevant with such high charges, something that the EU is planning to legislate on next year.
A few other points of note
- Public Safety session was very popular. Those custom engineered solutions specifically for fire/police etc. with handsets that cost $3000 each are seen to be too expensive. There's a strong movement to replace those with standard LTE technology (adding a couple of new features into the standard as required), and this could displace TETRA and similar proprietary systems. It's another driver to have LTE in the 450MHz band.
- The device ecosystem won't support eICIC for another 2 years. Apparently Qualcomm's current smartphone chip doesn't yet have it.
- An average smartphone owner uses it 150 times a day according to KPN
- The next 3GPP Release 12 (specs frozen mid 2014) has objectives for each of capacity increase, energy saving, cost efficiency, coping with different/divers traffic types, improving the user experience and backhaul enhancement.
- ERCOM don't just simulate networks – their tool plugs directly into macrocells/small cells in the lab and uses their real-time response to emulate a complete network cluster. This allows realistic testing of new software features, configurations etc. before deploying in the live network.
- Flash networks (not part of Adobe) showed some IP optimisation which works around some of the problems caused by TCP/IP. Their fast start and careful handling of lost packets can improve throughput and efficiency substantially.
- LTE is quietly a resounding technical success, delivering fantastic high speeds and reaching the boundaries of Shannon's law
- The macrocell industry accepts small cells are needed, but is taking its time to first deploy SingleRAN LTE macrocells
- SON will become an extremely important capability for networks over the next few years
- There are still plenty of (indoor) locations which aren't well served
- The industry is on a journey. There are no quick fixes, but those who prepare and plan ahead will out manoeuvre their competitors
A few minutes of video captured from the show including short interviews with Radisys on TD-LTE, Siklu on why so few small cell vendors there, Tulinx on their new performance management system and (for the technically detail minded) Posix show us a lengthy CSFB call trace and reveal why LTE smartphones are left stuck in 3G after voice calls and what is done about it.