http://labs.spotify.com/tag/postgresql/
And MySQL Cluster is a cohesive, built-in, well supported version for scaling MySQL. PostgreSQL has no equivalent.
But the PostgreSQL team are doing their usual slow, stepwise refinement approach to implementing these features from the primitives and moving up. I expect that in a few versions they'll be at sufficient feature parity with MySQL on this front that anyone who cares enough is the sort of person who decides between Oracle RAC and Teradata.
While that will be good and I look forward to it, I will miss seeing you turn up in these threads like a bad penny.
My position is that there a lot of different products out there that cater for different needs and there is no "one size fits all" solution. I can scale Cassandra out to a hundred nodes in minutes on EC2 with no configuration changes. I also have queries in MongoDB that are literally 50x faster than on a SQL database.
I think that for any case where you might be building a new system based on a relational backend, it's basically true (modulo local constraints like "we're an Oracle shop"). The chances that you will need to run a website that needs 100 Cassandra servers any time soon is ... well it's unlikely.
PostgreSQL is a stable, proven workhorse. That's why I like it. My point of view is that you should start with high safety and features and relax those constraints as circumstances demand.
And I comment on lots of different things but have a stronger interests in Apple and Database. So sue me I guess.
"Later came Postgres 9 and with it the excellent streaming replication and hot standby functionality. One of the most important database clusters at Spotify, the cluster that stores user credentials (for login), is a Postgres 9 cluster."
Hum, your link kind of defeat your point. That or I did not understand the point you were trying to make.