Adaptive connection management for mission critical applications over ATM-based networks

This paper focuses on connection management for mission critical real-time applications over ATM-based networks. Traditional connection management generally requires Quality-of-Service (QoS) parameters to be specified as fixed values, and can only provide a QoS that is constant throughout the lifetime of an admitted connection. Such simplistic specification and consequent resource management offer no flexibility to user applications. The applications cannot receive the best possible QoS, and system resources are grossly under-utilized. We take an adaptive approach. With our adaptive connection management, QoS of connections is specified over a range of values. Resources are reallocated and redistributed in response to dynamic fluctuations in resource availability. In addition to providing a QoS adaptation framework, in our design we identify and address two issues which are critical for the success of adaptive connection management: sensitivity to mission specific requirements, and fast and accurate computation of delays. With our adaptive strategy, we demonstrate dramatic improvements in both the QoS offered to applications, and the effective utilization of system resources. Our approach is practical and compatible with current networking standards. We have implemented adaptive connection management in our real-time toolkit, NetEx. NetEx provides delay guaranteed communication services for mission critical real-time applications over high-speed networks.

Adaptive connection management for mission critical applications over ATM-based networks Academic Article