Performance Analysis of Constrained Device Virtualization Algorithm

Abstract

Internet of Things aims to automate and add intelligence into existing processes by introducing constrained devices such as sensors and actuators. These constrained devices lack in computation and memory resources and are usually battery powered for ease of deployments. Due to their limited capabilities, the constrained devices usually host proprietary protocols, platforms, data formats and data structures for communications and therefore, are unable to communicate with devices from different vendors. This inability leads to interoperability issues in Internet of Things which, is in fact against the spirit of Internet of things which, envisions interconnection of billions of devices and hence, results in an isolated, vendor-locked and close-loop deployments of IoT solutions. Various approaches have been made by the industry and academia to resolve the interoperability issues amongst constrained devices. However, majority of the solutions are at different layers of the communication stack but do not provide a holistic solution for the problem. In more recent research, there have been theoretical proposals to virtualize constrained devices to abstract their data so that its always available to applications. We have adopted this technique in our research to virtualize the entire Internet of Things network so that virtual TCP/IP based protocols can operate on virtual networks for enabling interoperability. This paper proposes the operations of the Constrained Device Virtualization Algorithm and then simulates it in CloudSIM to derive performance results. The paper further highlights open issues for future research in this area.