At Flickr Engineering our favorite thing is rainbows. But fast stable websites come a pretty close second. So last week some of us drove down to San Jose to take part in the 2009 O’Reilly Velocity conference.
The conference consists of two tracks, performance and operations. The attendance was a great mix of people working on optimizing the whole web stack – everything from the download size of CSS to the PUE of data centers. There were talks on automated infrastructure, Javascript, MySQL, front-end performance recommendations and CDNs, but the most significant theme was the direct business benefits that companies like Google and Microsoft are seeing from microsecond speed improvements on their sites.
Some of the most interesting sessions for for us were:
- John Adams talking about Twitter operations (details , slides , video )
- Eric Schurman and Jake Brutlag’s joint presentation about performance benefits seen at Google and Bing (details , slides , video )
- Andrew Shafer’s discussion of Agile Infrastructure (details )
- Adam Jacobs and Ezra Zygmuntowicz’s talk about Cloud Infrastructure (details )
John and I also gave a talk on how Flickr’s engineering and operations team work together to allow us to iterate quickly without causing stability problems. The full video is available, and here’s the slides:
Thank you to Jesse and Steve for putting together a great conference. We’re already looking forward to next year.