Implementation of Multi agent and Multi source dynamic Resource allocation for IoT based cloud computing environment

Abstract

IoT based cloud computing system faces new challenges every day, due to the complex structure of system clusters and high volume of data processed by the systems. The ability of acquiring resources in an elastic manner is considered as the primary rationale for adopting IoT based cloud computing system. Elasticity mainly supports the facility to grow and shrink the virtual resources dynamically according to the requirement of IoT based cloud users. This article proposes a Multi-source framework using multisource QoS based Resource Allocation (QRA) and Multi agent Dynamic resource allocation (MADRA) Algorithm for increasing the flexibility and efficiency of resource allocation using virtualization. In the proposed framework, each source monitors and investigates all requests and processor availability before finding and allocating the resource. QRA algorithm is used to utilize the resources effectively and reduce the congestion by using vii multiple intermediate layers. On the average, the proposed framework approach provides 20.52% improvement in response time, 13.29% reduction in power consumption, and 1% error in prediction when compared to existing Efficient Resource Allocation (ERA) approaches. From simulation results, when the input load frequency is 200 Hz the percentage of resource allocation is 99.18% which is high when compared to that existing approaches. Experimental results indicate that the proposed approach is capable for more user requests and it improves QoS parameters such as completion time, response time and power consumption.