CAREER: Interoperation Among Heterogeneous Global-Scale Storage Systems

将存储系统推广为一门不同于数据库、操作系统和分布式系统的学科，这一共同主题将这个项目的研究和教育目标结合在一起。作为一门独立的学科，存储系统将在学术界获得与存储对计算的重要性相称的突出地位。存储限制了大多数应用程序的性能，拥有企业最有价值的资产，并消耗了大部分IT支出。培训存储专家为IT社区服务，并创建软件来提高存储的可管理性和性能，这是非常必要的。该项目的研究部分解决了大规模存储系统无法共享持久数据的问题。每个全球规模的系统都拥有数据，并将这些数据与无关的计算机隔离开来。该项目将开发一个联合协议，允许存储系统以交互方式迁移、共享和复制数据。联合不是一种访问协议；它不向用户和应用程序提供数据。相反，联合使存储服务器和数据存储能够传输和共享持久数据的管理。辛迪加将数据从软件系统中分离出来，这打破了在全球范围内访问数据和在全球任何地方访问数据之间的界限。这个项目的教育部分开发了一个课程来培训存储系统专家在八门课程的硕士学位。课程序列组织了特定于持久数据管理的知识，将其从许多其他学科中剔除。该课程将课堂作业与基于互联网的教育软件系统集成在一起，支持移动、临时和资源有限的教室。

The common theme of promoting storage systems as a discipline distinct from databases, operating systems, and distributed systems unites the research and educational goals of this project. As its own discipline, storage systems will attain prominence in academia commensurate with storage's importance to computing. Storage limits the performance of most applications, holds enterprises' most valuable assets, and consumes the majority of IT dollars spent. Training storage specialists to serve the IT community and creating software to increase the manageability and performance of storage is imperative.The research component of this project addresses the inability of large-scale storage systems to share persistent data. Each global-scale system owns data and isolates that data from unaffiliated computers. This project will develop a syndication protocol that allows storage systems to migrate, share, and replicate data interactively. Syndication is not an access protocol; it does not provide data to users and applications. Rather, syndication enables storage servers and data stores to transfer and share the management of persistent data. Syndication decouples data from software systems, which breaks down the boundaries between access to data on a global scale and access to data anywhere on the globe.The educational component of this project develops a curriculum to train storage systems specialists within an eight course Master's degree. The sequence of courses organizes knowledge specific to the management of persistent data, culling it out of many other disciplines. The curriculum integrates classwork with Internet-based educational software systems that support mobile, ad-hoc, and resource limited classrooms.