Small cells featured prominently at last week's LTE World Summit, where there were diverse views about the pace and shape of deployment. This premier event attracted several thousand delegates from around the world, including many operator presentations reporting their experience to date. Small cells and HetNets were firmly an accepted part of the technical solution. By contrast, there were more diverse views about the business case and value of LTE services.
There can be little doubt that LTE is the chosen technology for wireless communications of the future. It has been designed to be extremely efficient and adaptable, with many different modes and options. It's also proven in the field, with 40 Mbit/s speeds regularly being achieved even in high traffic environments. But the speed of deployment and the form it will take vary in different regions.
A snapshot of LTE today
While we read of over 60 commercially live LTE deployments, the reality is that three networks dominate. Verizon in the US has some 65% of all LTE subscribers, and the top five networks exceed 90%. While quite a few different LTE devices have been launched, the actual numbers in each network are fairly restricted. This is partly due to the different frequencies and modes assigned to LTE, for example where the LTE capable Apple iPad doesn't work outside the US.
Some regions are aggressively investing in LTE. Those without the option of 3G/HSPA, such as Verizon, have little alternative, causing their competitors AT&T and Sprint to advance their rollout plans. Japan and Korea have comprehensive plans to engineer extremely high capacity mobile broadband. Korea Telecom forecast capacity growth of up to 1,000 times, and detailed their plans for up to 650 small cells/radio heads per square kilometre in peak traffic areas.
By contrast, we heard that Orange Spain did not see a business case which justifies separate standalone LTE network deployments – with the economic situation today, they don't see their customers paying additional monthly fees for faster mobile broadband. Perhaps they will follow the lead of Yota, the Russian WiMAX network, which has almost completed its switchover to LTE and operates on a wholesale/white label basis for Megafon, another Russian national network.
So why have more than 60 networks launched commercially already? One reason may be the extra value this provides to the operator's brand. Consumers may choose to remain and/or upgrade with a network that demonstrates it is keeping up with the latest technology, even if they choose to remain on 3G for now. There may also be some regulatory requirements to launch within a certain timeframe.
The price of an LTE bit is the same as a 3G bit
Mark Newman, Chief Research Officer at Informa Telecoms and Media, observed that operators haven't been able to charge a premium price for LTE. Verizon data bundles are the same regardless of whether 3G or LTE is used, and Nordic operators have had to drop the premium price attached to LTE when it was launched. There's still more money to be made though, because LTE users consume more Gigabytes of data per month. With most operators charging by volume, this leads to higher revenues.
Some operators still believe that unlimited data packages make commercial sense. Sprint retains its unlimited data tariff, and Hutchison 3 offers unlimited data for £15 (about $25) in the UK. Ed Candy, their CTO, explained that telecoms is a capital business and that the cost of delivering the extra traffic was marginal. I'm not sure that everyone with congested networks would entirely agree.
Manish Singh, CTO of Radisys, believes that video traffic may be the differentiator for LTE. With video content taking such a high (and growing) proportion of mobile broadband traffic, while being sensitive to delay, perhaps this could be prioritised and priced differently to "best effort" web browsing and email traffic. He believes that policy management is here to stay, with Policy Enforcement, Video Optimisation and related functions all combined into a single box.
His view is that operators will need to make a clear choice between "embracing the pipe" and "delivering differentiated services". it will be difficult to succeed in both. For those focussing on the pipe, they should aggressively partner with Skype, Google and Facebook while lowering the cost per bit. For those seeking to differentiate on service, adopting RCS (Rich Communication Services), excellent video experience supported by IMS should allow them to create their own added value.
The shape and networks to come
Despite LTE having significant cost and performance benefits, the overall picture emerging is one where networks will comprise multiple technologies – 3G, LTE and Wi-Fi – with each used where appropriate.
Everyone seems to buy into this HetNet (Heteregenous Network) and Small Cell vision, where there are large numbers of small cells augmenting the existing macrocell sites. The nirvana is a combined 3G/LTE/Wi-Fi box, possibly with integrated wireless backhaul, but I heard operators talking about deploying 3G small cells first. Typical backhaul connection speeds of 100Mbit/s were suggested for LTE small cells (compared to 500Mbit/s to 1Gbit/s for macrocells) will give some advantage to those operators with their own fibre networks to connect them.
Selected soundbites from some other operator presentations:
- Orange Spain: Data traffic is growing, but signalling traffic is growing faster. They don't see the business case for a standalone LTE network of their own, instead believe that network sharing is a better option. Mobile broadband costs today are around $6/Gbyte on 3G dropping to $2.80/Gbyte on LTE by 2016
- Softbank Japan: Today, they have 200,000 Wi-Fi hotspots and continue to deploy more, but these become unusable in dense urban areas. There is a serious capacity issues on 3G due to smartphones, which is also an issue for other Japanese networks. Softbank launched a TDD-LTE network and have their CFO's blessing to adopt LTE as fast as possible because it is more cost effective than 3G. They particularly like TDD mode because it provides asymmetric downlink/uplink traffic ratio of 70/30, and so is much more efficient for internet access.
- Sprint USA are imminently launching LTE in 6 markets in 2012 with 3 devices, increasing to 15 devices by the year end. WiMAX will not be sold after 2012 (although will operate until 2015) and iDEN/Nextel will be decommissioned. They plan to upgrade to LTE Advanced in 2013.