Building the BBC's Olympic site

I was really blown away by how awesome the BBC's coverage was. You could watch every event, live and delayed, and they would even bookmark the live stream as you watched it (and preserved for replays) with event markers. This was particularly useful for track and field where you might want to skip around on various qualifiers and other events.<p>We never had a single issue with buffering or playback, and it was truly better than TV. It definitely felt like the future. I was really sorry our American friends in the USA couldn't experience the coverage like we had in the UK (without a VPN).

For those of you interested in hearing more about this, there's a great episode of the podcast Framerate where they talk to Andy Armstrong, who was behind all of the streaming of the BBC olympics.<p>There's some really interesting tidbits in there and he offers a candid perspective of how it went and the future of streaming TV, etc.<p><a href="http://twit.tv/show/frame-rate/89" rel="nofollow">http://twit.tv/show/frame-rate/89</a>

The full diagram is available here: <a href="http://www.bbc.co.uk/blogs/bbcinternet/2012/08/21/2012_Architecture_Overview_Diagram_20120727.pdf" rel="nofollow">http://www.bbc.co.uk/blogs/bbcinternet/2012/08/21/2012_Archi...</a>

The interesting part to me was the fairly heavy use of RDF triple stores. The BBC have used these to good effect before on the world cup site and waxed lyrical at the time about what a difference it made modelling the data in this way. AFAIK they use BigOwlim from Ontotext as their store, run in cluster mode.

I see they are talking about this at Velocity Europe <a href="http://velocityconf.com/velocityeu2012/public/schedule/detail/25792" rel="nofollow">http://velocityconf.com/velocityeu2012/public/schedule/detai...</a>

PHP - another victory.

The Paralympics coverage is abysmal by comparison. It's really sad.