Learning the Optimal Synchronization Rates in Distributed SDN Control Architectures

Abstract To support advanced coalition operations like secure information exchange between teams of soldiers that belong to different command centers in a tactical network, the control plane of the Software Defined Network (SDN) should be distributed. However, having multiple controllers managing the network brings additional challenges. One such challenge is how to coordinate the management decisions made by the controllers which is usually achieved by disseminating synchronization messages in a peer-to-peer manner. While there exist many architectures and protocols to ensure synchronized network views and drive coordination among controllers, these protocols typically induce significant traffic overheads, as we show through emulation measurements on OpenDaylight and ONOS controllers. These overheads hinder the application of SDN in wireless resource constrained environments like the army's coalition tactical ad hoc networks. To address this issue, in this paper, we introduce the SDN synchronization problem: how often to synchronize the network views for each controller pair. Our objective is the maximization of the performance of applications of interest which may be affected by the synchronization rate. Using techniques from learning theory, we derive an algorithm with provable performance guarantees. Evaluation results demonstrate significant benefits over baseline schemes that synchronize all controller pairs at an equal rate.
  • Konstantinos Poularakis (Yale)
  • Qiaofeng Qin (Yale)
  • Liang Ma (IBM US)
  • Sastry Kompella
  • Kin Leung (Imperial)
  • Leandros Tassiulas (Yale)
Date Sep-2019
Venue Annual Fall Meeting of the DAIS ITA, 2019