Performance of FDDI in parallel processing systems

Investigates the use of the Fiber Distributed Data Interface (FDDI) as a solution to the problem of interconnecting processing elements in a homogeneous processing environment. The authors present a detailed description of FDDI and compare the communication performance to that of the following topologies: the spanning multiaccess channel hypercube (SMCH), the generalized hypercube (GHC), the nearest neighbor mesh hypercube (NNMH), the Boolean cube (BC), and the cube connected cycles (CCC). It is concluded that, based on the communication performance of the networks, FDDI can be considered as an inexpensive solution to the problem of interconnecting smaller networks for applications which are computationally extensive