Fault tolerance logical network properties of irregular graphs

  • Authors:
  • Christophe Cérin;Camille Coti;Michel Koskas

  • Affiliations:
  • LIPN UMR CNRS 7030, Université Paris 13 - PRES Sorbonne Paris Cité, Villetaneuse, France;LIPN UMR CNRS 7030, Université Paris 13 - PRES Sorbonne Paris Cité, Villetaneuse, France;Département de mathématique, UMR 518 (MIA. INRA), Institut National de la Recherche en Agronomie, Paris Cedex 05, France

  • Venue:
  • ICA3PP'12 Proceedings of the 12th international conference on Algorithms and Architectures for Parallel Processing - Volume Part I
  • Year:
  • 2012

Quantified Score

Hi-index 0.00

Visualization

Abstract

Assume a desktop grid middleware or a deployed cloud infrastructure that are both based on a large number of volunteers for computational-intensive applications or business applications. In this case, the Internet is the communication layer; hence, the communication graph is not regular. Scalability and fault tolerance issues are implicitly present on any platform. For instance, the overlay network that must be built to control the application as part of the run-time support system needs to be scalable and fault tolerant. In this paper, we compute fault tolerance properties of large, irregular graphs that may be used as models for the Internet. In a previous work, we presented algorithms and a framework for computing fault tolerance properties of different variants of randomly-generated binomial regular graphs (BMG). In the present paper we compute the metrics of the Node and Link connectivities and the fault diameter for four benchmarks. We also compare our implementation of the diameter computation with the work of Magnien et al.