Now What Do We Do?
Randall S. Hansen and R. Kevin Stone
AT&T Labs, Middletown, New Jersey USA
Representing and providing ready access to large amounts of complex data has always been a difficult design issue for those who design the data structures and their representations in computing systems. Nowhere is this issue more apparent than in the telecommunications industry, where each call generates its own trace in the network and in legacy systems in which storage space is always a consideration of paramount importance. Each of these calls results in the generation of a Call Detail Record (CDR) containing information about the call. The AT&T network processes over 250 million calls -- and their associated CDRs -- on an average business day. The CDR information is used not only for customer billing, but also for network planning, engineering, marketing, and for security investigations. CDRs are the key to providing information about customer network usage. Because of their business value, AT&T stores each CDR for two years in a single database. In the near future, this database will contain approximately 200 billion call records.
Recording Applications Manager (RAM) is a system for monitoring and managing the call detail recording process (for a more complete description of RAM and the user-centered design method that produced it, see Somberg, 1999). RAM provides the user with an integrated view of the various data sources and systems involved in call detail collection, management, and use. It provides a consistent way of interacting with those data sources and systems, and it places common functions on a platform that can be used by all data