RE: Multiple databases performance
- From: Linchi Shea <LinchiShea@xxxxxxxxxxxxxxxxxxxxxxxxx>
- Date: Tue, 23 Jun 2009 06:24:01 -0700
The approach is fine if you expect the the load on the server will max out
the server resource before the number of databases becomes too large.
Personally, I'd consider hundreds of databases on a single server to be too
large.
But from your description of the load, the traffic is so light that the
number of databases may cause a maintnenance issue before any real computing
resource (CPU, memory, disk I/Os, etc) becomes a bottleneck.
If the second sceanrio is more likely, you may want to consider partitioning
your data inside the same database instead.
Linchi
"Jonny Bergdahl" wrote:
I have developed a system that potentially will be run by thousands of.
customers. In order to make it easy to partition the system on multiple
servers, I have opted to use a separate database for each customer. That
way, as the number of customers grow, I can easily add SQL servers to divide
the load.
The number of transactions for each database is low, typically we have 3
transactions/hour with a typical peak usage of 120 transactions/hour.
Are there any special issues in such a setup? What is the performance issues
of using a large number of databases? How many databases will I be able to
handle on each server?
Regards;
/jb
- Follow-Ups:
- Re: Multiple databases performance
- From: Jonny Bergdahl
- Re: Multiple databases performance
- References:
- Multiple databases performance
- From: Jonny Bergdahl
- Multiple databases performance
- Prev by Date: RE: understanding metrics/performance
- Next by Date: kill the long run job automatically
- Previous by thread: Re: Multiple databases performance
- Next by thread: Re: Multiple databases performance
- Index(es):
Relevant Pages
|
