[K12OSN] Server/Client Performance

Petre Scheie petre at maltzen.net
Tue Nov 30 13:59:19 UTC 2004


I started out with a configuration similar to what you mention. It's cutting it 
close but it's probably better than using that same hardware without 
implementing LTSP.  It's good for a proof of concept test, verifying that the 
clients boot and can bring up the applications.  And if you limit the clients & 
server to using, say, IceWM for the desktop (without Nautilus, so no desktop 
icons, but there can be icons on the IceWM toolbar), AbiWord for the word 
processor, and Firefox (not Mozilla) (or better yet, Dillo) for the web browser, 
I think the performance would be tolerable.  Yes, the performance on the clients 
would be the same as on the server, more or less.  So, try loading up AW and FF 
under IceWM on the server and see how it is.  If you at least set it up this 
way, then you'll be in a good position if someone else offers to donate an old 
400mhz machine, because you could dedicate that to running just one application, 
say Firefox, but everything else and all the machines would benefit.

Petre

Mark Cockrell wrote:
> Hello all,
>    I'm hoping to set up a very small K12LTSP deployment for a local 
> church in the near future using donated (trashed, actually) hardware and 
> I have a question regarding performance.  I've got a pretty low-end PC 
> here that I want to use as the server: a P-233 with 256MB RAM.  What 
> kind of performance can I expect for 2 clients connected to this machine 
> for use as a Web Browsing/Word Processing cluster?  Is it a safe 
> assumption that the clients would perform these tasks about as well as 
> the server itself?
> 




More information about the K12OSN mailing list