Zamm
Data warehousing is more of a concept that a type of application. In broad terms, you load archived data into a separate database, often used for reporting or analytical processing. The difference is that many of the table relationships are 'de-normalized', ie you might store a customer's main address in the customer table, instead of in a normalized way, where the address would be in an Address table, where the customer might have many addresses. In addition, many totals fields, ie. are added to tables as 'redundant' fields, ie Order total wouldn't be calculated by reading all detail records, it would have it's own field in the order header. Since the data is static, these types of redundancies are always accurate and you don't need all sorts of triggers and update programs to keep them in sync as you would in an live transaction processing database.
This serves to make the reports run faster, since fewer tables are read to get the data needed, and also it allows management to have an easier time writing their own reports, since they don't need to know all the intricacies of joining tables.
This is only a very surface level look at it, but there is a Data Warehousing forum here.
Oh, and Access can be used for data warehousing, in fact I'm using it as an interem solution for a client right now--with about 10 .mdb files all stretched to close to the 1gig max, some containing just a single table. (This until the client waits for Oracle to drop prices!)
--Jim