Hi all:I wanted to give the PostgreSQL community a heads up on ongoing
database research here at UC Berkeley, which may be of interest since
it uses PGSQL.
The last few years we've been building a system called Telegraph, which
has a number of research thrusts:
a) aggressively adaptive query optimization (based on the "eddy"
concept)
b) queries over external data sources, including "pull" (e.g. by
scraping web pages) and "push" (e.g. data feeds from sensors)
c) support for Continuous Queries over streaming push sources,
including multi-query optimization and fault-tolerant parallelism
The first version of the system was written from scratch in Java, which
caused us some pain About 8 months ago we abandoned our initial
version of the system, and decided to start over in C. Rather than
starting over from scratch, we decided to take PostgreSQL and enhance
it (and also disembowel it somewhat) for our purposes. We're planning
on releasing an alpha of the new system, called TelegraphCQ, in a
couple weeks. It will include some but not all of the above features,
and we'd be happy to have folks kick the tires a little.
It will be interesting to see how/if our modifications can be of use to
the broader PGSQL community. I encourage interested folks to have a
look at our website at http://telegraph.cs.berkeley.edu, read some of
the papers, etc. The best overview is our recent CIDR paper,
http://www.cs.berkeley.edu/~franklin/Papers/TCQcidr03.pdf
(Please see the related note on how we're now using PostgreSQL in our
database classes both at Berkeley and CMU.)
Regards,
Joe Hellerstein
--
Joseph M. Hellerstein
Professor, EECS Computer Science Division
UC Berkeley
http://www.cs.berkeley.edu/~jmh