Article Preview
Top1. Introduction
Nowadays, there is a change in the Internet concept from a set of connected devices to be a set of connected surrounding daily life objects such as TVs, Lamps, transportations, medical devices, and sensors. The effective interaction between these heterogeneous objects through the Internet leads to the Internet of Things (IoT). Each object in IoT is called a ‘thing’ that should be defined using the unique address. The radio-frequency identification (RFID) is used to achieve communication between things. Furthermore, many intelligent web applications are used to manage the IoT systems. The new IoT applications depend on the promotion of wireless sensor networks (WSNs) and inspire experts and engineering equally with the potential to achieve smart proactive actions. These IoT applications contain the big amount of devices, which are connected with different requirements and technologies.
IoT can even be seen as a loosely coupled network of networks connecting things/objects surrounding us. Each constituent in IoT is called a ‘thing’ and should be defined using unique addressing e.g. Uniform Resource Name (URN). The radio-frequency identification (RFID) is used to achieve communication between things. IoT is different from the internet in the sense that the internet connects machines to machines or web pages to web pages. In contrast, using the internet, IoT can connect machines as well as anything or everything surrounding us using sensors referred to as objects. This brings into the picture the use of RFID technology either reactively or proactively. There is a definite amalgamation of distributed computing with the IoT under the cloud-based condition. Therefore, Cloud computing provides efficient resource utilization and along with IoT can allow users to scale up to solve larger science problems. The cloud-based IoT model is quite suitable for the current applications being Service Oriented. (Hwang, Dongarra, & Fox, 2012; Sundmaeker, Guillemin, Friess, & Woelfflé, 2010; Uckelmann, Harrison, & Michahelles, 2011; Jayayardhana et al., 2013; Atzori et al., 2010; Zanella, Bui, Castellani et al., 2014; Guo, Zhang, & Wang, 2011; Vermesan, Friess, & Friess, 2011; Dennis & Dominik, 2010; Rao et al., 2012)
The IoT applications generate data through sensing, collect the data for analysis and send this data to the administrator nodes or even a cloud at the backdrop for data analytics and other possible actions. The routing protocols for the same works in two ways. The first way is to improve data transmission and scalability. The second way is to minimize energy consumption. Even though it faces the challenges like the unreliability of low-power nodes, limited resources and data should be transmitted through low delay and loss rate as the minimum Quality of Service (QoS). IoT produces an enormous amount of data, but this data is collected by network nodes under different network topology. This heterogeneity restricts the TCP/IP to be the best policy demanding proper resource allocation both for computing and data routing.
Owing to the above-mentioned constraints, different people based on their perception and requirements making it an NP-Complete problem can view IoT differently. The different job scheduling methodologies reported are either heuristic or metaheuristic based. Heuristic-based methodologies are more helpful when we look for local optima while metaheuristic approaches try to explore the solution space further to attain global optima. Despite the fact that, metaheuristic approaches look very engaging, a large number of parameters to be tuned on account of IoT limits the utilization (www.airclearenergy.com, n.d.; Evans, 2011; Vermesanovidiu et al., n.d.; Aggarwal, Ashish, & Sheth, 2013; Abdullah & Yang, 2013; Sumit & Zahid, 2017).