ZeroJunk: How do the credit card companies do it?
Everybody gets fixated on how something can't work. Can't ,can't, can't.
You didn't read my last post. With current technology you can either have a large database or a complex database. The only way you can have a large complex database is with several smaller databases being tied together with some form of software.
The credit card companies do need to have large amounts of data in their databases, but their databases are not that large and complex as they are separated into different databases. The database that does credit authorizations only has very basic information and thus is able to be very large. It likely only has a few fields that relates directly with credit authorizations.
The billing database would be separate from that and divided into several databases, I believe it is by ZIP code. So each ZIP code is a separate, smaller database that is able to have far more details. Using ZIP codes is a good idea because the postal service never allows ZIP codes to grow too large without being split up. So the databases for the CC company never grows too large. The authorization database is able to use the billing ZIP code in its database to send the transaction information to the correct billing database and get an updated remaining balance figure back for future authorizations.
So, in short, the credit card companies don't have a large central database. The software that ties those databases together may make it appear so. But the data remains separate.