Pros

• CPAN available
• Implement in a short period of time if you have existing databases
• Keeps search information inside the same container as the data

Cons

• Requires knowledge/familiarity with DBI, MySQL, and Perl data strucutures. Which may make it difficult to use for newer users of Perl.
• Only works with MySQL
• Could impact database performance if search is perfromed on the same database as the data in heavily loaded systems.

Performance Results

There is really no reason for me to do this at the scale of the application this was applied to. There is no noticable latency even when the search is performed over 100Mbps connected servers.

Worthwhile

The exercise worthwhile for several reasons:

• Reused existing code
• Provided results that were much more accurate
• Scripts produced will allow for easier future implementations on other projects
• Provided support for search modifiers (AND, OR, +)

Where to use it (will it scale)

• Since I have not done any performance tests I can't possibly predict how far this would scale, but speaking from my own install with roughly 5MB of searched content the seat of the pants performance is acceptable

The system has been online now for 6 months and it has satisified the clients search requirements as well as aided site visitors in locating the information of interest to them. The solution is implemented on a mod_perl enabled Apache server with ~30,000 visitors per month at peak.