An effective HPSO-MGA Optimization Algorithm for Demand based Resource Allocation in Cloud Environment

Cloud computing (CC) is the growing technology for providing multiple virtual infrastructures. Nowadays it is becoming and providing dynamics scaling of resources that are accessible to users as required. Dynamic Resource allocation means allotting the computing resources to the various services needed by the user without any time constraint with the help of virtual machines. It is also known as demand-based resource allocation. In the dynamic resource allocation method, the number of users remotely requesting the computing Resources. So resources are allocated to virtual machines on-demand basis. Here feature prediction is the important part by using the streaming data. So both the process was concentrate on this work. As a result, we propose a hybridization of Hybrid Particle Swarm Optimization and the Modified Genetic Algorithm (HPSO-MGA), which are developed to make the entire process for allocating resources dynamically by scattering tasks or requests in VMs. By using the rule generation process the feature outcome also predicted. The process begins from the User or Consumer by getting data from various resources through on-line. These data will be given to task manager. From the task manager, we can extract parameters like Cost of the task, Speed, Size of data and weight, etc. Similarly, we can extract the details of system resources like CPU Utilization, Memory Usage, Processing speed, and Process Cycle, Bandwidth, Number of requests, load on VM, Disk space from the cloud storage. Here the required features are selected by using Hybrid Particle Swarm Optimization and Modified Genetic Algorithm (HPSO-MGA) to decrease the processing time required for allocating resources dynamically. From all the features only necessary features will be used. Hence time taken for resource allocation is reduced and accuracy becomes low. So our proposed Hybrid Particle Swarm Optimization and Modified Genetic Algorithm (HPSO-MGA) reduce the execution time and increase the performance of resource allocation. Our proposed RA methodology shows better performance when compared with the existing methods.