> Although Jim has a point in that we are seeking solutions that scale
> beyond a single machine

True but that doesn't mean that all parts of the cloud are peers (all
doing the same thing). A good design would have a set of roles which
computers would fulfill. Data store is one of those roles which should
be given to specialized equipment.

As already discussed the cost of coordinating data changes across the
"cloud" is the real issue.

The Google cloud has a large number of roles, which are performed by
separate machines. In fact, it is common for the same operation to be
performed by simultaneously multiple machines.

Where the Google cloud has a significant advantage over a general
database solution is in the fact that they don't need to be concerned
about transactions and data consistency for searches.

The fact that an indexer station has parsed a web site and has sent out
updates to be applied against the search indexes; doesn't mean that the
search results which are generated 1 msec later are invalid -- hey it's
only a search!

So, rather than think of the cloud as meaning that all of the engine
functions are performed by each machine, it is better to think of
de-constructing the engine into its raw pieces and seeing how to
distribute those functions.


Sean