Cloud computing, a buzzword of today's that combines the power of both parallel and distributed computing. It delivers its output in the form of service(s) that can be IaaS, SaaS and PaaS (Infrastructure, Software and Platform-as a Service). In Cloud computing, we won't compute on local machines, but on someone premises operated by someone else. Actually Cloud environment deals with dissimilar kinds of virtualized resources. So, to allocate and schedule resources efficiently it requires noticeable efforts. One of the core phases is task scheduling which plays a vital role. It can be seen as the finding an optimal assignment of set of task(s) over the available resource set to obtain desired goals like: cost, quality of service and makespan etc. Even, most of the organizations already started implementing CTQ model (less COST, minimum TIME and assured QUALITY) for attaining maximum return with assured quality. The objective of this paper is to review various independent task scheduling techniques under heuristic mapping category so that we can apply techniques according to current requirement.

Due to constraints along with profit margins in background, service provider’s sometime neglect to feed essential services to their respective clients. Such compulsion raises the demand for efficient task scheduling that can meet multiple objectives. But without any prior agreement, again makes a casual approach. So this dispute can be addressed when competent scheduling executes right over the Service Level Agreement. It acts as hotspots to define set of rules to assure quality of service. At this time, there is a huge demand of SLA opted scheduling that can produce profitable results from provider’s and client’s as well. This article presents a fundamental approach that can be applied to existing scheduling techniques on the fly. Result shows drastic improvement in terms of average waiting time, average turnaround time without comprising provider’s cost margin at all along with fairness.

Job scheduling process under the roof of Cloud is consisting of three phases: Resource Discovery, Resource Selection and Task Scheduling (Yousif, A., et al., 2011. A taxonomy of grid resource selection mechanisms. International Journal of Grid and Distributed Computing, 4(3), pp.107.117). Among them, task scheduling is always treated a cumbersome activity because it mapped task(s) to their assigned resource(s) based on various constraints and impositions according to requirements. Task scheduling in Cloud environment is broadly categorized into two streams that are: Heuristic and Meta-Heuristic (sometimes combination of both). Heuristic approach is further categorized into two streams: Immediate mode or online approach and Batch mode or offline scheduling technique. MaxStd, heuristic mapping, is one of the efficient batch modes scheduling technique for independent task(s) due to its inherent efficiency and performance. In this paper, we have proposed an improved version of MaxStd (I-MaxStd) that refines the mapping process of conventional MaxStd to yields an efficient output in the form of reduced makespan and better resource average utilization rate without compromising its legacy. The validation of proposed work has been done for heterogeneous types of ETC matrices being used as dataset.

No abstract

No abstract