Programmable architecture based on Software Defined Network for Internet of Things: Connected Dominated Sets approach

In this paper, we propose a new programmable architecture based on Software Defined Network (SDN) paradigm for network control functionalities in Internet of Things (IoT) using the Connected Dominating Sets (CDS). In order to reduce the traffic load and to avoid a single point of failure at the controller node, we distribute the controllers role by introducing three levels of control: Principal Controller (PC), Secondary Controller (SC), the Local Controller (LC). The PC has a global view of the network infrastructure which is not the case of SC where it focuses only on one network technology. The LC acts locally by managing and relaying signaling messages from ordinary nodes to the SC. In order to select the LC nodes, we propose a Distributed Local Controller Connected Dominating Set algorithm (DLC-CDS). The DLC-CDS is a distributed algorithm with single phase and supports the dynamic network topology. The selection strategy of DLC-CDS is based on an important function named score, which is computed using the fuzzy logic and it depends on several parameters such as: the connectivity degree, the average link quality, and the rank. The performance of the proposed DLC-CDS are evaluated and compared with another solution named Distributed Single Phase-CDS (DSP-CDS) using many scenarios with different parameters: the node density and the radio range. The obtained results show that the DLC-CDS converges rapidly with a minimum CDS size compared to a DSP-CDS.