Last night YouTube TV went down for an hour. That’s not YouTube the platform, but the premium TV service that YouTube offers customers in the US a range of broadcast TV channels in exchange for a monthly fee. The service went down right in the middle of the England v Croatia World Cup semi-final in Russia.
Every time a set of major sports rights comes up for sale, there is more and more discussion about whether a major internet platform like Amazon, Facebook, Google or Apple will be bidding. So far, there have been a few toes dipped in the water. Amazon has a small package of Premier League games from the season after next; Amazon also has ATP tennis in the UK from next year, and has had a few tennis tournaments this year; Amazon has streaming Thursday Night NFL rights, sharing them with free-to-air and pay-TV ; Facebook has bought Premier League and La Liga rights for a handful of Southeast Asian countries.
But at the same time, there are ongoing problems with many of these streaming technologies. In Australia, Optus had massive issues with its World Cup rights as I’ve mentioned previously. They’ve ended up refunding subscribers, and allowing all their games to be shown on free-to-air broadcast TV. ITV Hub has had various issues during earlier games in this World Cup (although I’ve seen few reports for the semi-final last night). Hulu’s stream of this year’s Super Bowl went down towards the end of the game. There are plenty of other examples.
Streaming is hard, and the resources to ensure no breaks are not to be understated. You might get angry if you can’t stream an episode of GLOW on Netflix because something between Netflix and your ISP isn’t working right. The worst that might happen is that you have to wait a bit and watch it later. But that’s not a remotely satisfactory solution for live sport.
If a company the size of Google can still have a major outage during a global event like the World Cup, then you know that this isn’t easy. During the Sweden v England quarter-final, the BBC reported a record 3.8m live streams at one point. And of course, there were also reports that the stream fell over towards the end of the game for some.
It’s notable that for the World Cup, the BBC’s UHD streaming experiment was initially limited, to ensure that those who got a stream weren’t going to be disappointed half way through when too many other viewers caused the whole system to fall over (Of course, viewers would quickly find out that they were well behind other versions of the picture meaning that you could be hearing your neighbours cheering a goal minutes before you saw it yourself).
The same fixture had broadcast viewing figures of over 19m, with many more watching in pubs and at outdoor events. And while we need to be careful about comparing audiences (1 stream does not equal one viewer; they are not measuring exactly the same thing), it’s clear that the vast majority still watch via the more robust broadcast systems. The question is, for how long?
Talk to a TV engineer and you’ll begin to understand why broadcast is still better. The Freeview transmitter network is very robust with built-in redundancy to ensure that TV channels’ signals reach local transmitters. While local transmitters can fail, these tend to be extraordinary events, and their “up time” is high. If the transmitter is working then the only reason you don’t get a picture at home is down to your set-up (e.g. a faulty antenna on your roof). Satellite transmission is also remarkably robust – with perhaps only extreme weather causing picture degradation.
With IP, there are many places that the system can fail. Broadcasters are reliant on large Content Distribution Networks (CDNs) to distribute programming. And that complexity increases with live. Then there might be a local problem with your “exchange”, or even the local fibre cabinet near to your street. Perhaps your the free router your ISP gave you has failed. It can be hard to diagnose, and there are many potential points of failure.
For the most part, service will probably resume quickly. But just how quickly is another question.
I’m not arguing that IP can’t fix some of these problems, or be more robust. But I do think that it’s going to be a significant technical challenge, with many parties involved, and broadcast is better in many respects. From a broadcaster to transmitter might only involve a couple of specialist companies. The pictures arrive faster, and there are fewer places for things to break. One viewer or 30 million viewers? It makes no difference.
On the other hand, some future live event will take the record for streaming again, but these will be more worrying moments as systems are put under bigger pressure than ever before.
I’m not ready to give up broadcast as efficient video and audio propagation methodology just yet.