W. Kießling and B. Hafenrichter (Germany)
Preference query optimization, preference algebra
Personalization of Web services requires a powerful pref erence model that smoothly and efficiently integrates with standard database query languages. We make the case for preferences as strict partial orders, supported in database query languages like Preference SQL and Preference XPATH. Performance of Web services will crucially de pend on various architectural design decisions. We point out that, given current bandwidth restrictions, a central server architecture is desirable. Concerning the implemen tation of preference queries we investigate the tightly cou pled architecture, presenting a novel approach for alge braic optimization based on preference algebra. We pro vide new transformation laws and give evidence for the power of this heuristic optimization technique. This forms the basis for a new preference query optimization meth odology, promising good performance even for complex Web services.
Important Links:
Go Back
