QoS-Aware Dynamic Caching for Destroyed Virtual Machines in Sensor-Cloud Architecture

Abstract

In this work, we propose a scheme, named Quality-of-Service (QoS) Aware Dynamic Caching for Destroyed Virtual Machines in Sensor-Cloud Architecture, which enables efficient caching in sensor-cloud, in the presence of heterogeneous sensor nodes. This work is one of the first attempt of its type, in which a special cache is introduced for the efficient use of sensor data in the sensor-cloud architecture, in order to maintain QoS. Considering the reutilization of sensor data, the proposed scheme is capable of keeping data of a Virtual Machine (VM) for a certain duration of time, even if it is destroyed. Therefore, the data from SDC can be used in future, if any further requests arrive, which consists of same configurations of physical sensors inside a virtual sensor. We compared the proposed caching mechanism, Dynamic Caching for Destroyed VMs, with the existing mechanism proposed by Chatterjee and Misra [1]. We observe that the cache hit percentage increases at least double the number of times exhibited by the existing scheme of caching. On the other hand the energy consumption and message overhead decrease by 50% and 17% respectively.