Optimal Consistent Network Updates in Polynomial Time

Software-defined networking (SDN) enables controlling the behavior of a network in software, by managing the the forwarding rules installed on switches. However, it can be difficult to ensure that certain properties are preserved during periods of reconfiguration. The widelyaccepted notion of per-packet consistency requires every packet to be forwarded using the new configuration or the old configuration, but not a mixture of the two. A (partial) order on switches is a consistent order update if updating the switches in that order guarantees per-packet consistency. A consistent order update is optimal if it allows maximal parallelism, where switches may be updated in parallel if they are incomparable in the order. This paper presents a polynomial-time algorithm for computing optimal consistent order updates. This contrasts with other recent results, which show that for other properties (e.g., loop-freedom and waypoint enforcement), the optimal update problem is np-complete.