A scalable database architecture for network services

In this paper we describe how databases to support network services can be viewed in terms of a single integrated model and accessed via a common, application independent interface. We discuss the performance limitations of conventional database technology in realizing this model and propose a novel alternative: the Datacycle architecture. By employing high bandwidth fiber optic channels to distribute database contents past processing elements in the network the Datacycle approach both supports a general highly functional interface to the contents of the database and scales to support very high transaction volumes. Based on the performance of a working laboratory prototype, we estimate the hardware complexity required to implement a common database for several network services. We conclude that the Datacycle architecture is a feasible technology to realize an integrated, network-wide database and can scale to support database access on every call attempt, thus making possible a broad array of new communications capabilities in the public network.