支持消费转移的云资源分配与定价机制研究

In view of the phenomenon on serious unbalance of cloud resources consumption, the main goal of this project is to design a mechanism with consumption shifting ability to allocate and price cloud resources efficiently. By considering consumers’ alignment and the real situation of consumption shifting in the cloud market, as well as the cost for unit computing and the different payment of peak and valley time period, this funding focuses on researching the key issues about mechanism design for cloud resource allocation, shifting quantitative calculation and cloud resource pricing with the properties of individual rationality, budget balance and incentive compatibility. By studying the above contents, it will not only form a methodology for the effective use of cloud services, but also provide a theoretical guidance for other allocation and scheduling problem with the property of resource competition. This project is committed to achieving four specific purposes: 1) Modeling the dynamic environment cloud market, and designing an agent-based cloud resources consumption shifting architecture, sequentially achieving the goal of allocating and adjusting the consumption of cloud resource according to needs; 2) Designing a set of bidding language with the abilities to report users' preference, and effectively express the consumers' alignment willingness and personalized requirements; 3) Designing the cloud resource allocation and pricing algorithm for computing task, and designing the shifting mechanism which satisfies with the properties about both optimal revenue and competitive equilibrium; 4) Based on the case study and simulation experiment, testing and applying the cloud resources shifting mechanism on the real platform of ‘Alibaba Cloud’.

针对云资源消费的严重不均现象，本项目拟设计并实现支持消费转移的云资源分配与定价机制，以有效均衡平台负载提高资源利用效率。项目从云市场架构、用户结盟和竞价语言入手，综合考虑峰谷时段的单位计算开支及消费转移成本，重点研究兼顾个体理性、预算平衡和激励兼容等属性的云资源分配、转移定量与资源定价等问题。通过本项目研究，不仅为云服务的均衡利用提供方法论，而且对其它具有竞争属性的资源分配、调度问题提供理论指导。该项目致力于实现下述四个具体目标：1）对云市场动态环境建模，设计一种分层的云资源消费转移架构，实现云资源消费量动态按需分配和调整；2）定义一种具有偏好表示能力的用户竞价语言，有效表达云消费者的结盟意愿及个性需求；3）设计面向计算任务的云资源分配与定价算法，实现满足整体收益最优及竞争均衡的转移机制；4）在案例分析及模拟实验的基础上，实现云资源消费转移机制在真实的“阿里云”平台上的测试应用。

为了能够有效地均衡平台负载提高资源利用效率，本项目重点研究了兼顾个体理性、预算平衡和激励兼容等属性的云资源分配、转移定量与资源定价等问题。本项目取得了以下四个成果：1）对云市场动态环境建模，设计了一种分层的云资源消费转移架构，实现了云资源消费量动态按需分配和调整；2）定义了一种具有偏好表示能力的用户竞价语言，有效表达了云消费者的结盟意愿及个性需求；3）设计了面向计算任务的云资源分配与定价算法，实现了满足整体收益最优及竞争均衡的转移机制；4）在案例分析及模拟实验的基础上，实现了云资源消费转移机制在真实的测试应用。本项目研究不仅为云服务的均衡利用提供方法论，而且对其它具有竞争属性的资源分配、调度问题提供理论指导。

内容获取失败，请点击重试