Flexible services for the support of research.
Flexible services for the support of research.
复制标题
支持研究的灵活服务。
DOI：
10.1098/rsta.2012.0067
复制
发表时间：
2013
期刊：
影响因子：
--
通讯作者：
Turilli M
中科院分区：
文献类型：
--
作者：
Turilli M
用户和提供商越来越多地采用云计算来促进对计算资源的灵活、可扩展和量身定制的访问。尽管如此，这种范式的巩固也暴露了它的一些局限性。云计算最初是由直接控制大量计算资源的公司设计的，现在得到了资源有限或对这些资源有更明确、更不直接控制的组织的支持。这些组织面临的挑战是在处理有限且通常分布广泛的计算资源的同时利用云计算的好处。本研究的重点是高等教育机构对云计算的采用，并解决了两个主要问题：灵活和按需访问大量存储资源，以及跨异构云基础设施集的可扩展性。建议的解决方案利用一种联合的方法来获取云资源，在这种方法中，用户可以通过高度可定制的代理层访问多个基本上独立的云基础设施。这种方法支持统一的身份验证和授权基础设施、细粒度的策略规范以及会计和监视的聚合。在松散耦合的云基础设施联盟中，用户可以访问大量数据，而无需跨云基础设施复制数据，并且可以在本地云资源不足时扩展其资源供应。
Cloud computing has been increasingly adopted by users and providers to promote a flexible, scalable and tailored access to computing resources. Nonetheless, the consolidation of this paradigm has uncovered some of its limitations. Initially devised by corporations with direct control over large amounts of computational resources, cloud computing is now being endorsed by organizations with limited resources or with a more articulated, less direct control over these resources. The challenge for these organizations is to leverage the benefits of cloud computing while dealing with limited and often widely distributed computing resources. This study focuses on the adoption of cloud computing by higher education institutions and addresses two main issues: flexible and on-demand access to a large amount of storage resources, and scalability across a heterogeneous set of cloud infrastructures. The proposed solutions leverage a federated approach to cloud resources in which users access multiple and largely independent cloud infrastructures through a highly customizable broker layer. This approach allows for a uniform authentication and authorization infrastructure, a fine-grained policy specification and the aggregation of accounting and monitoring. Within a loosely coupled federation of cloud infrastructures, users can access vast amount of data without copying them across cloud infrastructures and can scale their resource provisions when the local cloud resources become insufficient.