Tunable survivable spanning trees

Jose Yallouz, Ori Rottenstreich, Ariel Orda

Research output: Contribution to journalConference articlepeer-review

Abstract

Coping with network failures has become a major networking challenge. The concept of tunable survivability provides a quantitative measure for specifying any desired level (0%-100%) of survivability, thus offering exibility in the routing choice. Previous works focused on implementing this concept on unicast transmissions. However, vital network information is often broadcasted via spanning trees. Accordingly, in this study, we investigate the application of tunable survivability for efficient maintenance of spanning trees under the presence of failures. We establish efficient algorithmic schemes for optimizing the level of survivability under various QoS requirements. In addition, we derive theoretical bounds on the number of required trees for maximum survivability. Finally, through extensive simulations, we demonstrate the effectiveness of the tunable survivability concept in the construction of spanning trees. Most notably, we show that, typically, negligible reduction in the level of survivability results in major improvement in the QoS performance of the resulting spanning trees.

Original languageEnglish
Pages (from-to)315-327
Number of pages13
JournalPerformance Evaluation Review
Volume42
Issue number1
DOIs
StatePublished - 20 Jun 2014
EventACM SIGMETRICS International Conference on Measurement and Modeling of Computer Systems, SIGMETRICS 2014 - Austin, United States
Duration: 16 Jun 201420 Jun 2014

Keywords

  • Fault-tolerance
  • Reliability
  • Spanning-tree
  • Survivability

All Science Journal Classification (ASJC) codes

  • Software
  • Hardware and Architecture
  • Computer Networks and Communications

Cite this