It seems to me that this has nothing to do with Google. Facebook is not a Google competitor...and Google would see no benefit from open sourcing their infrastructure. Google is an infrastructure company--the value of Google is in the strength of their search and related technology. The value of Facebook is the network of people--technically, it's simply not a difficult problem. Take away the user community and Facebook has almost zero value. Facebook happened to hire a lot of smart people who wanted to work on interesting things, and so NIH has led to lots of new technologies being created.

Their solutions may be optimal for Facebook, and they may even be generally useful to folks building large scale websites. But, their Open Source tools are no threat to Google, and would have no impact on Google's decisions about what to Open Source.

Hadoop/HBase has already been offering BigTable like capabilities in an Open Source project. This is merely another option--and frankly, one that sounds a bit half-baked at the moment. But I could be wrong about that. And, of course, Thrift is yet another RPC system. It makes pretty bold claims about being uniquely lightweight and cross-language, but the RPC and data interchange problem has been solved many times in the past in many ways. I don't think the world has been set on fire by Thrift.