EE will use LTE Broadcast as part of a range of tools it will deploy as it builds a network optimised for video distribution.
EE will have live LTE Broadcast services in 2016, according to Matt Stagg, Senior Manager of Network Strategy.
The operator will carry preliminary trials at a sporting event in 2014 [TMN has heard independently this may be at the Glasgow Commonwealth Games] and further trials through in 2015 as well as a stadium deployment. A wider commercial eMBMS rollout would then take place in 2016, Stagg said, dependent on there being sufficient compatible devices in the market to justify the business case.
Stagg said that LTE Broadcast would be viewed as a tool within a box that would deliver the best video distribution network. Its best use case is in cells where multiple people are likely to be watching the same content, such as at music festivals or sports events. That means we will not see a network-wide LTE Broadcast deployment from the operator.
With video dominating data traffic volumes across LTE networks, EE has recently tasked Stagg's team with looking at what a video distribution network would look like.
So what technologies does EE think will form part of this video distribution network.
Current video optimisation techniques will not scale to future demand, Stagg warned, and operator controlled manipulation of content could also bring risks of litigation from content owners.
"Optimisation platforms will move away from the mobile network doing traditional transcoding and transrating. It will be too expensive to do that in the core. We have to make use of all the capacity we've got in the RAN so we don't want to be spending a lot of money on the core. I also think we will start to get litigation on that in the next couple of years [from content providers]. It's not the mobile operators' content, so un-wrapping it and changing it... I think we'll start to see some litigation on that now."
Stagg said he is trying to move from a view of video optimisation to customisation: a configuration based on a basic set of parameters, such as device type, access technology and time of day and what the user has paid for.
"These are levers in the network and then they become business rules, so you can start to get some innovative tariffing and services with content customers want, at the quality they want and that the content provider wants to see them in."
2. Adaptive Bit Rate
One thing Stagg turned his attention to is the limitations of Adaptive Bit Rate (ABR) technologies in LTE. "ABR is designed to work on networks with low capacity, and it works really well, but LTE networks don't have low capacity."
ABR technologies tend to mean that all video is constantly trying to stream at a higher definition. So a device will always tend to receive content at the highest available bit rate - meaning that a user may be receiving high def content on a device that they can't tell the difference on, yet still be paying for the extra data. They also make it impossible for operators to charge for, or users to understand video usage, on a volume basis. "We have no way of knowing in advance how much data half an hour of video on ABR is," Stagg said. "We need to take a step back and say, 'You have 10hrs standard def, 5 hours hi def and unlimited hi def outside of busy hour'."
3. Edge caching
Edge caching - ie caching popular content at the base station - is intended to reduce backhaul load and also provide a more responsive customer experience, but it is a technology that comes with its own issues. It means that billing, content filtering (adult filters for example), policy and subscriber related decisions that currently happen in the core ("Gi services" as Stagg termed them) will have to happen at the edge, requiring increased processing support. Stagg said the time had not yet come for edge caching but in backhaul-constrained locations, for example, edge caching will have a role. It will come, but not just yet.
4. LTE Broadcast
LTE Broadcast bring spectral efficiency that makes the business case for its deployment, Stagg said. Once 2-3 people are watching the same streamed content in a cell, you can dynamically move the users to a cell broadcast of that content - serving up to 100 people in a cell in 8MHz of spectrum. Everyone else in the cell will get a better experience for their usage, as the streaming users are moved to broadcast. That spectrum can be re-dedicated to unicast services once demand for broadcast has reduced.
"We see a lot of live TV that is sporadic, and you can never build capacity for that peak. So for us one things that's good is reducing one off investments, such Glastonbury. At stadiums the biggest traffic is people watching the game they've gone there to watch," Stagg said.
Stagg added that this per-use case deployment would allow EE to use it as part of a toolkit. "We're not advocating deploying it network-wide."
The next steps for Broadcast are to get handset density, with handset developers supporting the technology. EE will do a live stadium demo this year, with further live trials and live stadium launch in 2015, then further "device dependent" launches in 2016.
As well as eMBMS chip support in the device, LTE Broadcast requires a software update at the cell site, as well as some changes in the EPC and also OSS support to handle device firmware updates.
Read more on LTE Broadcast:
Qualcomm-iGR white paper
Ericsson white paper
Matt Stagg was speaking at What Network is Fit for Purpose?, an event organised by Cambridge Wireless.