云计算可伸缩性的测试、评价和优化研究

项目名称： 云计算可伸缩性的测试、评价和优化研究

项目编号： No.61472197

项目类型： 面上项目

立项/批准年度： 2015

项目学科： 自动化技术、计算机技术

项目作者： 白晓颖

作者单位： 清华大学

项目金额： 82万元

中文摘要： 云计算已成为互联网环境下，计算机系统体系架构的一个主要发展趋势。可伸缩性（或弹性）关注计算资源动态调度的效率和性能，是云平台性能评价中的一个重要指标。但由于云计算系统的负载规模大且难以预测、多租户体系结构设计、按需动态资源分配等特性，给可伸缩性的评价、测试和优化带来了挑战。本项目旨在针对基于云平台、面向服务的互联网软件系统的特点，从可伸缩性度量和评价模型、典型负载模型、资源动态重构和优化部署几个方面开展相关研究。考虑异构资源、异构平台对性能的影响，兼顾性能的经济性等因素，研究多指标、多维度、综合化的系统评价方法；利用数据分析、机器学习等技术，建立负载时变性和波动性的模型，实现基于现实场景的负载模拟；基于博弈论、线性规划等优化分析方法，研究最优化资源配置选择推荐算法和动态部署技术。 具有良好可伸缩性的系统结构设计和高效的计算资源调度，对保证系统服务水平至关重要。本项目工作将为云计算性能研究探索新的理论、方法和技术，具有重要的意义。

中文关键词： 云计算；可伸缩性；软件测试；评价；优化

英文摘要： Cloud computing aims to support an advanced level of massive scalability so that it can provide necessary resources on demand following a pay-per-use pricing model. Scalability (or elasticity) is one of the most promising benefits of Cloud computing, and a critical measurement of Cloud performance. However, Cloud introduces many new challenges to conventional scalability research, such as highly fluctuated internet workload, sharing and dynamic allocation of large-scale virtualized resources, and multi-tenancy architecture. To address the challenges, the research aims to investigate new methods and techniques from three perspectives: scalability measurement and evaluation, Cloud workload characterization, and dynamic reconfiguration and provisioning of Cloud resources. A multi-dimensional, cost-aware scalability model will be provided, considering heterogeneous resources and platforms. Workload models will be investigated based on performance data analysis using analytic techniques. It will also propose methods and algorithms for cloud resource configuration and provisioning, based on optimization theories like Game theory and linear programming. Well-design scalable architecture and efficient resource scheduling are critical to ensure system service level agreement. The project will research on the theory and innovative techniques on cloud performance engineering. It will be great beneficial to Cloud computing.

英文关键词： Cloud Computing;Scalability;Testing;Evaluation;Optimizaiton