Failsafe End-to-End Protocols in Computer Networks with Changing Topology

E ND-TO-END communication is clearly the object of any communication network. Naturally, therefore, end-to-end protocols have been the subject of research for a long time. While many protocols achieve their goals by using timeouts [l] , [2]. Finn, in an original paper [3], demonstrated the existence of end-to-end protocols without making use of timeouts. Essentially, a failsafe end-to-end protocol is one that should deliver, all packets in the correct order and without duplication. In addition, flow control‘and error recovery must also be addressed. In his paper, Finn presented a ptotocol which properly handlCd .ordering and duplication avoidance in a network,in which a resynch procedure operates. In this paper, we extend that protocol, to. accommodate windows, i.e:, allow mores than one packet to be in transit between source. and destination nodes, allow to control the flow, and hadale erroneous packets. All this is done while using bounded counters. The paper is structured in the ,following way. We first introduce a model for a network with changing topology,, the concept of resynch procedures, and the assumptions .underlying the operation of end-to-edd protocols. These are based on previous work [4], [3]. A windowed end-to-end protocol is then presented and its correctness proven. The use of windows for flow, control is also included, as is the special case of circuit-based networks.