What will the key themes be for a 4G conference today?
Last week, I was over at the 4G Wireless Evolution conference presenting on the planning and back-office systems needed to plan build and operate them. I also moderated an extremely popular panel session on femtocells (which had the best attendance of all sessions at the conference).
Data traffic is swamping the networks
That old chestnut is always top of the agenda. Ali Tabassi from Clearwire predicted enormous growth in wireless data consumption, claiming that the 400MB/month used by many iPhone subscribers today could rise to as much as 14GB/month. He believed that this would give advantage to providers with large bands of spectrum (Clearwire has 120MHz in most parts of the US), because this would be needed to deliver this capacity. What he didn't say was that this spectrum is at extremely high frequencies above 3GHz which doesn't travel so far - meaning that more cellsites will be needed to give good services, especially penetrating in-building.
But is the answer lots of spectrum or very small cells?
For my perspective it's not about bandwidth but instead about cell size. My own femtocell gives me up to 7Mbit/s (i.e. as much as my broadband DSL line) for my exclusive use. No wireless system using outdoor cellsites can directly compete with that. And the argument of offloading high data users from the outdoor sites to free up capacity for those who really need it also remains strong.
Therefore, I'd argue that the 4G technology debate is actually between WiFi and femtocells. With some 2G femtocell vendors suggesting that they can be used in conjunction with WiFi for data from smartphones, it still seems to me that these are the only commercially viable approaches to deliver these enormous data volumes at a sensible price and quality.
How is voice going to be delivered over LTE
The other lively debate around 4G at the moment concerns how voice is going to be delivered. There are broadly three main techniques for delivering voice over the new 4G LTE radio system:
Circuit Switched Fallback (CSFB): Send a paging message to the phone when a call arrives, so that it switches from 4G back to 2G/3G and uses the existing voice systems to take the call. Adds a delay when making/receiving calls, and there is also a risk of dropping the call if the handover to 2G/3G doesn't work. But this is 100% compatible with all existing voice and SMS services and requires little additional investment other than a software upgrade to the MSC switch.
IP Multimedia Subsystem (IMS): Use a completely new call control signaling protocol (SIP = Session Initiation Protocol) to setup and teardown calls. Interworking/conversion boxes can make this compatible with existing 2G/3G systems, so that the same phone numbers and services can be reused. Alternatively, newer handsets with native SIP/IMS capability could make use of the latest SIP features which include many found in Skype or Microsoft Communicator today (e.g. concurrent chat, file/picture transfer etc.) Requires substantial investment in the new IMS core network, interworking and back office provisioning/billing etc. Does not currently support text messaging (SMS).
Voice over LTE Generic Access (VoLGA): Bit of a mouthful this one. Reuses the existing protocols developed to carry mobile phone calls over WiFi known as UMA. Provides full compatibility with existing voice and SMS 2G/3G services. Requires a VoLGA gateway.
We saw both Kineto Wireless and Genband (both vendors of today's femtocell gateways) promote the VoLGA option, AcmePacket and IMS Forum promoting the IMS solution. CS Fallback was not actively promoted at that event.
It seemed to me the VoLGA Forum have a good story to tell for today's GSM/UMTS operators, but haven't yet got a good one for CDMA operators. But it's the CDMA operators like Verizon who will likely be launching LTE first because:
- CDMA operators have no other evolution path, with LTE being proclaimed as their next step
- They are annoyed at missing out on lucrative roaming revenues from foreigners who use GSM/UMTS handsets
- The range of available CDMA handsets is more limited/later than for the larger GSM/UMTS market
- No CDMA iPhone is ever likely to be made
They (CDMA operators) are committed to the IMS option because it's the only show in town for them. I'm not entirely sure what their equivalent fallback solution is when out of LTE coverage areas (it may use IMS running over the 3G EV-DO data service), so this will set a precedent for IMS rollout. (Please clarify this point by commenting below if you know the answer).
I reckon the VoLGA forum have about 6 months to persuade the community to take up their solution - the debate about the choice for voice over LTE is certainly one to watch.
Realistic timescale for LTE commercial service
The last theme is around the realistic timescale for LTE to become commercially available. With each new generation of wireless technology coming out about every 10 years, we go through the same hype cycle suggesting that each time the commercial launch will be quicker. One presenter said that history shows the typical timeframe from freezing of specifications to commercial live service (2 operators in more than 2 cities) is around 3 years. He felt that the LTE specs were "rushed" (presumably due to competitive threat seen from WiMAX) and still needed some further clarification. So his prediction was for significant commercial launch around 2012. He also thought that LTE would be an outright winner over WiMAX in the technology stakes, suggesting by 2020 that LTE would have more than 80% of the market share worldwide.
My guess is that the Japan and US will launch LTE first - both Verizon and ATT desperately need more bandwidth and have bought spectrum for this purpose. Europe broadly seems to be happy to continue investing more into 3.5G for the moment, using HSPA and HSPA+ technology with an incremental rather than disruptive growth path.
Femtocells offer 4G data capacity at an earlier opportunity
For femtocells, making the case as the preferred short term capacity solution (even over and above LTE/WiMAX) remains important to win the hearts and minds of network operators struggling to meet the data growth demands of their customers. For me, the battle is not between WiMAX and LTE, but instead betwen WiFi and Femtocells.
So it's entirely appropriate that both are included in 4G conferences as a complementary technology solution.