Customer Name Matching

[Raghavendra]

We have a requirement to search a database for a customer based on Name and Address fields. The search happens in real time from a web service.

We have customer database with 10 million rows (Name and Address are not standardized) and it can be updated in real time with new customers.

When we invoke a search service, it has to read free format Name and Address fields and search in our customer database and pass the customer ids.

The solution I am thinking is, read the customer data (from database) and standardize it and store in a temp table. Add real time updates also to this temp table and keep the standardized data available for the search.

When search service invoked, read the free format text and standardize it and search in the temp table to pass the customer id. If we do not find any match just send a flag saying that match not found.

Please let me know if this approach is fine and also advise me to tune the approach for better performance. I feel the time taken to search the temp table would be very huge and need your inputs in tuning the performance here.

Is there a way where we can store the data internally within QualityStage to improve the performance.