Re: PostgreSQL, OLAP, and Large Clusters

On Thu, Sep 27, 2012 at 11:34 AM, Ryan Kelly <rpkelly22(at)gmail(dot)com> wrote:

>
>
> > At this time I would try:
> > - Postgres-XC
> From what I understand, more of a write-scaleable-oriented solution. We
> mostly will need read scalability. I also don't think it really handles
> redundancy.
>

From my understanding it gets around the key read scalability issue in
PostgreSQL, which is a lack of intraquery parallelism. Since components of
a query can run on different storage nodes concurrently, this helps a great
deal. It doesn't do the things a column store would help with but it is
still a major step forward.

As for redundancy, Postgres-XC handles redundancy on the coordinator side,
but on the storage node side, I believe you could use streaming replication
and other standard PostgreSQL approaches to redundancy there.

Hope this helps,
Chris Travers