Horms (Simon Horman)
To be presented at
The 1998 Atlanta Linux Showcase
Atlanta, Georgia, USA
Friday 23rd - Saturday 24th October 1998
``A distributed system is one in which I cannot get something done because a machine I've never heard of is down.''
I would like to acknowledge my employer Zip World www.zipworld.com.au for their assistance and patience that enabled this presentation to come together. I would also like to thank Computer Generation Incorporated www.compgen.com for sponsoring my trip to the 1998 Atlanta Linux Showcase to present this paper. Finally this paper would not have been possible without the support of my friends. In particular Little Miss Fun, Mandrake, Raster, Mr. O'Brien, Gus, Miss Kim and K have helped to bring this work to completion.
This paper is freely distributable under the terms of the Gnu Public License.
Redundancy of all forms is an important issue when administering servers. One aspect of this which is of particular interest is having highly available content. If your HTTP server dies your site shouldn't go down with it. This area is particularly challenging as solutions must provide for both redundancy and high performance.
One way to achieve high content availability is to house the content on a file server enabling you to hot-swap machines such as FTP servers using methods such as ARP spoofing. This solution poses interesting performance issues but the file server is still a single point of failure.
Regardless of whether the content is housed on a separate server or not, providing duplicate file servers is also important and synchronisation between servers can be implemented in various ways. Again performance is a major issue particularly during any synchronisation procedures.