Multi-Sink Load-Balancing Mechanism for Wireless Sensor Networks

Wireless Sensor Network (WSN) is gradually taking the center stage as the foundation network for the Internet of Things (IoT), however, it is not without its challenges. These include lossy wireless links, limited processing capacity, limited memory and storage, and most importantly limited battery life. These limitations are inherent because of the small form-factor of a sensor device, which could be as small as a finger-tip. Of these challenges, the most pressing is limited battery life or power as this directly affects the network lifetime. Subsequently, the standard WSN architecture presents only one sink, serving as a gateway between the local sensor network, and the external network which in most cases is the Internet. The demanding pressure of routing data through a single sink makes it a Single Point of Failure (SOF), this means if the sink goes down, the entire network goes down. This practice is not only suboptimal but inefficient. In this paper, we propose a Multi-Sink Load Balancing Mechanism (MSLBM) for WSN. This mechanism adaptively and dynamically assigns network loads or traffic to sinks with less instantaneous operational tasks. It operates in such a way that no sink is overloaded with tasks which may result in early battery drainage and inadvertently untimely shutdown of the network. MSLBM was benchmark with related mechanisms with an outstanding performance of 98% Packet Delivery Ratio (PDR), least Latency of 21.7ms, and least Power Consumption of 1.6mW. This, therefore, presents MSLBM as an optimal energy-efficient load-balancing mechanism for a prolonged operational lifetime.

I. Introduction

The applications of the Internet of Things (IoT) are becoming a common sight nowadays. The word “Smart” is now associated with IoT applications, as seen in Smart homes, Smart factories, Smart military, Smart logistics, and the list goes on [1]. In this instance, Smart means the ability of an object to execute an expected task automatically without the intervention of humans. IoT has made the interaction between man and machines a reality, especially with the application of Artificial Intelligence (AI) and Machine Learning (ML). Sizes of IoT devices vary from very small to as small as the tip of the finger. The practical operations of IoT are made possible by Wireless Sensor Networks (WSN) [2], [3] These are networks of spatially dispersed and dedicated sensor devices installed in locations that are meant to be monitored. These sensors, sometimes referred to as nodes are limited in power and capacity due to their miniaturized form factor. While in operation, they intentionally collect sensed data and route them to locations where they can be used to trigger another operation or stored for further analysis.