Toward a Flexible Design of SDN Dynamic Control Plane: An Online Optimization Approach

Abstract

With a centralized control over the forwarding devices and the embedded flows, Software Defined Networking promises to increase the flexibility of communication networks. Meanwhile, a dynamic control plane would adapt itself in a timely manner to sustain flow setup performance in the face of traffic variations. Such adaptation depends on a careful decision of the controller placement, which is challenging because we need to consider two contradictory objectives, namely the cost of operating the control plane and the cost of its adaptation. In this work, we model the problem of operating the control plane as a multi-period offline optimization problem to minimize the total cost induced by the flow setup performance and the control plane adaptation. We leverage the lookahead control scheme and decompose the intractable offline problem into smaller instances, which are solved in an online fashion efficiently with an algorithm based on simulated annealing. We perform extensive simulations on real world topologies and show that our proposed algorithm can reduce the total cost by up to 20% compared with the reference algorithms. Further, we analyze the need of frequent control plane adaptation, and compare different control plane design choices according to a novel flexibility measure.

Optimizing the Flexibility of SDN Control Plane

“…Initial literature focuses on static objectives, e.g., average control latency [7], [10] and average inter-controller latency [11], which depend only on the topology and the number of controllers. Dynamic objectives, however, depend on varying networking conditions such as traffic in the network [5], [12], and node/link failures [13]. From the solution's perspective, the existing works are classified into two categories: optimal solution and heuristic algorithms.…”

“…We need to solve different CPP instances over multiple time slots, where the movement of controllers are subject to certain constraints. Online optimization techniques are preferred for such a problem [5], [19]. A comprehensive survey of CPP can be found in the work of Das et al [2].…”

“…Formally, given a changing traffic input, the adaptation of distributed control plane means to place the controllers dynamically so that particular performance metrics such as average flow setup time or load imbalance [5], [6] can always be optimized. The optimization problem is first introduced as the Controller Placement Problem (CPP) by Heller et al [7].…”

“…This problem sounds trivial: a naive approach could place DCs close to large cities where most of the traffic originate and terminate. However, in a multicontroller scenario, inter-domain flows would also trigger flow setup requests at intermediate switches [5] which are not necessarily in the proximity of large cities, not to mention the flows variate from time to time. Fig.…”

“…Our main objective is to maximize the flexibility of the control plane under the constraints of adaptation time and cost. The control plane adaptation consists of controller migration and switch re-assignment [5]. Implemented as VM migration, the overhead of controller is proportional to the size of the VM and the available bandwidth for migration.…”