面向跨云协同计算的资源融合与优化共享研究

Cross-Cloud Collaborative Computing is a new stage of the cloud computing development. It integrates various forms of clouds with abstract method and realizes a more flexible and optimized configuration among mutiple clouds using an open-sharing cross-cloud integrated architecture. Currently,the Cross-Cloud Collaborative Computing is still confronted with many problems and challenges such as elasticity of service capability, convertibility of resource value, scalability of computing schema and ability of availablity reconfigurability. This project intends to use the cross-cloud integration and collaborative computing, and the adaptive resource optimizating schema based on the demand-respose mechanism, to build cross-cloud collaborative computing system with interoperable configuration among mutiple clouds. The research contents include the cross-cloud collaborative computing architecture that supports resource fusion and optimized sharing, the uniform modeling of resources in heterogeneous cloud platforms and cross-cloud sharing method, the adaptive resource optimizating schema based on the demand-respose mechanism and the cross-cloud interoperation arrangement and high availablility guaranteeing mechanism. This project will innovatively propose the cross-cloud seamless management mechanism under the heterogeneous cloud integrated schema, the cross-cloud sharing method based on the nested virtualization, the resource configuration and management based on the interoperable configuration among mutiple clouds, the adaptive resource optimizating schema based on the demand-respose mechanism and the high availablility guaranteeing mechanism adopting the technology of cross-cloud reconfiguration. The research results will improve the ability of cross-cloud resource fustion and sharing as well as the collaborative operation level for the next generation cloud systems.

跨云协同计算是云计算发展的新阶段，其采用抽象方式整合各种云形态，通过开放共享的跨云融合架构，实现云服务更加灵活与优化的多方配置。目前跨云协同计算还面临着服务能力可伸缩、资源价值可转换、计算模式可扩展以及可用能力可重构等诸多问题与挑战。本项目拟采用跨云资源融合与协同计算的思想，基于需求-响应机制的资源自适应优化模式，通过多方互操作配置方法构建跨云协同计算系统。研究内容包括：支持资源融合与优化共享的跨云协同计算架构；异构云平台资源统一建模与跨云共享方法；基于需求-响应机制的跨云资源自适应优化方法；跨云互操作编排与高可用保障机制。本项目将创新性地提出异构云资源融合模式下的跨云无缝管理机制、基于嵌套虚拟化的跨云共享方法、基于多方互操作的跨云配置与管理、基于需求-响应机制的资源自适应优化方法以及运用跨云重构技术的高可用保障机制。研究成果将提升新一代云系统跨云资源融合与共享的能力以及协同运行的水平。

跨云协同计算是云计算发展的新阶段，其采用抽象方式整合各种云形态，通过开放共享的跨云融合架构，实现云服务更加灵活与优化的多方配置。目前跨云协同计算还面临着服务能力可伸缩、资源价值可转换、计算模式可扩展以及可用能力可重构等诸多问题与挑战。本项目采用跨云资源融合与协同计算的思想，基于需求-响应机制的资源自适应优化模式，通过多方互操作配置方法构建跨云协同计算系统。研究内容包括：支持资源融合与优化共享的跨云协同计算架构；异构云平台资源统一建模与跨云共享方法；基于需求-响应机制的跨云资源自适应优化方法；跨云互操作编排与高可用保障机制。通过四年的研究取得的研究成果如下：提出了面向跨云协同优化云服务的调度方法、跨云平台核心技术存储设备虚拟化技术、跨云协同计算的资源共享与收益优化方法、基于存算传融合的分散式计算架构及其优化方法、云平台核心技术高效能可调度NVMe虚拟化架构、基于NVMe存储的虚拟化性能隔离和分级服务方法、基于NVM混合内存的自适应前缀树构建方法。本项目的支持下发表论文17篇，申请国家发明专利4项、美国专利1项。