DEVELOPMENT OF METAHEURISTIC BASED WORKFLOW AND TASK SCHEDULING ALGORITHMS IN CLOUD ENVIRONMENT

Abstract

Cloud computing is a network-based high-performance computing model which delivered access-based on-demand, scalable, fault tolerance, pay-per-use, and reliable services. It has a wide range of applications, including business, research, academic, and de-centralized interactions. It is a collection of resources such as powerful computing devices, software, storage, networks, and database that are available as a cloud service to its users with distinctive Quality-of-services (QoS) requirements. Scheduling is one of the prominent issues in cloud computing that truly exploited the parallelism of the system to achieve the high speed up and efficiency of cloud resources. It deals with assigning each task to the available cloud resources under the deadline and cost constraints in the cloud environment. Scheduling algorithms are classified into two categories based on the assigned tasks: workflow scheduling and task scheduling. Workflow scheduling performed the mapping of the dependent task to the virtual machines on different functional and non-functional requirements, whereas task scheduling maps the independent tasks to the unallocated cloud resources. Unfortunately, the scheduling problem belongs to the NP-complete category and challenging to achieve the optimal solution reasonably. It is preferable to produce the adequate sub-optimal solution of the scheduling problem quickly because there are no techniques that generate the optimal solution in polynomial time. Metaheuristic-based approaches produce the near-optimal solution for the workflow and task scheduling problems in a reasonable time frame. Therefore, we need a metaheuristic-based scheduling technique that can minimize the completion time of parallel applications under the deadline and cost constraints. There exist many scheduling techniques for the cloud computing models. The systemic illustration and analysis are required to analyze the various approaches of the task and workflow scheduling and to extract the research gap. It is worth proposing, designing, and developing the scheduling technique for dependent and independent tasks using a metaheuristic approach. The major objectives of the thesis are as follows. 1. Systematic illustration, classification, and key attributes of the scheduling techniques and to investigate the exiting scheduling techniques and their extensions for possible improvements. 2. Development of a management system for servicing workflow application of multi-organization on the community cloud model in a secure cloud environment. 3. Design and development of a secure framework and scheduling algorithm for independent tasks in a cloud environment. 4. Design and development of a secure framework for the virtual machine allocation for efficient VM placement on the physical machine in the cloud computing model.