Article Preview
TopIntroduction
Sensor cloud is a new trend in the present scenario to overcome from the drawbacks of Wireless Sensor Network (WSN). The issues of WSN are node deployment, battery energy, bandwidth, routing, QOS, robustness, storage, fault mechanisms and security requirements (Akyildiz, 2002). The energy consumption is a trending issue in WSN from past many years and still many researchers are working on different scenarios to create an optimal system which consumes less energy for transmission and reception. Multiple research has been carried out to create energy efficient WSN (Sulieman, 2018; Anastasi, 2009; Jones, 2001; Karaki, 2004; Kumar, 2009; Ehsan, 2010; Tang, 2010; Zaman, 2016). Energy efficient WSN leads to improved network efficiency and network lifetime.
On the other hand, cloud computing is an efficient information storage paradigm which facilitates the end users to access the information at any point of time irrespective of where they are located. Cloud computing can improve the storage and processing capacities of WSN with great extent. Hence the sensor cloud is used to improve the overall system like gathering the data, accessing the information, processing the data and importantly storing the data (Madria, 2014; Alamri, 2013), this removes the long term energy consumption issue of WSN and improve its lifetime. Sensor cloud is a unique sensor data storage, visualization and remote management platform that support cloud computing to afford data scalability, rapid visualization and user programmable analysis (Lan, 2010). The sensor cloud is used in wide variety of applications like, health monitoring, environmental monitoring, transportation, agriculture, military and industrial monitoring (Alamri, 2013; Jit 2010).
Load balancing is one of the network energy saving technique in sensor cloud where it decides which node to be used for proper load distribution in the network. The load balancing is required to enhance the network lifetime and to make all the nodes operable in the network (Dumbrava, 2010; Zhang 2009). Load balancing technique not only concentrates on all nodes but it also concentrates on specific nodes which plays a vital role in the network. The considerable parameters for measuring performance efficiency of load balancing are reliability, adaptability, fault tolerance, throughput and waiting time (Yana, 2018). Commonly load balancing algorithm has five major components, transfer policy, selection policy, location policy, information policy and load estimation policy as discussed in (Amar, 2011). The highlighting features of load balancing are, reduced task waiting time, minimized task response time, maximum utilization of resources and throughput, improves reliability, stability and allows further modification if system prefers so (Manekar, 2012). The load balancing techniques are broadly classified into two types, static load balancing technique and dynamic load balancing technique (Rajguru, 2012). The static load balancing refers certain predefined information and does not consider status of the system. The dynamic load balancing on the other side improves better load distribution with additional communication and computation overhead. The static load balancing further has many techniques namely, round robin algorithm, randomized algorithm, central manager algorithm and threshold algorithm. The dynamic load balancing types are ant colony algorithm, central queue algorithm, honeybee foraging behavior, local queue algorithm, least connection algorithm and nearest neighbour algorithm (Kushwaha, 2015; Randles 2010).