使用參考佇列之高效能任務排程演算法於雲端運算環境之探討

Efficient Task Scheduling Algorithm with Reference Queues in a Cloud Computing Environment

陳彥儒(Yen-Ru Chen)

指導教授 ： 江茂綸 ； 林傳筆

朝陽科技大學/資訊學院/資訊與通訊系碩士班/碩士(2012年)

摘要

由於網際網路的蓬勃發展及電腦硬體的快速成長，使得現行的網路應用服務需求日漸增加，進而造成網路服務供應商的龐大負擔。因此，為了要應付大量的客戶需求，網路服務供應商必需提升相關硬體的計算能力及網路頻寬。然而，傳統的分散式系統無法因應現今龐大的應用服務需求。因此Google在2007年提出以分散式系統為基礎的網格運算新概念，稱之為雲端運算。雲端供應商透過虛擬化的技術建置了大量的服務節點及儲存資源。然而，在這些大量的雲端資源下，有效率的任務分配以及資源管理將是一個的重要課題。因此，本研究設計了一個參考佇列雲端服務架構Reference Queue based Cloud Service Architecture (RQCSA)及Reference Queues及Fitness Service Queue Selection Mechanism (FSQSM)排程演算法去提升服務效能，如服務任務量。此外，總執行時間及任務等待時間也將被降低。

關鍵字

分散式系統 ； 使用者導向 ； 排程演算法 ； 雲端運算

並列摘要

Due to the rapid growth of Internet and computer hardware, the demand of Internet application is getting growing, resulting in a large overhead for Internet Service Provider (ISP). To cope with a lot of user’s requirement, Internet Service Provider needs to improve the computational ability of hardware and network bandwidth. As a result, the traditional distributed system can not satisfy these huge demands for Internet application. A new concept of grid computing which is based on the distributed system is proposed by Google at 2007, called as cloud computing. Cloud Provider provides a lot of service nodes and storage resources by virtualization technology in a cloud computing environment. However, a number of scheduling algorithms are not efficient due to the task dispatch without considering the loads of clusters. Therefore, this study design a Reference Queue based Cloud Service Architecture (RQCSA) and Fitness Service Queue Selection Mechanism (FSQSM) to enhance the service performance, such as throughput of number of service tasks. Besides, the makespan and queue waiting time can be reduced in this study.

並列關鍵字

Schedule ； Cloud computing ； User-oriented ； Distributed Computing

參考文獻

[2]R. Armstrong, D. Hensgen, and T. Kidd, “The Relative Performance of Various Mapping Algorithms is Independent of Sizable Variances in Run-Time Predictions,” 7th IEEE Heterogeneous Computing Workshop, 1998, pp. 79-87.

[3]A. Bedra, “Getting Started with Google App Engine and Clojure,” Internet Computing, Vol. 14, No. 4, 2010, pp. 85-88.

[5]R. S. Chang, M. H. Guo, and H. C. Lin, “A Multiple Parallel Download Scheme with Server Throughput and Client Bandwidth Considerations for Data Grids,” Future Generation Computer Systems, Vol. 24, No. 8, 2008, pp. 798-805.

[6]Y. S. Dai, M. Xie, and K. L. Poh, “Availability Modeling and Cost Optimization for the Grid Resource Management System,” IEEE Transactions on Systems, Vol. 38, No. 1, 2008, pp. 170-179.

[8]R. L. Grossman, “The Case for Cloud Computing,” IT Professional, Vol. 11, No. 2, 2009, pp. 23-27.