With the work on Mediawiki I've done recently, I was wondering about the characteristics of web-based applications such as Mediawiki.
What I see there is a lot of reads and few writes. Looking at other types of web applications, we can see the same thing. Think about reservation systems. Someone goes on the web and does a search on flights and hotels. What comes back is a list of possible flights and a list of possible hotels. Out of the list of flights and hotels, maybe one flight and one hotel will be selected. I say maybe because many people may only be browsing. If a selection of possible 10 hotels is returned, that means that there is a 10 to 1 ratio of reads vs. writes. The ratio is most likely higher.
Web-based applications appear to be the perfect candidates to take advantage of the IDS continuous availability features. For a fast growing demand, in minutes, you can add additional machines to share the same database. The benefits are quick access to scalability, quick recovery in the case of a node failure, and disk savings compare to on-site replication.
If you have a successful web based application, this is something to look into.