Re: [tlug] [tlug-digest] Large Network Systems

[Tim Hurman <kano-tlug@example.com>]

On Fri, Mar 18, 2005 at 07:32:22AM -0500, Josh Glover wrote:
> On Fri, 18 Mar 2005 16:50:17 +0900, Robert Fisher <robert@example.com> wrote:
> 
> > I have been recently plagued with many questions on building large scale
> > redundant high performance server clusters for the company I am working
> > for.
> 
> I am afraid you may have posed the question poorly, as I am not
> *exactly* sure what you are looking for.
> 
> My last job was sysadmin for a biotech software company, [1] so to me,
> "high performance" means clusters and message-passing systems and
> massively parallel algorithms. If that is what you are after, start
> here for a decent overview:
> 
> http://www.oreilly.com/catalog/hpc2/index.html
> 
> If you just want a scalable webserver platform, that would be a better
> question for Tim Hurman, the BBC's sysadmin. (Tim, you still
> subscribed?)

Hi,

Yes, I am still subscribed, just really busy. Meant to reply sooner, just
forgot.

As Josh implied, there are about million and one ways to make high
availability systems. It usually depends on the application you are
"targeting". Generally some form of load balancing/failover is required
and this can be done by various kit between layer 2 and layer 7. I can
easily tell you how the BBC do it, although it is probably best said off
list as it is quite long and off topic (unless of course people are
curious). If you have some more details, I am sure there are a few people
here who could help you.
Yours,

Tim