REAL-TIME TRAFFIC MEASUREMENTS FOR HIGH-SPEED NETWORKS

BT requires timely information about high-speed network performance in order to manage these networks and to provide the best possible service to customers. Network management uses this information to evaluate, in real time, the state-of-network resources. Despite some claims to the contrary, traditional data collection techniques fail to support useful or accurate real time analysis because they either overflow available disk storage ol exceed the real time processing capabilities of embedded processors. As network media speeds increase and as end users communicate across network architectures with differing protocols, the inadequacies of traditional techniques become more pronounced. This paper presents a new information collection architecture (ICA) for real time traffic measurements that solves the data collection and analysis problem for high-speed networks. The ICA is applicable across network architectures and is scalable to SONET/SDH speeds and beyond. Also described is a prototype implementation of the ICA that provides performance measurements for token ring networks. This ICA prototype device includes programmable hardware to detect and collect traffic patterns of interest in real time, and an expert system to compute network performance using the collected data. Performance results computed using ICA prototype device traffic measurements taken from several operational token ring networks are presented and these results are compared with those computed using traffic, measurements from a traditional network traffic monitor and with simulation results. These comparisons demonstrate that the ICA device accomplishes accurate performance measurement for real time network management by collecting only a small fraction of the data required by traditional approaches.