[htdig] using htdig/mifluz for mid-sized database search

[<alo...@ya...>]

Hi

 I am planning to use htdig/mifluz to search a
database of documents. The size is about 1 million
records, each of 4K. The search is multi-dimensional
(4-8 search criteria). I am thinking of using "system
generated keywords" (so if A,B,C are possible drop
down values, then use SYS_A, SYS_B, SYS_C as system
keywords put into a "generated" document in addition
to all the original document contents) to mimic
multi-dimensional search. Currently we use MySQL on
Linux and use SQL queries but
1. The speed is beginning to suffer as the database
grows (2000 updates a day, 1000 new records)
2. It doesn't allow mix of structured and full text
search

 I wanted to know if using htdig/mifluz in the above
manner is a good idea. We do not have a lot of
hardware -- in fact, currently the entire application
runs on a dual cpu PIII with 1GB ram. We need response
times of upto about 5 seconds, doing about 100,000
searches a month. The queries will have to be
generalized boolean, and with all fields put together,
might have 6-10 individual search terms (connected
with AND and OR)

 Any other suggestions are also welcome.

Thanks!
Alok

________________________________________________________________________
Want to sell your car? advertise on Yahoo Autos Classifieds. It's Free!!
       visit http://in.autos.yahoo.com