I concur the data loader is a fine way to look at loading massive amounts of data into a Salesforce org. This is dependant on external IDs though, and the data being formatted into reasonably correct CSV files.

You may wish to look at a third party data loading tool, Talend is a great - fully customisable (free) ETL tool, based on the Eclipse IDE, and can produce easily repeatable jobs, in the form of automatically generated java web services, and even offer the option of hosting and running them on a schedule for you. (i should say that it is just as easy to make a one off dataload via Talend, and it gives you awesome powers like drag and drop field mapping - loading related objects and using them too (ie. you can set lookup fields just from the related record name or any other SOQL-available field) it also reads literally any source, from CSVs to web services or even an MS access database.

Another tool (paid for) is called Jitterbit, this again has a drag and drop interface for building mappings between one data source (databases, CSVs, web services, you name it) and pumping them into your Salesforce org. it even watches out for batch size for you to make sure no single request breaks an obvious goveror limit),

One thing to always remember when starting out on this road is how much data-space you have in the organisation? This can be found under company info. Typically a record of any single object  (no matter how big, small or number of fields) takes up 2Kb (the main exception to this being if records have long-text fields) - so you can do some quick forward-calculations to make sure you have space.