Abstract

This article presents, by analysis and measurement, the probability density functions (pdfs) for the latency from origination to message delivery for the Totem multicast protocols, in the presence of message loss and token loss. The Totem protocols provide reliable totally ordered delivery of messages across single and multiple local-area networks (LANs), using a logical token-passing ring on each LAN with gateways that forward messages selectively between LANs. The analysis of the pdfs for the latency involves decomposing the latency into independent components and convolving the pdfs of those components. A comparison of the performance of single-ring, two-ring and four-ring networks shows that, with message filtering in the gateways, multiple-ring networks achieve lower mean latency, less variability, and shorter tails of the latency distribution than an equivalent single ring. The experimental measurements of the pdfs for the latency are similar to those obtained from the analytical formulas, particularly for values of the latency with high probabilities. Limitations of the analytical model, and insight into communication and scheduling of the operating system, are discussed.

Author Keywords: Multicast protocols; Performance modeling and analysis; Message delivery latency; Probability density functions; Latency measurements; Clock skew elimination; Density estimation

^*Corresponding author. E-mail: thomo@alpha.ece.ucsb.edu.

¹E-mail: moser@ece.ucsb.edu.

²E-mail: pmms@ece.ucsb.edu.