This paper describes a distributed computing system composed of multiple microprocessors. Each processor runs a single process to completion without context switching or interrupts. The processors exchange messages over an unusual transport system that allows messages to be directed to a group of like processors so as to attain reliability.

The emphasis in the paper is on the transport system, a variant of a ring network that combines token passing and a bus structure. It includes a temporary storage for messages that cannot be delivered immediately because all destination processors are busy.

No indication is provided as to the type of applications suited to this kind of distributed architecture. Although the transport scheme appears unique, no argument is given for its advantage over conventional microprocessor buses.