Multistage interconnection networks (MINs) can be designed for parallel computing and switching fabrics. Multistage sorting networks can be one of the many applications using interconnection networks. In this monograph, the generalized odd-even merge design is discussed in detail. Designing switching network is another important application of the multistage networks. Generally, interconnection network can be blocking, rearrangeably nonblocking, or nonblocking, or with variable connecting capabilities, from rearrangeable for permutation to nonblocking for multicast services. Novel multistage switching fabric designs, the PINIUM and PINIUM+ switches, for packet switching and multicasting are discussed. A novel switching framework of high-performance hardware schedulers for various buffering techniques based on interconnection networks are discussed.

This three-volume work presents a compendium of current and seminal papers on parallel/distributed processing offered at the 22nd International Conference on Parallel Processing, held August 16-20, 1993 in Chicago, Illinois. Topics include processor architectures; mapping algorithms to parallel systems, performance evaluations; fault diagnosis, recovery, and tolerance; cube networks; portable software; synchronization; compilers; hypercube computing; and image processing and graphics. Computer professionals in parallel processing, distributed systems, and software engineering will find this book essential to their complete computer reference library.

In the existing research literature a large number of regular MINs with their performance characteristics are available. However, there is relatively smaller number of irregular MINs. The present research work is an endeavour to study the importance and design of fault-tolerant MINs. It is with the use of various techniques based on high degree of parallelism, self-routing and modularity that three new dynamic irregular MINs have been proposed in order to improve their fault-tolerance, performance and reliability. The irregular networks have been considered because of their advantage of the shorter path length in comparison to their regular counterparts. The irregular dynamic MIN supports many redundant paths in the event of faults in order to dynamically reroute the data through a fault-free path and passes most of the permutations even under faults while improving the Mean-time to failure (MTTF).