This paper describes a class of partitioning networks, called banyans, whose cost function grows more slowly than that of the crossbar and whose fan-out requirements are independent of network size. Such networks can economically partition the resources of large modular systems into a wide variety of subsystems. Any possible partition can be realized by paralleling several networks or by multiplexing a single network in a manner to be described later. Results will be given indicating that a cost/performance advantage over the crossbar can be obtained for large systems and that the crossbar can, in fact, be considered a non-optimal special case of a banyan network. Inherent fail-soft capability and the existence of rapid control algorithms which can be largely performed by distributed logic within the network are also important attributes of banyans.
This paper presents fundamental properties and preliminary simulation results of banyan partitioning networks. A more detailed treatment, including proofs of theoretical properties, is reserved for reference (5).
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.