Fast BGP convergence following link/router failure

  • Authors:
  • Swapan Kumar Ray;Susmit Shannigrahi

  • Affiliations:
  • Computer Science and Engineering Department, Jadavpur University, Kolkata, India;Computer Science and Engineering Department, Jadavpur University, Kolkata, India

  • Venue:
  • ICDCN'10 Proceedings of the 11th international conference on Distributed computing and networking
  • Year:
  • 2010

Quantified Score

Hi-index 0.00

Visualization

Abstract

A Modified Border Gateway Protocol (MBGP) has been proposed towards achieving faster BGP convergence in the Internet following link/router/ network failures. MBGP adopts the overall strategy of distributed fault detectioncum-identification, fault notification and rediscovery-cum-readvertisement of valid routes. In the assumed simplified model of the Internet, the sole MBGP router in each autonomous system (AS) identifies any failed component using the novel concept of special neighbors and notifies the identity of the failed component to all the MBGP routers in the Internet. Six new messages, including a query-response message pair and four permanent withdrawal messages, have been proposed in MBGP, without changing the BGP message format. The path exploration problem is significantly reduced because some failures cause no path exploration, the others do but only in a small number of nearby routers and, finally, no invalid messages are ever exchanged. Simulation studies have demonstrated significantly faster convergence of MBGP over BGP.