The increase in the number of SDN-based deployments in production networks is triggering the need to consider fault-tolerant designs of controller architectures. Commercial SDN controller solutions incorporate fault tolerance, but there has been little discussion in the SDN community on the design of such systems and the tradeoffs involved. To fill this gap, we present a by-construction design of a fault-tolerant controller, and materialize it by proposing and formalizing a practical architecture for small to medium-sized networks. A central component of our design is a replicated shared database that stores all network state. Contrary to the more common primarybackup approaches, the proposed design guarantees a smooth transition in case of failures and avoids the need of an additional coordination service. Our preliminary results show that the performance of our solution fulfills the demands of the target networks. We hope this work to be a first step in what we consider a necessary discussion on how to build robust SDNs.

Maintaining a strongly consistent network view in a Software Defined Network has been usually proclaimed as a synonym of low performance. We disagree. To support our view, in this paper we argue that with the use of modern distributed systems techniques it is possible to build a strongly consistent, fault-tolerant SDN control framework that achieves acceptable performance.The central element of our architecture is a highly-available, strongly consistent data store. We describe a prototype implementation of a distributed controller architecture integrating the Floodlight controller with a data store implemented using a stateof-the-art replication algorithm. We evaluate the feasibility of the proposed design by analyzing the workloads of real SDN applications (a learning switch, a load balancer and a device manager) and showing that the data store is capable of handling them with adequate performance.

No abstract