Engineering performance for the future of intelligence

Abstract

A Ingham, J Graham and P Hendy Current intelligent network (IN) platforms support high capacity real-time call processing applications, tuned to provide virtually instant service, and with comprehensive network overload control schemes to cope with huge (often media stimulated) traffic spikes. In order to meet these requirements, IN platforms tend to be 'stove-pipes' (platforms dedicated to a specific service). BT's aim is to deploy a zero-touch, self-service platform supporting all services. BT's proposed 21st century architecture has determined that the intelligence space and the systems space shall converge. This convergence poses some serious challenges for the performance engineer.Does the converged intelligence layer need to provide performance equal to, or better than, the current platforms? Or will end users tolerate a degradation, if the price is right? How can stringent performance requirements actually be met in a converged intelligence space, with its rich new service features? This paper discusses the issues, and potential solutions.