Optimised clustering algorithm-based centralised architecture for load balancing in IoT network
Thair Al-Janabi, Hamed Saffa Al-Raweshidy · 2017
Considerable efforts have been expended on the centralised administration of WSNs. However, only very small functions, such as processing, memory, battery unit, and communication ability, can be configured using WSN nodes due to the associated resource restrictions. Moreover, unbalanced cluster construction and unbalanced energy dissipation can reduce network lifespan by a large extent. Therefore, this paper proposes an efficient clustering algorithm based on load adjustment for the IoT under the SDN architecture. The basic idea is to utilise cloud resources such as data-centres and storage units by a centralised SDN controller located over the cloud to calculate a load-balanced PSO clustering algorithm. The SDN controller implements the PSO that considers load-balancing, communication cost and remaining energy factors to construct a clustering table (CT). This CT contains information utilised in cluster formulation, such as an optimal set of cluster heads (CHs) and cluster members (CMs). Simulation results demonstrate that the suggested load-balancing algorithm is comparatively more efficient in terms of network lifetime, energy dissipation and volume of data send to the sink.