Concept for archiving huge database

I need a concept how I can archive all data in D-Pool which is older than one year. At the moment we have more than 3 million records in the D-Pool. Because of this huge data foundation searches and filters over the database takes quite too long, because most searches are done over the whole D-Pool data, but in most cases I am only interested in current data. So I want to archive regularly all data from D-pool which is not needed for current work and evaluations. But for some functions it should be possible furthermore to do a search over the whole D-pool, the current and the old data.

Could you tell me some ideas for this probl开发者_开发问答em?

This describes the typical data warehousing solution. Most owners of large datasets that change daily have a transactional database and a historical or reporting database. The historical or reporting database allows users to mine for insight against everything - except the data added since the last extract. That's usually sufficient.