We have 10 denormalized tables that contains large amounts of data (11,000,000) in many of them. The client wants the user to be able to download up to 100K of records at a time. What I am wondering is the design of a background process to handle this. I was thinking of possibly setting up a queue for each of the tables and having multiple threads handling the requests. I'm definetly not an expert on designing multi-threaded components. Do you guys think I'm on the right track here? Any suggestions on implementation?
Thanks!