Audience: Networkers
Technical Level: Deep Geek

The session will deal in transaction clustering primarily for serving web pages. I will discuss the initial stages of what building a cluster is supposed to accomplish and how to determine organizational need. We will look at deciding on an operating system, and what packages are available to ease the transition. I will explain in broad terms, using the Linux Virtual Server project as an example, how to design and implement a cluster. I will go into detail about load balancing, redundancy and fault tolerance. We will discuss the administrative overhead associated with maintaining data in parity on the cluster and give examples on how this might be accomplished. I will give live demo of a functioning cluster showing how it load balances and imitate a failure to show redundancy. I will briefly describe how other services might be applied to the cluster model. There will be time at the end to take questions... or eat pie.

MP3: NagyLR3Th400.mp3

Play It Now!

Watch the presentation.


Daniel Nagy
Sys Admin, Legal Information Institute
Cornell Law School