Enhanced fault identification and optimal task prediction (EFIOTP) algorithm during multi-resource utilization in cloud-based knowledge and personal computing

Abstract

Virtualization technology is playing an important role in cloud computing for efficient task scheduling and application deployment. Cloud computing offers a platform to store and retrieve a large volume of information without any restriction on time or location. The system optimizes the available resource based on the user application requirement. Server and data storage devices can access distributed data residing in remote places via virtualization mechanism, where cloud applications are easily migrated from one server to another. Issues related to fault identification and resource optimization problems often occur in a cloud environment. To resolve these issues, an enhanced fault identification and optimal task prediction (EFIOTP) algorithm are proposed for finding and preventing faults during task execution with multiple resources. The research work objective is to design a deadline-determined resource allocation model with the VM resource isolation method in a cloud. The proposed work evaluates the maximum amount of task execution by considering different types of resources to identify and predict the faults at various levels and to minimize the occurrence of faults and task execution time. Based on the experiment evaluation, the proposed EFIOTP algorithm reduces 775 task completions (TCT), 0.237 datacenter server utilization (DCSU), 2% virtual machine cost (VMC), and improves the 0.39 hypervolumes (HV) on several parameters and scientific workflow application.