Using clustering approaches for response time aware job scheduling model for internet of things (IoT)

Abstract

Internet of things (IoT) is the next evolution of internet connecting and transferring information between smart things/objects covering daily aspects of life. This is realized with the involvement of large number of sensor nodes which monitors, generates and enables aggregation of data. In such an environment, clustering becomes essential by grouping a structure of objects with similar attributes. Clustering, helps in establishing topologies which in turn can be used for optimizing the quality of service (QoS) parameters while managing the resources in the underlying dynamic and heterogeneous IoT network environment. This work proposes to study and compare K-means, hierarchical clustering and fuzzy C-means clustering (FCM) algorithms to design a response time aware scheduling model for IoT. The work intends to improve the QoS by routing the data through clusters formed using the above three algorithms to observe the effect of clustering on the response time aiming to minimize the same. Establishing the clustering scheme with optimum response time results in optimizing the scheduling performance of the underlying network too by minimizing the overall execution cost. The effect on message scheduling to account for the prioritized message delivery has been studied. Simulation study proves the efficiency of the K-means clustering approach under various test conditions.