Hi all,
I've been using Firebird for years, but mainly keeping a very low profile.
I'm planning a Health Informatics project with the University of
Tasmania, and as I've used Firebird to power all the other projects, I'm
planning on using it again.
The question I have is about data partitioning across servers.
Essentially, we're wanting 10 servers to handle the backend - not
because it can't do it with one, but we need to show that the DB can
scale (investors love clusters).
Initially, as every table is based on bigint primary keys, I could have
server 0 handling every record ending in a zero(0,10,20,etc), with
duplicated table structures on top.
The middle layer handles the MapReduce, determining which server to ask
for data, and combining data sets(reducing) for the app on top. (No,
it's not BigData, just an entity mapping layer).
When it comes to data partitioning across servers, what other strategies
have people done?
Cheers,
NW.
--
KarbonIT
M: 0404 351 640
E: nigel.we...@karbonit.com.au
W: karbonit.com.au
