2nd Event Processing Symposium – Summary

Following the 1st Event processing symposium that IBM hosted in March 2006, and the establishment of three workgroups, Oracle hosted the 2nd Event Processing Symposium in November 2006. The symposium included: two keynote talks about “Data Mining and Event processing”, and about “Event Processing in the Global information Grid”; two panels about standards and about benchmarks, report of the three work groups, presentation of 17 use cases, and some general discussions.

View the complete schedule and download presentations here 

Here are the decisions and action items.

1. We shall continue to act as a technical society – the name we adopted is “Event Processing Technical Society” (EPTS), which consists of collection of stakeholders in this area – vendors, academic people, analysts and any other person interested in promoting the area. The goal of this technical society is to promote the understanding of the “event processing” area and its potential impact, sponsor community-wide workgroups, encourage academic research in this area, sponsor meetings and conferences, and serve as a catalyst to the growth and understanding of the area by the general technical and business community. The technical society does not intend to develop into a standard organization and will work with various standards organizations when results of its work groups will mature.

2. The work will continue in four work groups, three of them will be continuation of existing work groups and one is new.

a. The terminology workgroup (David Luckham and Roy Schulte) has published its glossary draft. This site will also be the vehicle for the community to post comments and discuss terms. A deadline for posting comments (probably around mid January 2007) will be announced. At that time the comments will be considered by the workgroup, a new version will be produced and put for approval in the EPTS steering committee. After approval it will get the stamp of the consensus glossary, and will be taken as a proposal for one of the standard organizations.

b. The reference architecture workgroup (so far: Tim Bass and Opher Etzion) – Several reference architectures were submitted and discussed, and this workgroup will issue a call for more submissions. It was also decided to extend the scope of this workgroup to determine what is the underlying functionality behind event processing (“the platform independent model”), as part of this work we shall advertise call for use cases, where the use cases will be described as a structured questionnaire. We solicit volunteers to participate in this workgroup, and probably have a Face-2-Face workshop of this workgroup in spring 2007 to consolidate the input and work on concrete proposal that will be presented to the community. The workgroup activity can be found here.

c. The interoperability group (Bill Hobbib and Mark Palmer) will re-organize to deal with various topics:

• Related standards in the messaging area
• Event formats
• API’s for interoperability
• Relations of event processing and SOA; this topic was added at the meeting.

The issue of language will be discussed at later phase after we’ll advance in the understanding of the functionality requirements of the other workgroup

d. The awareness and research workgroup. This is a new workgroup. Initial members will be: David Cameron, Shailendra Mishra, Pedro Bizzaro, Beth Plale and Opher Etzion. This workgroup will work on the message to the wider community, awareness, and relations with the research world.

3. We have decided that it is a premature to have work on standard benchmarks in this area; however, we encourage research work into the current approaches taken by vendors towards measuring the performance.

4. The next symposium will be held in summer of fall of 2007, depends on the workgroup progress, and we assume that significant progress will be achieved by these four workgroups by that meeting. Anybody who wishes to participate or contribute to one of this workgroups is welcome
to address the appropriate workgroup leader.

5. This is also a reminder that we’ll hold a Dagstuhl seminar in May 2007. This will be one of the places in which we can do in-depth discussions to help us progress in our understanding of the event processing area, by bringing together people from industry and academia. If you have received an invitation and did not respond – please do, if you want to participate, please let me know and we’ll put you in the waiting list (typically some percentage of the original invitees decline, and there is a second wave of invitation).

I would like to thank my colleagues who took upon themselves to lead the workgroups, those who moderated the various sessions, all the presenters of use cases, all the panel participants and the hosts – Oracle Corporation in general, and our colleagues Dieter Gawlick and Shailendra Mishra and
particular.