Global cellular networks are reeling under remarkable growth rates that look like they will be sustained over the next five years, and this will be driven by video, says a recent report.
According to the Cisco Visual Networking Index (VNI) report, with its quarterly update out this week, video accounted for over 50 per cent of all mobile traffic for the first time in 2012 and this will rise to 66.5 per cent by 2017, growing at a CAGR (compound annual growth rate) of 75 per cent, faster than any other mobile application category, driven both by increased consumption and higher resolutions, pushing up bit rates.
Overall global mobile data traffic will increase 13-fold between 2012 and 2017 at 66 per cent CAGR to reach 11.2 exabytes (1018 bytes) per month by 2017, of which 7.4 exabytes will be video. Of course such projections are never exactly right, and it could well be that video will account for an even bigger proportion, as that part of the forecast looks slightly conservative.
At least Cisco VNI is in tune with Akamai’s State of the Internet Report in its view that total mobile data traffic almost exactly doubled during 2012 (Cisco reckons 96 per cent), and our own Faultline report on the subject, about to be issued next week. But Cisco drills down further by pointing out that a lot of the extra traffic was diverted onto the fixed broadband infrastructure, mostly via Wi-Fi, to relieve congested cellular backhaul networks. Globally, 33 per cent of total mobile data traffic was offloaded through Wi-Fi or in some cases femtocells in 2012, so that the total load on cellular core networks grew at the slower rate of 70 per cent rather than 96 per cent.
Cisco does not believe this growing trend towards offloading ever more data onto fixed networks will continue indefinitely, arguing that mobile and fixed network growth will converge to the same growth rate. The assumption presumably is that cellular operators will invest sufficiently in backhaul and core infrastructure to meet the growing data demand. But this reckons without the reality that 4G/LTE does not on its own deliver the extra radio bandwidth that will be demanded by future mobile services. The Swedish mobile connectivity and security software vendor Birdstep has calculated that while LTE gives a radio efficiency improvement of about 12 times, that has already been more than soaked up by mobile traffic growth, which has increased by about 30 fold since 4G was first proposed.
So already 4G/LTE is lagging behind the bandwidth curve and the gap will continue to grow. The only recourse will be to increase cell density via pico and femtocell deployment, or by harnessing the emerging global network of Wi-Fi hot spots, using ANDSF (Access Network Discovery and Selection Function ) and HotSpot 2.0 in combination to facilitate the automated network discovery and selection currently performed by cellular networks. Together ANDSF and HotSpot 2.0 allow mobile devices to discover Wi-Fi roaming relationships more readily, determine access point capabilities and loading conditions, and connect to Wi-Fi networks securely and seamlessly.
Even where femtocells or picocells are deployed, mobile data will tend to be offloaded onto broadband infrastructure largely because it is there and will usually provide a far more cost effective backhaul option than upgrading the operator’s own network. So as mobile data proliferates, an increasing proportion of it will be offloaded onto broadband, perhaps as much as 90 per cent eventually. Some even suggest it will approach 100 per cent, becoming an end game for fixed/mobile convergence. Cisco itself predicts that 46 per cent of traffic generated by smartphones will be offloaded onto broadband by 2017, but that the figure for tablets will be 71 per cent as they will continue to be used more for consuming video in hotspots.
In the meantime the Cisco VNI Forecast does provide some interesting numbers on 4G/LTE deployment and the amount of data it will create. Firstly it noted that mobile network connection speeds more than doubled in 2012 to reach 526 Kbps, but that was mostly because of 3G build out. For smartphones the average speed was 2.064 Mbps, plenty fast enough to watch high quality video at that screen size.
But Cisco found that 4G connections generated 19 times more traffic on average than the rest, comprising a mixture of 3G and 2G. As a result 4G services already account for 14 per cent of all mobile data traffic, even though they represent only 0.9 per cent of connections. By 2017 Cisco predicts that the average mobile network connection speed will have increased more than seven times the 2012 value to reach 3.9 Mbps, but with still only 10 per cent of connections being 4G, and even by then not quite half the traffic, at 45 per cent.
Like the Akamai report, the Cisco VNI devotes a lot of space to IPv6, but with some hard projections as well as current numbers. Across the whole sector Cisco predicts that 4.2 billion or 41 per cent of global mobile devices will be IPv6-capable by 2017, up from 14 per cent or 1 billion in 2012. Two market segments with strong affinity for IPv6 but for different reasons are laptops and machine-to-machine (M2M) devices, the former because they have IPv6 enabled by default when they connect to a mobile network, and the latter because IPv6 will be necessary to support the exploding number of connections in the “Internet of Everything.” With the IPv4 address space on its last legs, Internet Registries have been rationing IPv4 addresses, and there are certainly not enough to sustain much growth in M2M applications such as smart metering and remote environmental monitoring in the home.
There will also be significant penetration of IPv6 in the big consumer segment of smartphones and tablets, with 73 per cent or 2.2 billion being IPv6 capable by 2017, compared with 41 per cent or 479 million in 2012.
Of course a device can be IPv6 capable but still be connected to an IPv4 network as many will, and Cisco is careful to avoid any forecast of the breakdown between the two of mobile data. Nobody seems prepared to stick their neck out on that, and it is possible that the applications generating the most data, including video, will still often use IPv4. That said, IPv6 does offer compelling advantages for video services in terms of QoS and multicast support.
But whatever the extent of IPv6 take up, Cisco like Akamai is convinced that mobile video adoption will continue unabated over the next five years and put big pressure on backhaul. It is vague what form this backhaul will take and who will provide it, and it fails to identify the huge role that Wi-Fi offload will play.
Copyright © 2013, Faultline
Faultline is published by Rethink Research, a London-based publishing and consulting firm. This weekly newsletter is an assessment of the impact of the week's events in the world of digital media. Faultline is where media meets technology. Subscription details here.