2003•Unpublished venueRequires access
A-mediAS
Annika Hinze
Open publisher page 10 citations
Abstract
Event Notification Services inform about the occurrences of events that are of special interest for their users. Upcoming systems for event notification cover multiple applications and integrate event data from different sources. A Main challenge is the evaluation of user profiles under various and changing application requirements. Of particular interest are (1) the profile evaluation using differing semantics and (2) high filter performance under changing system load. This paper introduces the design and implementation of A-mediAS -- an adaptive event notification system. We present a case study of a remote monitoring application, which shows the effective adaptation of the evaluation semantics in A-mediAS. We describe different event filtering strategies and show their adaptation to varying event and profile distributions. Finally, we analyze the filter performance for different workload scenarios.