Investigating access to heterogeneous storage systems using linked data in UNICORE Grid middleware

The Grid provides access to execution and storage resources - through middleware - by normalizing the access to resources. UNICORE calls this quality dasiaseamlessnesspsila. One result of this is that a request for some action at one resource can, with relative ease, be re-targeted to another, making possible the coordinated usage of multiple resources. In this paper we describe a seamless interface to heterogeneous storage systems, using the (Semantic) Web as middleware for the Grid. Data is accumulated by experiments and simulations running on execution systems, instruments and sensors, and by importing data from other locations. Storage systems are regular file systems connected to the execution systems, and also distributed storage systems such as iRODS, or Cloud storage such as Amazon S3. All should be available seamlessly. Given increasing volumes, sophisticated tools to help manage data are needed. Furthermore, following the Grid idea, we should be able to perform this across multiple storage systems in a coordinated manner. By Web-enabling e-infrastructure it will make it significantly easier and more accessible for both users and developers wanting to develop new applications for users, and will provide improved functionality for eScientists in key areas such as global data management, search, metadata and collaboration support. The message is that ubiquity and simplicity is better. Semantic Web technologies and the principals of LinkedData are used to describe resources and relationships with other resources, both generally and for specific domains. Detailed information can be extracted by navigating and querying the resulting graph. In addition collaborative capabilities are built into the middleware allowing resources to be selectively (or indiscriminately) shared.