I just completed my talk at the 2007 Connectathon, the yearly NFS gathering for interop testing. I’m thankful to Tom Haynes for having me. It was a nerve racking exercise to prepare for it… what do I have to share with the guys that wrote NFS? I described it to my wife this way: “You use Post-It Notes right? – Yeah. -What if 3M asked you to give a presentation on them? – Your screwed.” Jason Hoffman came to the rescue and helped me brainstorm a presentation where I focused not on NFS itself, but the ecosystem in which NFS exists… because thats something I know much better. All in all things went well. It was definitely a learning experience and I’ve always dreamed of attending a Connectathon… I never imagined I’d be speaking at one.
Here are the slides if anyone is interested: Connectathon2007.pdf
The Connectathon floor isn’t what I expected. I imagined racks of screaming systems in a booth arrangement with engineers frantically darting back and forth across the room with laptops… sorta like the NYSE floor for storage interop. The reality is quite different. A gigantic room (conferences are commonly held in the same room) with 3 long rows of tables with chairs. Laptops and workstations are sitting at almost every other chair and engineers plugging away largely silently. There were storage setups, often 1 or 2U systems and an array scattered around under the tables that you won’t even notice just walking around. Only two full server racks of (can’t say). I was surprised by how much of the testing is done on laptops and workstations. This is largely because performance isn’t what they are testing, its interop.
One other thing I learned that was interesting is that all talks at the Connectathon are open to the public! No registration needed. So if you too geek out at the idea of being at a Connectathon, you need only to show up. There is still a day left of talks, so if your in the Silicon Valley and want to listen to NFS goodness check out the talks list: Connectathon Talks Schedule.