A performance model of layered communication architecture

This paper proposes an analytical model for a layered communication architecture to aim at predicting the performance of protocol software implementation. Each layer is modeled by a queueing network. The model is constructed with focusing the behavior of protocol entity in a layer. The authors employ an FCFS server queue with pre-emptive priority for entity and an infinite server queue for response time of acknowledge packets. Respective layer models are piled up by embedding the lower services in the infinite server queue. From the perspective of end-to-end preformance, the present model provides a solution for the following problems which come up in implementing a layered communication protocol: 1) to which processor should each layer’s process be assigned?, 2) which priority should be given to respective processes in each processor? and 3) which layer should provide an error recovery service? The authors show an approximate analysis to obtain an average end-to-end transmission delay and discuss the possible solutions for the problems mentioned above. [In Japanese.]