Database Design for speed question

Hi all,

I'm thinking of a possible redesign of a database to increase speed of transactions. Any musings or corrections would be welcomed.

The database is feb by the Web. Hundreds of sessions are kept active and idling to help get data in and out as quickly as possible. Java opens and controls these sessions. Data is customer specific and the customer must be verified before query or input is performed. Currently the data is essentially in a couple of big tables, and the tables are getting bigger and bigger. To reduce the time to find and input specific customer data, I'm thinking along the following lines:

Create, say, 50 tables that are identical in structure to hold the customer data. Create a small 'lookup' table to point to which table a particular customer's data is in. As new customers come on-line, they are methodically put into the next available table. When table 50 is reached, is starts over at table 1. Pin the small 'lookup' table in RAM for speed (Is that possible?). Has anyone done this? Does Oracle provide this with some other method already?

Thanks for your time! Received on Wed Nov 17 1999 - 13:55:46 CST