[base] Re: BASE performance/limits

[Carl T. <ca...@th...>]

kh...@sf... writes:

> I am interested to find out what the limits of BASE are. 

Me too.

> I suppose that the amount of experiments that one can store and analyze 
> depends mostly on the underlying relational database, operating system 
> and computer hardware. But there might be some performance issues related 
> to the design of BASE itself as well, right? (Please correct me if I'm 
> wrong!)

There certainly could, and there are a few places that could really be
bottlenecks, such as the RawBioAssayData and GeneListGene tables.

> Specifically: Has anybody used BASE for dealing with data from more than 
> 10,000 hybridizations, accessed by 20-30 researchers (not necessarily 
> simultaneously)? 

300-some hybridizations and 10 million rows in RawBioAssayData here,
and no complaints from the users yet.

> Any estimates on what hardware setup would be required for that type of 
> scenario (CPU, RAM, disk-space, etc.)?

Two CPUs per concurrent user is at least an upper limit for what's
useful, I'd say. Disks fast as hell, and lots of RAM, but I really
don't know how much, what CPUs, etc. would be needed for reasonable
performance.

//Carl

-- 
 Carl Troein - ca...@th...
 http://www.thep.lu.se/~carl/
 BASE developer - http://base.thep.lu.se