Negotiating Data Access in Federated Database Systems

Abstract:

The ever growing need for information is putting pressure on organizations to share data with their partners. However, although the different entities would like to share information, it is clear that each individual system administrator would like to preserve his or her control over the system. This concern with each system's autonomy has led to the notion of federated databases. Previous work in this area has touched upon the topic of negotiating access in a federated database (i.e., determining what local information may be access by any particular remote user), but we feel that the topic has not been studied in depth. In this paper we propose a new scheme, based on the notion of quasi-copies, which may be used for interaction among autonomous databases. We also provide protocols for access negotiation, as well as simple cost models for estimating the expense involved in allowing remote access to information. One of the main advantages of our new scheme is that it provides a very precise way of establishing how much autonomy is given up by the owner of the information when he or she decides to share data. Finally, our approach may be used in both the case where the databases systems in question have a common query language (or there exist facilities for query translation), and when they do not.