Call for Papers
Goal
The
purpose of this workshop is to bring together people from academia and industry
interested in the foundations, implementation, and application of all aspects
of event-based middleware, from traditional topic-based and content-based publish-subscribe,
to event correlation, streaming queries, mediations, and systems integration.
We seek contributions from practitioners in industry and government, as well as
from academic and industrial researchers.
Workshop
Theme
Event-based systems are those in which information provided by producers
is distributed in a timely manner to interested consumers via messaging middleware.
A common service interface provided by event-based systems is the publish-subscribe
paradigm:
- producers and consumers can be mutually anonymous,
- producers
deliver events to topics,
- each consumer specifies interest by means of
a "subscription", and later receives events of interest via event notifications
The
middleware is responsible for consolidating subscriptions and propagating events.
Queuing and hybrids of publish-subscribe and queuing are also popular. Event-based
systems can be used to integrate a wide range of components into a loosely-coupled
distributed system.
For example,
- event producers can be application
components, post-commit triggers in a database, sensors, or system monitors;
- event
consumers can be application components, device controllers, databases, workflow
queues, etc.
These systems are seeing increasingly widespread use,
in applications ranging from time-critical systems, system management and control,
to e-commerce. Publish-subscribe services have been incorporated into standards
such as CORBA and JMS, and into commercial systems, such as offerings of IBM and
TIBCO.
Traditional event-based systems support subscriptions based on topics,
or based on filter predicates on message content; the middleware service delivers
copies of published messages to some subset of all subscribers. An emerging new
area of interest is the extension of these services to include transformation,
aggregation, and correlation of events. For example, a subscription to published
events of stock trades and offers to trade may ask for
- a continuous
running average of stock trades on a certain company; or
- alerts indicating
that more than 100 trades for the same stock issue have been made in the same
hour; or
- notifications that a request to buy and a request to sell have
been made on the same issue with matching offer price ranges.
As
with traditional publish-subscribe, it is still desired that
- publishers
and subscribers remain anonymous,
- subscribers are independent and can
possibly dynamically enter and leave,
- the system should push events from
publishers to subscribers as quickly as possible, and
- the implementation
should scale to large numbers of clients.
Many applications of this
nature are built today by combining publish-subscribe systems with other client-side
systems to perform the additional computation. The interest in these problems
overlaps with other system areas -- specifically streaming systems, continuous
query and continuous view update systems in databases, correlation engines, and
system monitoring and management tools. It is our goal to provide cross-fertilization
between researchers and practitioners in these areas with researchers and practitioners
in traditional event-based middleware.
The purpose of this workshop is
to bring together people from academia and industry interested in the foundations,
implementation, and application of all aspects of event-based middleware, from
traditional topic-based and content-based publish-subscribe, to event correlation,
streaming queries, mediations, and systems integration. We seek contributions
from practitioners in industry and government, as well as from academic and industrial
researchers.
Papers do not have to be based on complete and comprehensive
works. In fact, we welcome position papers, requirements for real-world applications,
as well as papers based on preliminary results, provided that they are forward-looking
and that they remain well-argued and justified in terms of existing work.
Workshop
Topics
The goal of this workshop is to share and discuss original and innovative
ideas in the area of event-based systems. Therefore, we invite authors interested
in event-based applications, event-based infrastructures, and event correlation
and transformation to submit a paper to this workshop. Below is a list of possible
topics of interest grouped by areas. (The list should not be seen as exhaustive.)
Event-Based Models and Paradigms
- Query/advertise vs. publish/subscribe
vs. queueing vs. hybrids
- Event correlation languages
- Streaming/continuous
query languages
- Event schemas and type systems
- Supporting XML
documents in queries and transforms
- Qualities of ServiceCorrectness guarantees
- Security
specifications
- Application interfaces
- End-user programming
- Tools
for specification and verification
Middleware Infrastructure
- Architectures
for distributed event-based infrastructures
- Content-based routing algorithms
- Fast
event matching algorithms
- Fault-tolerance, reliability, and recovery
- Performance
monitoring and system management of event-based middleware
- Dynamic reconfiguration:
load balancing and rebalancing, subscription propagation
- Implementing
streaming queries, transforms, or event correlation engines
- Consolidating
subscriptions, rules, streaming queries
- Performance evaluation of event-based
infrastructures
Integration
- Domain-specific deployment
of the event-based paradigm (e.g., workflow management systems, mobile computing,
pervasive and ubiquitous computing, sensor networks, user interfaces, components
integration, web services, embedded systems)
- Attaching event-based systems
to:
- Point-to-point communication systems
- Browser interfaces
- Application
GUIs
- Databases
- Federating event-based systems
- Converting
formats and protocols
- Security and privacy issues
- Mobility issues
User Experience and Requirements
- Real-world application
deployments that use event-based middleware: types of subscriptions, types of
pre- and post-processing, languages used, number of subscribers and publishers,
message rate and latency requirements
- Combining event-based systems with
other systems in a total solution
- Requirements from future event-based
applications
Submission Instructions
Submissions
must conform to the ICDCS
formatting guidelines (follow the instructions for "Formatting your Paper"
but not for "Submitting your Paper") and must not exceed six (6) pages,
including all text, references, appendices, and figures. Submissions must be in
Portable Document Format (PDF).
Please submit your paper by e-mail to
robstrom@us.ibm.com. Your submission e-mail must contain the PDF file as a
MIME attachment. The sender of the submission will be the contact person, unless
otherwise requested in the submission.
The submission deadline is 11:59
PM EST, January 24, 2005. Click here for other important
dates.
Authors of accepted papers are expected to participate in the workshop.
Location
The workshop will be held in Columbus,
Ohio as a co-located workshop of the 25th International Conference on Distributed
Computing Systems ICDCS 2005 (http://www.cse.ohio-state.edu/icdcs05/).
Workshop
Co-Chairs, Organizers and Program Committee
The workshop co-chairs are:
- Juergen
Dingel, School of Computing, Queen's University, Kingston, Ontario, Canada
- Rob
Strom, IBM Thomas J. Watson Research Center, Yorktown Heights, NY, USA
Contact
details for the co-chairs, as well as for additional organizers and the program
committee are listed on the Organization page.