SNA route generation using traffic patterns

Abstract

This paper describes a procedure used by the IBM Information Network to generate optimum routes for a complex Systems Network Architecture (SNA) network by utilizing communication traffic patterns. The Route Table Generator and an understanding of customer locations and available facilities had been the basis for route generation. However, this approach became overwhelming as the network grew. The lack of flexibility required an increasing need to manually override generated routes. The resulting approach could not ensure that network delay had been minimized. The Network Design and Analysis (NETDA) tool developed at the IBM Yorktown Research Center was used as a solution. NETDA orders routes based on static indicators such as number of hops, route distance, and speed of the path components. However, NETDA also selects optimal routes based on network traffic patterns. Traffic data were easily incorporated into NETDA, and the IBM Information Network has optimized its SNA routing using NETDA and actual traffic data. The process was challenging because of the number of network components involved and the difficulty in obtaining portions of the traffic data. The use of NETDA for route generation is discussed, and the data collection methodology is described. Network component utilization and network delay changes are reviewed as a means of showing the benefits of such optimizations.