Fast convergence mechanisms and features deployment within operator backbone infrastructures

Abstract

The goal of backbone operators like: Orange/France telecom, Cegetel , AT&T, Sprint, BT…etc, is minimizing impact on customer applications due to failures and interruptions within their backbone. Data backbones based on MPLS /VPN infrastructures are widely deployed over the world, giving connectivity to enterprises and public users with on demand added services. Ideally: customers will see no interruption at all, tending to make backbone failures as transparent as possible to end applications. That is why tuning configuration of routers and routing mechanisms, introducing features that help to minimize interruption durations within the backbone. In this paper we will give an overview of fast convergence mechanisms and features deployment within data network infrastructures. Will be discussed standard features like: “bidirectional failure detection” (BFD) [1] which provides fast failure detection for all media types, encapsulations, topologies, and routing protocols, and how it is the way to go when the physical media does not provide any way of detecting a failure, fast enough to meet fast convergence design requirement, Mpls ldp-igp synchronization [5]: which ensures that LDP is fully established before the IGP path is used for switching. It achieves this by not building IGP neighbor-ship before the LDP [6] neighbor-ship is established, will be discussed also some other relevant features ….