Last week I attended Velocity 2011 in Santa Clara, CA; this was the 4th edition of Velocity and my 2nd one. This means that even if I was still very much impressed with everything what happened during these 3 days, it was not quite so mind blowing like last year, and you will be able to see this from my impressions bellow.
First impression was that this was much bigger than last year, and the numbers just released by O’Reilly demonstrate it was almost double with about 2,000 participants and another sell out (compared with 1,200 in 2010). Considering the high price of the conference (~1k) this is very impressive and interesting to see how they will host next year even more people. From a question a speaker had many people where at their first Velocity, but even so I met many friends and people I met last year. From the content point of view there was a lot of focus on mobile performance that I haven’t seen last year. Myself I mostly followed the operations track, and only a few talks in the performance track.
Tuesday was the “workshop” day, and I feel that they should change that name as none of the workshops I attended (OpenStack, Chef, Postmortem, etc.) were nothing else that talks (some quite interesting ones) that were maybe longer than a usual conference talk, but not workshops. This is confusing and people I spoke didn’t really liked it either. Not even the Chef workshop that could have had a practical exercise (and even if they sent instructions for people to configure their laptops to prepare for this), had nothing for the people in the audience to test and get their hands dirty. Even if this was not a real workshop either the talk by John Allspaw: “Advanced Postmortem Fu and Human Error 101” (slides) was the best of the conference in my opinion and I highly recommend it to anyone interested in the topic (if of course O’Reilly will publish the video). The day ended with a very good Ignite session (including a karaoke part that was quite entertaining).
Wednesday and Thursday were the main days of the conference with even more people arriving in Santa Clara. I’ve met people that traveled from all over US, but also from other parts of the world (Europe in general). Each day started with plenaries and sponsored talks in the morning until lunch. Personally I would cut down on these plenaries and sponsored talks, or at least give people some alternatives; even allowing for BoFs or any other talks would be great. From the actual presentations, I liked Mark Imbriaco talk “Building for the Cloud: Lessons Learned at Heroku”, Patrick Debois and Andrew Shafer “Measuring the devops gap” (slides). In general all the talks I attended were very good. I had high hopes for Adam Jacob’s “Choose Your Own Adventure 2” (slides) but this was not even close as funny as the one from last year (not sure why; maybe the room was too big, but also the content was not quite so funny).
On another note, Velocity Europe was announced to be happening sometimes in November in Berlin. Also Jesse Robbins passed the co-chair duties he had for the last 4 years to John Allspaw.
Overall, I would say that Velocity is still the conference to go for anyone serious about operations (even though there are some other good options appearing: DevOpsDays, Surge). Velocity is a great, high quality conference that gathers the best people in operations and web performance field. Still for me, I will probably follow next year only the ‘hall/bar track’ as that is what I found to be the most valuable; and of course DevOpsDays that followed Friday and Saturday, is something I wouldn’t miss for anything.