An Smarter Multi Queue Job Scheduling Policy for Cloud Computing

摘要

Cloud computing is an extension paradigm of grid and distributed computing. Cloud providers mainly focus on managing computing power, energy consumption, storage and services that are assigned to external users via internet. Managing users requirements have created several challenges in optimize job scheduling and on-demand resource allocation. Cloud job scheduling can be viewed as NP-hard optimization problem. An efficient job scheduler should choose scheduling policy to increase the performance of system. In earlier research an efficient multi queue scheduling (MQS) algorithm was build which divide user jobs in multiple queues and carry out dynamic selection of user jobs for execution. It successfully plummet the problem of fragmentation associated with the tradition job scheduling algorithms like First Come First Serve, Round Robin etc but left behind some drawbacks of higher switching time between multiple queues and dynamic selection posses high probability of indefinitely postponement of different types of user jobs causing long job waiting time therefore results in higher energy consumption. To address this issue, inspired by the concept of multi queue scheduling we introduce a Smarter MQS model which effectively separate user jobs into two job queues then give more importance in formation of merging jobs pattern by merging user tasks from both queues for execution, so the technique will empower us to reduce energy consumption while naturally to some degree will reduce job completion time and the overall cost. The proposed technique will achieve a high degree of job scheduling in cloud computing environment.