Reliability and Performance Modelling of Hypercube-Based Multiprocessors

摘要

New technologies of integration now enable the design of computing systems with dozens and even hundreds of independent homogeneous Processing Elements which can cooperate on the solution of the same problem for a corresponding improvement in the execution time. However, as the number of Processing Units increases, concerns for reliability and continued operation of the system in the face of failures must be addressed. A commonly used network topology, the hypercube, is analyzed in this paper. It is shown how the disconnection of a subset of the machine would yield a total system failure. Overall reliability measures as well as computational availability, etc., are derived by a combination of analytical modelling and simulation approaches (Monte Carlo simulation). Finally, a comparison with other topologies (mesh connected network, etc.) shows the influence of the connectivity of the network on the reliability of the system.