-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1

thus David S. spake:

Hi Timo, I mean parallel cloud computing, do you have solution for this case?

First thing that comes to my mind is

http://hadoop.apache.org/

which we use ourself for cloud computing. There's a nice tutorial here:

http://www.ibm.com/developerworks/aix/library/au-cloud_apache/

(don't let 'AIX' shock you, it's about Linux).

HTH,

Timo

-- Best regards, David http://blog.pnyet.web.id

On 12/14/2010 05:24 PM, Timo Schoeler wrote:

...

thus David S. spake:

...

Dear All, I got a new project to build cloud computing base on centos clustering (clustering and cluster storage). whether failover, load balancing can be applied? I've read about CentOS clustering and cluster storage but I'm still confused, any help or advice in this thread will be appreciate.

Hi David,

what exactly are you trying to achieve? Is it just 'plain loadbalancing' of services (such as HTTP requests, IMAP or similar) or massively parallel cloud computing stuff?

...

Cheers...

Timo

On 12/14/2010 4:41 AM, David S. wrote:

Hi Timo, I mean parallel cloud computing, do you have solution for this case?

There are many different meanings for terms like that. One is a set of resources that you can easily carve into virtual machines that then act like unrelated physical machines except for the ease of cloning and migrating them. Another is a scientific supercomputer that can solve certain types of problems with specialized libraries that can divide that task into work units that run on different hosts and collate the results. This tends to be limitied and difficult, so another approach is to use a scalable database where you push everything that needs to run in parallel into a 'map/reduce' function. That tends to be difficult by itself, and handled differently by the different implementations. If that's what you are interested in, I'd look at 'riak' first (http://wiki.basho.com/display/RIAK/Riak)and their comparisons to others to understand the tradeoffs in approaches.