Sudden spikes in demand for video streams, usually caused by major events, can cause "content storms" -- posing a significant challenge for operators. These spikes are becoming more common, and their peaks are rising higher, as smartphone penetration and mobile video consumption increase.
As EE's video chief, Matt Stagg, is responsible for preparing for and dealing with these spikes. He heads up content technology in the mobile arena for EE , which is now owned by BT Group plc (NYSE: BT; London: BTA). In this role, he evaluates user behavior and adoption of new mobile content and services, and explores new technology solutions such as LTE Broadcast and edge computing in the mobile network. (See EE Mobile Video Chief Dissects Big Video, All Video Towards Broadband and EE Video Chief Matt Stagg Dissects Big Video, Part II.)
Stagg is also the chair of the Mobile Video Alliance, a community of operators, broadcasters, CDNs and content owners working to improve mobile video; and is part of the LTE Broadcast Alliance, looking at LTE Broadcast as a way to manage video delivery at peak times, along with
KT Corp. , Verizon Communications Inc. (NYSE: VZ) and Telstra Corp. Ltd. (ASX: TLS; NZK: TLS).
Stagg is also a founder member of Telco Transformation's VTAB -- the Video Transformation Advisory Board, advising us on major industry developments and sharing his insights with our readers. (See Introducing the Video Transformation Advisory Board .)
Telco Transformation recently connected with Stagg, to follow up on a presentation he gave at the Mobile World Congress earlier this year, on "content storms."
Telco Transformation: Your presentation at Mobile World Congress addressed content storms -- what are they exactly?
Matt Stagg: It's really any event that drives a large increase in demand within a limited time-frame. Large events will always drive a traffic spike; we've always had appointment viewing in a linear TV environment. But in VoD this is quite new. We are used to people watching video of missed shows or older footage, but now new content comes out weekly from Netflix, Amazon, etc., and they've created demand for these shows so people start watching within seconds of each other. For example with The Grand Tour [one of Amazon's flagship shows], we saw people starting to watch within a very, very short time-frame from each other. And we also see this with software, such as iOS or Android updates.
TT: What challenges does this create for network operators?
MS: The main challenge is that we can't scale infinitely to support huge demand spikes -- it doesn't make commercial sense. We have to build for the busy hour, we understand that. We build network capacity, and also manage offers and tariffs with that in mind. But when something comes along, like a major football match that just drives a huge audience above that peak, it's a problem.
We have that capacity for now, but even our network can't scale infinitely. But that excess capacity has helped us, and is helping us now, understand user behavior. And this in turn allows us to better understand and potentially manage content storms in the future, where capacity really will be strained. We can see where and how big these spikes are, so we can better engineer solutions and direct new technology development to cope with that demand in future.
TT: Are there particular events that you can recall that led to a recent content storm?
MS: Yes, I think a good example would be the England versus Wales game [at the Euro 2016 soccer tournament] last summer. Even though it was last year, it's a good example, because I think it's the most traffic we have ever seen on our network. We're getting close now, with some sporting events, but at the time it was a massive spike.
Content Storm: England v. Wales, July 16, 2016
See, with sports, there's a sort of hierarchy in terms of viewing preference. Most people's number one choice is to see the event in the stadium, live. If you can't do that, then the second-best option is to see it on a large screen, maybe in 4K, at a pub or at home. And it's only if you can't do that either, then you stream the game to a mobile phone.
So why was mobile viewing so high for this game? Of course, England and Wales are both UK teams and don't play each other very often, so you would expect interest to be high. But that wasn't enough to explain why it was so high. We really wanted to go through the analytics, and really understand what the reasons for this spike were, so we could be better prepared the next time it happened.
One reason was that in the UK most major soccer matches are broadcast on pay channels like BT Sport or Sky [since they usually buy out the rights]. Traffic is limited as a result, because they have only so many subscribers. But this was broadcast on BBC, so everyone had access to the coverage.
Weather can also be a factor, though it wasn't in this case. But on warm, sunny days people want to get out, and you will have people watching video on mobile while they are in parks, or gardening, barbecuing etc.
Another really important factor for this game was that the kick-off was at 2:00 pm I think, and it was midweek. Normally, a lot of streaming is offloaded to Wifi networks, but at this time a lot of people were at work. And many corporate networks limit Internet access on their networks. So people couldn't watch it on WiFi, they had to switch to 4G.
We have the [soccer] World Cup coming up next year in Russia, and the rights are again with the BBC and ITV [another UK broadcaster], so they will be open to everyone. We could face a similar scenario, so we picked up some useful learnings from the England-Wales match.
Also the time difference with Russia means we'll probably have matches on during the afternoons, and often during the week.
So we can expect to see more of this behavior.
— Aditya Kishore, Practice Leader, Video Transformation, Telco Transformation