This year’s 4Gworld conference in Chicago emphasised the momentum for LTE, but also highlighted the need for small cells and that the mobile industry is generally seeking how best to deal with the rapidly growing levels of data traffic profitably.
Last year’s conference was more of a shoot-out between WiMAX and LTE technologies, with both having similar number of exhibitors and presence. This year, fewer WiMAX vendors were present and it’s clear that the industry (at least in North America) is backing LTE. Really it was only Clearwire championing the technology, and several questions were asked on how WiMAX networks might transition to LTE at a later date.
LTE is more urgent for CDMA operators than GSM because CDMA doesn’t have the ongoing roadmap of speed and capacity improvements available to their competitors. True, KDDI in Japan is (uniquely?) deploying multi-carrier CDMA in an attempt to offer higher headline data rates, but this doesn’t achieve significant additional capacity. AT&T and T-Mobile are both continuing to invest heavily in 3G in the short term, although AT&T will launch their LTE network in 2011.
Making more money out of data (and specifically the data pipe) is a key concern for most operators today. With data traffic now consuming 70% of network resources (that’s excluding offload to Wi-Fi) and growing fast, but providing only about 30% of revenues something has to change. Various vendors offered policy management solutions to shape and control the traffic – watch out for premium data tariff plans at higher rates.
The fragmented spectrum allocated for LTE around the world - different frequencies and different bandswidths – will give device manufacturers a real headache. Universal handsets/dongles that work on every network may be difficult to engineer, leading to higher costs and less likelihood of universal data roaming.
Where will LTE appear first
US and Japan (in volume): The CDMA network operators have a more urgent need for LTE than their GSM/UMTS competitors. Japanese network KDDI is keen to deploy quickly, with it’s primary competitor NTT DoCoMo also likely to be one of the first to market. US operators Verizon and MetroPCS are likely to launch before the end of 2010, whereas AT&T continue to invest in 3.5G and launch during 2011. T-Mobile, the other nationwide US operator, has no immediate plans for LTE and wasn’t visible at the conference.
For those networks without their own LTE plans, LightSquared are building a nationwide US network to be resold wholesale through existing providers.
The first services will be data only, with a few super-smartphones able to handle voice using existing 3G, using the so-called Circuit-Switched FallBack (CSFB) approach. This means heavier battery drain, longer time to answer/make calls, but is a quicker and easier route to working devices. It will be 2013 before we are likely to see many native voice LTE smartphones.
The data tsunami
A recurring theme throughout the conference is that data demand is growing rapidly. Whether you agree with Cisco’s 40x growth rate or Yankee’s 20x, these are still huge and transform the shape and purpose of mobile networks.
Ericsson (not a fan of residential femtocells) recognise that small cells are essential to deliver the capacity required – LTE might give 3 to 5 x capacity , not the up to 1000x capacity required. So they forecast up to 10x the number of basestations in the future – not large towers of course, instead many picocells. If terminology weren’t important, then others might call them metro-femto or enterprise femtocells. They quoted figures of a basestation costing around 10-15% of the total cost of a cellsite. This leads me to think that when cellsite costs of power, backhaul and site rental being major ongoing expenses, locating small cells on customer’s premises using their own power, internet connection seems an attractive approach.
Many (most) network operators are actively looking at how they can control and price data traffic on their networks. Charging more for premium service is not yet commonplace, but higher prices for faster speeds are already on the menu for LTE from some providers.
Several well-known femtocell personalities participated in the femtocell track sessions. Simon Saunders (Femto Forum Chairman) gave a good summary of the state of the industry. Sprint explained their femto strategy which comprised five key points:
- Choice: can use with any CDMA phone
- Ease of Use: Plug and Play, no difficult configuration settings to setup
- Security: End to end encrypted
- Support: 24x7 remote support
- QoS: Known service quality (within the limits of broadband internet)
Cisco, who have led the deployment of ATT’s 3G Microcell rollout, said that the skeptics have been proven wrong. Consumers are now buying the product and liking it. There are a variety of different business models, even from the same operator. They recognise that deployment can be difficult becauses it crosses divisional boundaries, for example which budget should pay for subsidising what is effectively “network equipment”.
Manish Singh from Continuous Computing highlighted some of the practical issues around building the MME (Mobile Management Entity), which could control many (tens of) thousands of LTE basestations or femtocells. Some architectures won’t scale up to handle these volumes, leading to an option to frontend the MME with a signalling gateway. Manish also suggested that using TD-LTE for residential femtocell with a common frequency might be quite an attractive way of offloading substantial traffic in a more controlled way than using Wi-Fi.
Sprint have recently launched their 3G Airave femtocell, using the Airvana femto and Tatara core network, based on the 3GPP2 standards. They confirmed to me that they will continue to support their 2G product and have no plans to turn it off, but that growth will be “pimarily on the 3G platform”. They would welcome having more vendors to choose from [a strong hint that they’d like Samsung or others to offer a 3G CDMA standards compliant femtocell].
There are different views on whether LTE femtocells need to be dual mode 3G/4G or not. The initial models on display today are all LTE only. Some vendors believe that with so many 3G devices around, it makes little sense to deploy 4G only femtocells. Others suggest that LTE will offload the larger data traffic component, leaving 3G for voice and less demanding featurephones. I haven’t had any insight from operators on this yet other than that they are still concentrating on their wide area macrocellular coverage first.
Just in case you thought that LTE/4G was the end of the road for technical development, Advanced LTE is just around the corner. The 3GPP standards people will be publishing Release 10 in March 2011 which will include this revised and improved version of the specifications. Apart from going faster and being even more spectrally efficient, a major item is Hetorogenious Networks (or HetNet for short). In laymans terms, this means a mix of large cellsites and lots of smaller cells to infill capacity, all talking nicely to each other to optimise the network capacity. The synopsis in this IEEE call for papers explains it better than I could.