A 3GPP workshop, several vendor statements and some recent blog posts have been suggesting enormous data capacity increases and how they might be achieved – small cells being an essential part. A target of 1000x traffic throughput has been set. Is this realistic or even relevant?
Growing demand, capacity and expectations
Many industry pundits refer to the often quoted Cisco VNI report, which forecasts some 20 times more wireless data traffic in the next 5 years. Reports of actual traffic consumption in the past two years indicate that it is on track so far. For example, China Unicom just reported a 112% data traffic increase in the past year from their 219 million customers. Worldwide, the availability of low cost Android smartphones is growing prepaid data traffic volumes. The question is - will traffic will continue to double annually for years to come as predicted.
Gating factors such as price and performance would of course limit growth, but a network operator survey by Rethink Research reflects that most planners are actively working towards a 20-50x capacity expansion over the next five years. Consumers cleary expect more, with a CommProve survey reporting that 32% of UK respondents did not expect to be able to make phone calls and 58% use applications such as Facebook on their mobile at major sporting events.
By comparison, technology growth curves continue to astound – such as processor power, memory capacity or wireline broadband speeds. Why shouldn't wireless traffic growth match that?
Industry consultant Keith Mallinson believes virtually all commentators agree with this view of traffic doubling annually – a 15 to 30x growth by 2020, but there are others like Dean Bubley who have lower expectations.
The 3GPP standards organisation plans for the future
The 3GPP standards organization held a 2 day workshop in June. to consider their 5 year priorities. Major vendors and operators shared their outlook (view many of the presentations here). It predominantly addressed how this capacity might be achieved. Unsurprisingly, the outcome comprised a mixture of higher efficiency, more spectrum and smaller cells.
Specifically, large numbers of LTE small cells featured as the key component, contributing by far the largest gains.
What's surprising is the relative benefits from each of the three dimensions:
Their target was a 1,000 capacity increase based on 3 factors:
- 3x Spectrum increase
- 6x Spectral efficiency
- 56x Spatial efficiency (i.e. using small cells)
Why aim for 1,000 fold rather than 30 or 50x?
Looking at a graphic of peak traffic demand columns superimposed on national maps, you can quickly see that the highest demand by far is in the dense urban areas. It is said that some 95% of data traffic is handled by less than 10% of cells. It's these very high traffic zones, where many people congregate and want to use their smartphones, that present the greatest challenges. Whether it's 5 or 10 small cells per macrocell (rather than 56 as suggested above), the ability to add extra capacity precisely where needed then becomes a commercial decision (i.e can we make enough additional revenue from each specific new small cell installed) rather than a technical one (i.e. will the network break and/or performance be reduced if we install another small cell).
So an imporant element of the standards work program will include features on how to make the small cells co-operate efficiently and effectively, both with each other and with the existing macrocell network.
Network Optimisation Conference
Dr Kim Larsen from T-Mobile presented the keynote at Informa's mobile network optimisation conference this year, giving insight into the current and projected traffic bottlenecks and issues.
You can view his full presentation on Slideshare here
Interesting points from this presentation include:
- The introduction of CELL-PCH and fast dormancy (two optimizations to handle signalling traffic) has had considerable impact on signalling traffic, which previously had become a serious bottleneck. Signalling load/efficiency has dramatically improved with Apple iOS and Blackberry devices; Android less so.
- Strategic investment in Wi-Fi offload and small cells should be pursued more aggressively
- Traffic is highly concentrated in a small percentage of cells
- 3G will continue to grow dramatically until 2017 (because it will take time for 4G capable devices to become predominant)
- Traffic distribution across the user base is broadening, so that it isn't concentrated so much on a very few high traffic users.
Qualcomm's 1000x Views
Conclusion
I've read many articles about the impending data capacity issue recently. The consensus view of between 20x and 50x capacity growth in developed markets by 2017 seems to be a popular target. Regional factors will affect how relevant that figure is for each country.
This extra capacity won't be distributed evenly, focussing very high traffic throughput in hotspot areas. The target of up to 1000x capacity growth should meet the most demanding of locations, meeting and even surpassing the expectations of today's 3G smartphone users.
While LTE small cells are an important element, the large installed base of 3G smartphone devices will drive continued investment in 3G for several years. Wi-Fi offload will also have a major role to play.
A small cell architecture is an essential component, and the technology appears to be on almost every operator's roadmap.
