Automated network repair with Meta provenance

Abstract

When debugging an SDN application, diagnosing the problem is merely the first step - the operator must still implement a solution that works, and that does not cause new problems elsewhere. However, most existing SDN debuggers focus exclusively on identifying the problem and offer the network operator little or no help with finding an effective fix. Finding a fix is challenging because the number of potential repairs can be enormous. In this paper, we propose a first step towards automated repair for SDN applications. Our approach consists of two elements. The first is a data structure we call meta provenance, which can be used to efficiently find good candidate repairs. Meta provenance is inspired by the provenance concept from the database community. However, whereas standard provenance can only reason about changes to data, meta-provenance can also reason about changes to programs. The second element is a system that can efficiently back-test a set of candidate repairs using historical data from the network. This is used to eliminate candidate repairs that do not work well, or that cause other problems. We present initial results from a case study, which suggest our approach is able to efficiently find high-quality repairs.