Hi Christian,
This white paper might help...
http://www.mysql.com/why-mysql/white-papers/mysql_wp_cluster7_open_cgp.php
And especially this one...
http://www.mysql.com/why-mysql/white-papers/mysql_wp_subscriber_db.php
Both papers have recently been refreshed.
Regards, Andrew.
> -----Original Message-----
> From: Jonathan.Stephens@stripped [mailto:Jonathan.Stephens@stripped] On
> Behalf Of Jon Stephens
> Sent: 24 October 2009 20:54
> To: Christiaan den Besten
> Cc: cluster@stripped
> Subject: Re: Designed for cluster?
>
> Christiaan den Besten wrote:
> > Hi !
> >
> >> I think by "massive in parallel" he means that cluster allows many
> >> different instances of MySQLD, to interact with the cluster. So not
> so
> >> much the threads, but actual different processes, on different
> machines.
> >
> > Indeed ;)
> >
> >> So we cluster the SQL nodes, and also we cluster the storage of the
> >> data. If the access pattern for the data is suited, then cluster
> >> scales very very well. If the access pattern is not suited to
> cluster
> >> then it performs poorly and does not scale.
> >>
> >> Note that each SQL node sees identical copies of the data, the data
> is
> >> synchronously available to all SQL nodes.
> >>
> >> Given 4 nodes 2 copies of the data and 8 SQL nodes. You have the
> >> following capabilities.
> >>
> >> You have 8 SQL nodes that handle authentication, thread management,
> >> parsing of queries.
> >>
> >> You have 4 machines that handle transaction management, and data
> >> retrieval.
> >>
> >> Without going into details, accessing a row based on primary key is
> >> very fast, and scales very well. Although this _individual_
> operation
> >> might not be any faster than innodb, or MyISAM, it can perform
> equally
> >> fast from each of the connections, and it scales. Where sooner or
> >> later as concurrency increases MyISAM or Innodb will face a
> bottleneck.
> >>
> >> Other types of access beside single row, primary key. They do not
> >> perform/scale as well with cluster.
> >
> > I think Tom's description of when to use MySQL Cluster should be the
> nr
> > 1 faq entry :)
>
> I don't know why this never occurred to me before, but sometimes one
> can
> actually be a bit too close to something to see it clearly.
>
> Thanks for the excellent suggestion -- can't promise it'll show up
> first
> in the MySQL Cluster FAQ, but I'll get it in there soon.
>
> cheers
>
> jon.
>
> >
> > bye,
> > Chris
> >
>
>
> --
>
>
> Jon Stephens - jon.stephens@stripped
> Technical Writer
> MySQL Documentation Team
> Sun Microsystems AB
> MySQL and Software Infrastructure Group
> Liljeholmen (Stockholm), Sweden
> Summer: UTC +02.00 / Winter: UTC +01.00
> Mobile: +46 (0) 736 773 993
> Skype: plastic-fish
> MySQL: www.mysql.com
> Sun: www.sun.com
>
>
> --
> MySQL Cluster Mailing List
> For list archives: http://lists.mysql.com/cluster
> To unsubscribe:
> http://lists.mysql.com/cluster?unsub=1
