Group routing without group routing tables: an exercise in protocol design

We present a group routing protocol for a network of processes. The task of the protocol is to route data messages to each member of a process group. To this end, a tree of processes is constructed in the network, ensuring that each group member is included in the tree. To build this tree, the group routing protocol relies upon the local unicast routing tables of each process. Thus, group routing is accomplished by composing two protocols: an underlying unicast routing protocol, whose detailed behavior is unknown but its basic properties are given, and a protocol that builds a group tree based upon the unicast routing tables. The group routing protocol is developed in three steps. First, a simple protocol is obtained, and is proven correct. Then, the protocol is refined twice. Each refined protocol improves upon its predecessor by satisfying all of the predecessor's properties plus some additional stronger properties. The final protocol has the property of adapting the group tree to changes in the unicast routing tables without compromising the integrity of the group tree, even in the presence of unicast routing loops.