Exploiting multi-agent scheme for traffic management in ATM networks

The experience with the application of a multi-agent scheme to traffic management in high-speed asynchronous transfer mode (ATM) networks is reported. The scheme consists of a collection of problem-solving agents that autonomously execute and coordinately cooperate each other to achieve the ATM traffic management for effectively service access. To possess the features of the parallelism, community and cooperation, three key components are built within the scheme including: knowledge source, blackboard system, and control mechanism. It was implemented based on the Java expert system shell. We treat the application in an ATM testbed and compare the performance in terms of the throughput and data loss. The results show that the scheme is possible to improve the flexibility of an ATM-based operations that support multiple strategies for traffic management