[Web4lib] high availability web service infrastructure: your experiences?

[Tom Wood]

[Apologies to those who received this via LITA-L a few weeks ago.]

We are embarking on a redesign of our Library IT infrastructure, and I'm
looking into our web site/service architecture.  My redesign goals
include:

 * Improving the availability of web delivered services

 * Introducing a development -> staging -> production flow for content
and applications

 * Providing a flexible architecture that can accommodate new (and
unanticipated!) projects

 * Doing this all while spending as little money as possible :-)

While I'm an "old hand" at software, I am but a novice in the library
world and would benefit greatly from your experiences.

So, I would like to tap into the community's experience in building
*high availability* web systems for libraries.  I realize that is a bit
of a broad brush inquiry, so here are some specific questions that may
give you a better idea of what I'm after, specifically about high
availability systems:

  * What sort of high availability architecture do you use?  (e.g.,
clustering, load balancing, etc.)  What is the scope of your
architecture? (e.g., web server, application server, database, etc.)

  * If clustering or load balancing, how do you distribute your web
content to your servers? (e.g., replication/mirroring, network file
system, global file system, etc.)

  * How do you move your content/applications through
development/staging/production?  (e.g., home grown tools, version
control system, content management system, etc.)

  * Do you use a content management system in a high availability
configuration?  Any quirks, words of wisdom, etc.?

Thanks very much!


Tom Wood
thomas.a.wood at uconn.edu
ITS Applications Developer
University of Connecticut Libraries