10 June
2005, Columbus, Ohio, USA
in
conjuction with the
25th International Conference on Distributed Computing
Systems
ICDCS 2005 (6-9
June 2005)
Program
The workshop is organized as series of discussion
sessions, each one revolving around a specific research topic. Discussions will
be introduced and stimulated by presentations.
The final program is now available:
Brief Description
Event-based systems are systems in which
producers deliver events, and in which messaging middleware delivers events to
consumers based upon their previously specified interest. One typical usage pattern
of such systems is the publish-subscribe paradigm, in which producers and consumers
are mutually anonymous, and in which producers deliver events to topics, and each
subscriber separately and independently registers a specification of its particular
interest; the system then receives events from producers and in a timely fashion
propagates (possibly transforms) and delivers these events. A second popular event-based
usage paradigm is queuing; hybrids of publish-subscribe and queuing are also possible.
These paradigms are popular because of their ease of use and high degree of decoupling,
and because efficient and highly scalable implementations are available.
Event-based
systems are widely used for integrating loosely coupled application components,
including sensors, device controllers, and databases. The use of event-based systems
is expanding, and an increasing number of commercial systems are offering messaging
middleware components. There is a wide interest in techniques both for extending
the capabilities and for improving the performance and ease of use of such systems.
Traditionally, the DEBS workshop has covered all aspects of the design,
implementation, performance analysis, and application and deployment of distributed
event-based systems. We propose to continue soliciting submissions relating to
these core topics, and also to encourage submissions relating to (1) user experience
and user requirements in realistic application scenarios from industry and government,
and (2) requirements and solutions that extend the traditional event-based paradigms
(publish-subscribe and queuing) beyond the standard content-based filtering, to
include also transforms, aggregations, event mediations, and event-correlation.
We specifically seek to encourage cross-fertilizations with such research areas
as stream processing, continuous queries, continuous view updates, and rule-based
event correlation. We wish to avoid, for example, the database communities and
the messaging communities independently studying and solving overlapping problems
and developing separate architectures and vocabularies for related concepts. We
seek a balance between research studies relating to design and implementation
of such systems, and presentations of either actual experience with the use of
such systems for a real problem, or requirements for future systems to solve projected
real problems. Researchers need to make sure they are studying important problems;
potential users need to understand the directions in evolving technologies. Since
this is a workshop, we encourage position paper and works in progress rather than
polished results.