Re: mutex contention for the query cache

[Brian Aker]

Hi!

On Jul 4, 2008, at 4:55 PM, MARK CALLAGHAN wrote:

> There is a Stonebraker research project on an OLTP engine called
> H-store that partitions data so that one core gets each partition.
> Locking is not an issue at that point. MySQL Cluster could approximate
> that deployment model.

This is the one place where in the discussion of thread vs process  
that process based DB's (and what I am really talking about is  
Postgres) could bind processes to particular processors/sets of cores  
and arbitrate out access to those workers. By using processes you get  
a natural affinity for any sort of lock call to stay with the confines  
of a particular processor.

You could get this by using "node collections" of MySQL servers. This  
was something that was talked about after we looked at a new design  
for prepared statements so that proxies/anything could do a bit easier  
work on incoming queries.

Cheers,
	-Brian


--
_______________________________________________________
Brian "Krow" Aker, brian at tangent.org
Seattle, Washington
http://krow.net/                     <-- Me
http://tangent.org/                <-- Software
_______________________________________________________
You can't grep a dead tree.