دانلود رایگان مقاله انگلیسی توزیع موثر درخواست ها در محیط محاسبات ابری متعهد با استفاده از توزیع آماری - الزویر 2019

Abstract

One of the main questions in cloud computing environments is how to efficiently distribute user requests or Virtual Machines (VMs) based on their resource needs over time. This question is also an important one when dealing with a cloud federation environment where rational cloud service providers are collaborating together by sharing customer requests. By considering intrinsic aspects of the cloud computing model one can propose request distribution methods that play on the strengths of this computing paradigm. In this paper we look at statistical multiplexing and server consolidation as such a strength and examine the use of the coefficient of variation and other related statistical metrics as objective functions which can be used in deciding on the request distribution mechanism. The complexity of using these objective functions is analyzed and heuristic methods which enable efficient request partitioning in a feasible time are presented & compared.

Conclusion & Future Works

In this paper we looked at an efficient method of distributing customer re540 quests among cloud service providers in a federated cloud environment, with the goal of utilizing statistical multiplexing. We showed how the coefficient of variation can be considered as an important objective function in this regard, showing the smoothness of a collection of requests. We described the k-FEOptRP problem for efficient request distribution and examined the com545 plexity of various statistical functions being set as its objective function, with most of them being NP-Hard. Finally we looked at various heuristic algorithms for the k-FEOptRP problem with CoV being set as its objective function and compared them in various simulations scenarios. Our results show that our algorithm based on the Late Acceptance Hill Climbing method outperformed 550 others. In our work we looked at the behavior of each request in a single time frame, distributing them to various CSPs in one go. In reality each customer request or process will show much varying resource needs and correlation with other requests in its lifetime and therefore we can look at its behavior dynamically, 555 changing the CSP to which it is partitioned to multiple times. Hallac et al. [31] use a similar concept in looking at covariance-based clustering of multivariate time series data which can be of use in this regard. This is an important direction our future work can take. We would like to also test out CoV based request partitioning on larger data sets in real world scenarios in which large 560 scale enterprise grade processes are at play.