Simulation of Infiniband Networks and Effciency Calculation on Non-blocking Fully Populated Fat-Tree Topology

In this paper, we have proposed efficient algorithms for simulating Infiniband networks using a non-blocking fully populated Fat-Tree topology. These algorithms successfully model Fat-Tree Infiniband networks and introduce an efficient routing scheme for multi-source multi-destination routing. Our aim is to use maximum/minimum portion of the topology for each instance of routing and calculate efficiency of network used. Initially we find all possible routes for each source-destination pair. Then, one path is selected per pair such that the constraint of minimum/maximum network usage is satisfied. After final paths have been obtained for all pairs, the network efficiency is calculated. We have considered only latency of links assuming that switches in the network do not have any latency. Hence, this paper includes our algorithms for modelling, routing and performance evaluation of Infiniband networks for a Fat-Tree topology.

[1]  Charles L. Seitz,et al.  Myrinet: A Gigabit-per-Second Local Area Network , 1995, IEEE Micro.

[2]  P.M. Melliar-Smith,et al.  Recursively scalable fat-trees as interconnection networks , 1994, Proceeding of 13th IEEE Annual International Phoenix Conference on Computers and Communications.

[3]  Laxmikant V. Kalé,et al.  Scaling all-to-all multicast on fat-tree networks , 2004, Proceedings. Tenth International Conference on Parallel and Distributed Systems, 2004. ICPADS 2004..

[4]  Charles E. Leiserson,et al.  Fat-trees: Universal networks for hardware-efficient supercomputing , 1985, IEEE Transactions on Computers.

[5]  Robert W. Horst TNet: A Reliable System Area Network , 1995, IEEE Micro.

[6]  Robert W. Horst,et al.  A flexible ServerNet-based fault-tolerant architecture , 1995, Twenty-Fifth International Symposium on Fault-Tolerant Computing. Digest of Papers.

[7]  Leonid Oliker,et al.  Communication Requirements and Interconnect Optimization for High-End Scientific Applications , 2007, IEEE Transactions on Parallel and Distributed Systems.

[8]  Fabrizio Petrini,et al.  k-ary n-trees: high performance networks for massively parallel architectures , 1997, Proceedings 11th International Parallel Processing Symposium.