Heuristics for friendship selection in Social internet of things

Abstract

Internet of things is considered as the future of computing era where every newlineconventional object is equipped with sensing and communicating capabilities. newlineThe objects will co-operate with each other to achieve a common goal. Social Internet newlineof things is the recent advancement of Internet of Things in which social networking newlinetheories are adapted to address challenges in Internet of Things. In social Internet of newlineThings environment objects will autonomously establish relationships and newlinecommunicate with each other. The objects that are equipped to establish relationships newlineshould have intelligence to select its relationships. The intelligence should be provided newlinebased on the heterogeneity of the devices in the network, accessibility of the devices newlineand search latency of the network. newlineThe relationship management of Social Internet of Things administers the newlineselection of relationship. The friendship selection heuristics governs the overall newlinestructure of the network and hence it should be designed intelligently. This work newlineproposes two relationship heuristics and a fault tolerance algorithm along with newlinesimulation of Social Internet of Things network to test the proposed algorithms. newlineThe proposed relationship heuristic is the first work that introduces the type of newlinerelation attribute in friendship selection to enhance the feasibility of the network. newlineThe relationship heuristic accommodates a reachability equation to enhance the newlinenavigability of the network. The proposed navigability heuristic divides the threshold newlinefor determining the number of friendships into two, one half will include potentially newlineimportant relationships and another half will include relationships that will increase newlinenavigability. The proposed fault tolerance algorithm maintains the total network newlinenavigability achieved by a device even when some devices leave the network. newlineA network which mimics real world Social Internet of Things network is simulated to newlinetest the proposed algorithms. The performance of the algorithms are evaluated based on newlineaverage degree, giant component, average clustering co efficient, network diameter and newlineaverage path length of the resultant network obtained after applying the proposed newlinealgorithms on the simulated SIoT network. The network properties that are considered newlinefor evaluation influences the overall performance of the network. newline