Abstract:

This tutorial will give a brief introduction of the emerging ATM technology. Then the problem of applying ATM technology in designing distributed industrial plant control applications with integrated real-time communication requirements will be considered. In particular, it first examines (1) the design space using current ATM service categories and their corresponding traffic models, and (2) the mathematical and practical implications of the traffic models. It then presents a taxonomy of the cell scheduling algorithms available to implement the service categories that can provide performance guarantees in the ATM layer. Through this examination, the gap between what is usually assumed in research and what the actual ATM network will provide will be made clear. It next shows how real-time communication can be mapped onto ATM service categories, and demonstrates the limitations of ATM services. Schemes which can be employed to overcome the limitations and some current research issues will be discussed. Finally, issues in supporting the IETF QoS guarantee standards RSVP and Integrated Services over ATM will be briefly presented. A comparison between the Internet style QoS guarantee and the QoS guarantee required by distributed hard real-time systems will be drawn.