Availability and performability assessment in LAN interconnection

Topics related to the service quality of a network made up of interconnected LANs are discussed. As the size of a network increases, so do reconfiguration times, thus reducing service availability, increasing frame loss, and leading to performance degradation. This degradation may be substantial if circumstances requiring reconfiguration occur frequently. Two types of networks are compared by assessing the parameters which directly affect the quality of service of the network. The first uses media access control (MAC) bridges based on the standard spanning tree algorithm; the second uses parallel MAC bridges, by means of which the network can be organized into islands, thus making reconfiguration selective. In both cases the bridged LAN is modeled according to the schemes suggested by systems reliability theory, which allows assessment of performance parameters directly affecting service quality, such as the mean time to failure (MTTF), performability, and availability.<<ETX>>