Next up was Jay Parikh from Facebook to talk to us about the Open Compute Project. Jay jumped in to explain where the project has come since last year … but didn’t explain what the project was, so this is from the official site:
We started a project at Facebook almost two years ago with a pretty big goal: to build one of the most efficient computing infrastructures at the lowest possible cost. We decided to honor our hacker roots and challenge convention by custom designing and building our software, servers and data centers from the ground up – and then share these technologies as they evolve.
Jay talked about 2 projects that Open Compute is focusing (there are many more). The first was the OCP Networking project which focuses on creating a vendor agnostic switch to allow us to separate our hardware from our software – giving us more choices.
The next project focused on data – our photos in particular. There are over 200 billion photos on Facebook that are not accessed that often. Facebook can’t lose the photos and the users won’t accept if things slow down because the system has to store all of these photos in the same place. So what they decided to do was create a data storage center that was separate from the compute heavy server center. The goals of this new center were:
- handle massive scale
- they have to be durable – can’t lose data
- they have to make things highly efficient
Jay then went in to more hardware-type information than this software girl could follow along with … but it all looked pretty darn impressive.