Performance evaluation of group communication architectures in large scale systems using MPI

  • Authors:

  • Kayhan Erciyes;Orhan Dagdeviren;Reşat Ümit Payli

  • Affiliations:

  • Computer Eng Dept., Izmir Institute of Technology, Urla, Izmir, Turkey;Computer Eng Dept., Izmir Institute of Technology, Urla, Izmir, Turkey;Computational Fluid Dynamics Laboratory, Purdue School of Engineering and Technology, Indiana University-Purdue University, Indianapolis, Indiana

  • Venue:
  • ODBASE'06/OTM'06 Proceedings of the 2006 Confederated international conference on On the Move to Meaningful Internet Systems: CoopIS, DOA, GADA, and ODBASE - Volume Part II
  • Year:
  • 2006

Quantified Score

Hi-index 0.00

Visualization

Abstract

Group communication is an important paradigm for fault tolerance in large scale systems We describe various group architectures as pipelined, hierarchical, daisy and hypercube groups each consisting of separate clusters, investigate the theoretical performance bounds of these architectures and evaluate their experimental performances using MPI group communication primitives We first derive time bounds for multicast message deliveries in these architectures and then provide tests to measure the times taken for the same operation The multicast message delivery times are tested against the number of clusters within a group and the size of the multicast message We conclude that daisy architecture is favorable both in terms of delivery times and message sizes theoretically and experimentally.