Article ID: | iaor19991027 |
Country: | United States |
Volume: | 46 |
Issue: | 4 |
Start Page Number: | 515 |
End Page Number: | 531 |
Publication Date: | Jul 1998 |
Journal: | Operations Research |
Authors: | Lee Thomas (Yew Sing) |
Keywords: | queues: theory |
The evolution of communication technology and the competitive software design environment dictates that performance characterization involves iterations between real-world design and its representation for modeling purposes. In this paper, models and analytical techniques are developed to evaluate the performance of single-server polling computer and communication systems, e.g., multiplexers or local area networks with a single token passing medium access protocol. We present a novel methodology for the design and control of such communication systems and computer systems. Our approach allows the handling of smooth, correlated, and bursty traffic, whereas traditional analytical approaches have depended on a Poisson traffic model. The methodology also provides a unified treatment of several existing methodologies for controlling single-server polling computer and communication systems.