Robust Controller Placement And Assignment In Software-Defined Cellular Networks

Software-defined cellular networks (SDCN) have been recently introduced to enable flexible cellular network design that facilitates fulfilling SG design requirements. Placement of controllers within the SDCN plays a crucial role in optimizing its performance. In this paper, we study the controller placement problem in SDCN, considering the uncertainty in cellular user locations. Specifically, our contributions are as follows. First, we develop (CP2)-P-3, a static joint stochastic controller placement and evolved node B (eNB)-controller assignment problem. The objective of (CP2)-P-3 is to minimize the number of controllers needed to control all eNBs, while ensuring that the response time to each eNB will exceed 6 seconds with probability less than 1 fl. Second, we develop CPPA, a joint stochastic controller placement and adaptive eNB-controller assignment problem. In contrast to (CP2)-P-3, in CPPA the eNB-controller assignment adapts to variations in the eNB request rates, resulting from the variations in the cellular user locations. Finally, we use sample average approximation combined with various linearization techniques to solve and evaluate (CP2)-P-3 and CPPA under various system parameters. Our results demonstrate the advantages of (i) joint compared to sequential optimization, (ii) stochastic compared to deterministic optimization, and (iii) adaptive compared to static optimization.