Estimating the Connection Probability for 2-D Mesh Networks by Monte Carlo Methods

Fault tolerance is a central issue in the design and implementation of interconnection networks for large parallel systems. Connection probability of a network is a good network fault tolerance measure. For a mesh of given size and node failure probability, the gap between the known upper and lower bounds on the connection probability is often very large. In this paper we design algorithms to estimate the connection probability for 2-D meshes by Monte Carlo methods. The experiment is carefully designed and performed, and the simulation results give good estimates of the connection probability for 2-D meshes and can be used to evaluate the known lower and upper bounds on connection probability for 2-D meshes.

[1]  Christian P. Robert,et al.  Monte Carlo Statistical Methods , 2005, Springer Texts in Statistics.

[2]  Ajay K. Royyuru,et al.  Blue Gene: A vision for protein science using a petaflop supercomputer , 2001, IBM Syst. J..

[3]  Guojun Wang,et al.  Probabilistic analysis on mesh network fault tolerance , 2007, J. Parallel Distributed Comput..

[4]  Xiaodong Xu,et al.  Lower Bounds on the Connectivity Probability for 2-D Mesh Networks , 2009, 2009 5th International Conference on Wireless Communications, Networking and Mobile Computing.

[5]  Dhiraj K. Pradhan,et al.  Dynamically Restructurable Fault-Tolerant Processor Network Architectures , 1985, IEEE Transactions on Computers.

[6]  Werner Butscher.  The Dataparallel Computer MasPar MP - 1 , 1991, Supercomputer.

[7]  Ricardo Bianchini,et al.  The MIT Alewife machine: architecture and performance , 1995, Proceedings 22nd Annual International Symposium on Computer Architecture.

[8]  Jie Wu,et al.  Fault Tolerance Measures for m-Ary n-Dimensional Hypercubes Based on Forbidden Faulty Sets , 1998, IEEE Trans. Computers.

[9]  Jerry L. Trahan,et al.  Improved Lower Bounds on the Reliability of Hypercube Architectures , 1994, IEEE Trans. Parallel Distributed Syst..

[10]  Anoop Gupta,et al.  The Stanford Dash multiprocessor , 1992, Computer.

[11]  Jianer Chen,et al.  Locally Subcube-Connected Hypercube Networks: Theoretical Analysis and Experimental Results , 2002, IEEE Trans. Computers.

[12]  Yi Pan,et al.  Practical Deadlock-Free Fault-Tolerant Routing in Meshes Based on the Planar Network Fault Model , 2009, IEEE Transactions on Computers.

[13]  Shahram Latifi,et al.  Conditional Connectivity Measures for Large Multiprocessor Systems , 1994, IEEE Trans. Computers.