A Statistical Analysis of ARQ Protocols Operating in a Nonindependent Error Environment

In this paper the behavior of the Stop and Wait and Go Back- N error detection and retransmission (ARQ) protocols in an environment characterized by nonrandom errors are studied. Analytic models are developed for the case that the error process is modeled as a Markovian process. These models may be used to predict performance measures such as expected queue length and expected delay for the case of the Stop and Wait protocol and the Go Back- N protocol when N is very large. The models can also be used to determine maximum throughputs for both protocols. The results of these models are compared with simulation results for the Selective Repeat ARQ protocol.