Most commercial information systems are built to support heavy volumes of transactions
by many users on a daily basis. Examples include banking, insurance, and order
processing systems. These Online Transaction Processing (OLTP) systems typically
require quick throughput for their largely predefined range of transactions, especially
update transactions. To improve performance, historical data is often archived once it
reaches a certain age, reducing the size of the data sets used for daily operations. A
single organization may have several OLTP systems (e.g., purchasing, sales, inventory,
customer records), possibly implemented using different kinds of DBMS or other
software applications, and the coupling between such systems may be weak or even
nonexistent.