Where’s Waldo? practical searches for stability in iBGP

Abstract

Abstract-What does a child's search of a large, complex cartoon for the eponymous character (Waldo) have to do with Internet routing? Network operators also search complex datasets, but Waldo is the least of their worries. Routing oscillation is a much greater concern. Networks can be designed to avoid routing oscillation, but the approaches so far proposed unnecessarily reduce the configuration flexibility. More importantly, apparently minor changes to a configuration can lead to instability. Verification of network stability is therefore an important task, but unlike the child's search, this problem is NP hard. Until now, no practical method was available for large networks. In this paper, we present an efficient algorithm for proving stability of iBGP, or finding the potential oscillatory modes, and demonstrate its efficacy by applying it to the iBGP configuration of a large Tier-2 AS.

Humpty Dumpty: Putting iBGP Back Together Again

“…This constraint is a sufficient condition to prevent persistent oscillation and guarantees a unique solution [8]. However as the condition is not necessary, it can be overly restrictive and not satisfied in practice [10,4]. Removing this assumption removes the benefit of always converging to a unique solution -the timing of BGP updates can determine which of multiple solutions is settled upon.…”

“…The reliance rules in the three-level hierarchy are somewhat more complicated than the two-level case examined in [4]. We first present a brief recap of the notation used in [4] before outlining generalized reliance rules and showing an example of how they apply.…”

“…However using this approach, a significant number of intermediate router states are evaluated prior to converging to an arbitrary final solution (there may be multiple feasible solutions [4]). Consequently, determining if a network is in the convergence process or if it is persistently oscillating is difficult.…”

“…The measurement data can be used for tasks such as deriving traffic demands in operational networks [1], finding traffic matrices [2] and their dynamics [3], and oscillation detection [4]. A majority of such tasks require some knowledge of the path traffic takes through a network -hence the need for routing measurements.…”

“…In [4], we introduced a model to analyze the oscillatory properties of a twotier route-reflector (RR) iBGP topology. We now extend this model to determine the network solution of a general RR iBGP topology.…”