雲端運算系統中使雲端服務提供商利潤最大化之允入控制演算法

Abstract

雲端運算在技術蓬勃發展下已愈趨成熟，各式各樣的應用類型與日俱增，近年來被廣泛地採用。對於企業客戶或使用者而言，雲端運算使其得以透過以量計價的方式租賃第三方提供的軟硬體資源，令使用者能將資金做更有效的運用，因此，使用者希望雲端服務提供商能滿足其對軟硬體資源之需求，並提供高的服務品質。對於雲端服務提供商而言，他們試圖最大化資源利用率及其可所獲取之利潤，換言之，亦即在有限的資源中，創造最大的利潤。為了使資源在雲端運算環境中有效的被利用，雲端服務提供商須採用合適之負載平衡技術。因此，如何提供良好品質的服務予使用者，並最大化其利潤，是雲端服務提供商最具挑戰性的問題之一。 本論文著重於協助雲端服務提供商在有限的資源中，使其利潤最大化，並將此問題數學模式化，進而發展出使用者允入控制演算法。在求解的過程中，採用拉格蘭日鬆弛法協助求得最佳解。

Due to the cloud computing has gained rapid adoption in recent years, cloud service providers have begun to establish new data centers to implements cloud computing application. For end users, cloud computing is a scenario that users can access any kind of infrastructures, applications or platforms in a cloud. Therefore, they hope cloud service providers can offer high quality cloud environment and satisfy their requirements. For cloud computing service providers, they try to maximize resource utilization and maximize their profits simultaneously. In order to make efficient use of resources in cloud computing systems and ensure their availability to the end users, they need to adopt a best suited load balance technique to achieve it. Hence, how to provide efficient cloud services to users and maximize the cloud service providers’ profits at the same time has become an extremely important issue. In this thesis, we focus on helping cloud service provider to admit users as many as possible in the most efficient way. A generic mathematical programming model we proposed that can be used for developing an admission control and resources reallocation algorithm by simulating the role of the cloud service providers and users in cloud computing system. The Lagrangean relaxation method is adopted here to solve the problem mentioned above and obtain the optimal solutions.