Kolam inspired task scheduling for workflows in big data clouds

Abstract

Current advances in cloud computing and big data analytics have newlinepaved the way for modern technologies and innovative developments. The newlinenext generation of computing will largely be based on the efficient utilization newlineof resources and proper management of big data. Solving a well-established newlineNon-deterministic Polynomial-time (NP-Hard) problem of task scheduling and newlineload balancing in the cloud environment has been attempted by researchers newlineacross the globe for the past two decades. This is generally viewed as a newlinenon-linear optimization problem that distributes workloads across multiple newlinenodes to improve the response time and minimize the load imbalance among newlinethe nodes. With the advent of big data analytics and the need for timely newlineprocessing of humungous amounts of data, this problem has gained even more newlineattention. Even after realizing the moving computation to data paradigm, it newlineis observed that numerous data movements still occur due to task schedules newlinethat do not consider the data availability aspect. An optimal task scheduling newlinemechanism that simultaneously improves the response time and minimizes the newlineload imbalance degree with the least number of data migrations is the need of newlinethe hour. newlineBased on the extensive literature study conducted, it is observed newlinethat bio-inspired algorithms are largely applied in solving task scheduling newlineand load balancing in the cloud. By analyzing and applying certain peculiar newlinebehaviors of animals/insects like ants, honey bees, birds, swarms, fireflies, newlinegrey wolves, and so on, researchers have found breakthrough multi-objective newlineoptimization techniques for the scheduling problem. On close examination of newlinethese techniques, certainly, a scheduling and load balancing mechanism that is newlineaware of the underlying data placement is required. newline