A Collaborative Inter Data Grids Strong Semantic Model with Hybrid Namespace

Abstract

The intention is to design a hybrid namespace collaboration model for existing organizations to pay lower cost (initial and running) in order to join it. Data resources will be allowed to have colliding names. The contributing organization will be free to set any rules for its internal users to follow (such as taxonomy). However, it will have to take care of a very simple interface to the intended model in order to introduce new contributed resources or to query external ones. Traditional data grids will join this model (represented by proxy agents) to provide their content while their owners can join as users of relevant servers to explore relevant resources to keep their content up to date. A hierarchical namespace will be maintained in order to uniquely identify data resources. (Full names are the local names prefixed with the ID of owner data grid for uniqueness.) If the full name is queried, the corresponding resource is sought in the wide group of collaborating data grids. If the local name is queried, it is fetched in the local data grid. Similar and related topics in different places in the hierarchy will be linked for robustness. Otherwise, a comprehensive search should be performed all over the whole hierarchy