RCP: A Temporal Clustering Algorithm for Real-time Controller Placement in Mobile SDN Systems

TL;DR

This paper introduces RCP, a real-time controller placement algorithm for mobile SDN systems that uses temporal clustering and control theory to achieve near-optimal, fast, and scalable solutions suitable for high-mobility scenarios.

Contribution

The paper proposes a novel real-time control placement algorithm based on temporal clustering and control theory, suitable for high-mobility SDN networks like UAVs.

Findings

RCP achieves up to 25 times faster performance than traditional methods.

RCP provides high-quality solutions with exponential convergence to near-optimal placement.

The algorithm has linear computational complexity, making it scalable for large networks.

Abstract

Software Defined Networking (SDN) is a recent paradigm in telecommunication networks that disentangles data and control planes and brings more flexibility and efficiency to the network as a result. The Controller Placement (CP) problem in SDN, which is typically subject to specific optimality criteria, is one of the primary problems in the design of SDN systems. {\em Dynamic} Controller Placement (DCP) enables a placement solution that is adaptable to inherent variability in network components (traffic, locations, etc.). DCP has gained much attention in recent years, yet despite this, most solutions proposed in the literature cannot be implemented in real-time, which is a critical concern especially in UAV/drone based SDN networks where mobility is high and split second updates are necessary. As current conventional methods fail to be relevant to such scenarios, in this work we propose a real-time control placement (RCP) algorithm. Namely, we propose a temporal clustering algorithm that provides real-time solutions for DCP, based on a control theoretic framework for which we show the solution exponentially converges to a near-optimal placement of controller devices. RCP has linear O(n) iteration computational complexity with respect to the underlying network size, n, i.e., the number of nodes, and also leverages the maximum entropy principle from information theory. This approach results in high quality solutions that are practically immune from getting stuck in poor local optima, a drawback that most works in the literature are susceptible to. We compare our work with a frame-by-frame approach and demonstrate its superiority, both in terms of speed and incurred cost, via simulations. According to our simulations RCP can be up to 25 times faster than the conventional frame-by-frame method.