Performance evaluation of the distributed queue dual bus metropolitan area network.

The DQDB network has been adopted by the IEEE Standard Committee as the subnetwork for the IEEE 802.6 MAN. Performance evaluation of the network is crucial in the network design phase since it allows us to predict the performance of various design alternatives, such as the ordering of stations within the network. Only a limited amount of work has been reported on the performance evaluation of the DQDB network. The main difficulty in analyzing the DQDB network resides in the high degree of interdependence among the various processes that make up the operation of the network. In our research, we develop two approximate analytical models which accurately capture the performance of the network. In the first model, we consider a network where the most upstream station is saturated and all other stations are stable. Each station is modeled as an M/G/1 queueing system. Since the most upstream station is saturated, the service time for a packet has a special structure, which enables us to decompose it into five independent components. The mean and variance of each component in the service time are computed in four different modes of operations. Simulation experiments show that our model is very accurate under our modeling assumptions. Moreover, this model provides insights into how the interactions between data stations in the network can be described. In the second model (the single-station model), we consider a network in which all stations are stable. Each station is modeled as a two-queue single-server system which serves data packets and requests from downstream stations according to the DQDB protocol. We estimate the mean waiting time of the requests and, in turn, using the discrete time work conservation law, estimate the mean access time for packets. Simulation experiments show that the model is highly accurate, even under very high traffic loads.PhDIndustrial and Operations EngineeringUniversity of Michigan, Horace H. Rackham School of Graduate Studieshttp://deepblue.lib.umich.edu/bitstream/2027.42/103126/1/9303797.pdfDescription of 9303797.pdf : Restricted to UM users only