At the upcoming stream processing symposium (November 7-9 in Redwood City), a workshop will be held to discuss the topic of Interoperability Standards, and I am writing on behalf of this session's working group to request your input towards this program.

Below you will find an abstract for the session, as well as the specific call for input.  All members of the community are invited to submit input by the deadline of Friday, October 6 at 6pm EDT.  The results will be presented back at the November program.

INTEROPERABILITY WORKSHOP ABSTRACT

Interoperability has been one of the thorniest problems in enterprise-level systems over the last few decades.  In general, it refers to the integration of multiple, possibly heterogeneous, systems.  If credible solutions are to be delivered in the CEP space, we must ask to what extent our solutions can be made interoperable.

Interoperability is a very large topic that can mean different things at different levels.  Some of these things include:

1. Reliable message delivery protocol.  (e.g. JMS, MQSeries)
2. Data formats / wire protocols.  The format of incoming data. (e.g. XML)
3. API’s.  The way in which one system invokes the services of another (e.g. JDBC)
4. Processing language.  The programming language that is used to express complex processing on stored and streaming data. (e.g. SQL, C++, Java)

In each of these areas, interoperability can be achieved by one of two means: Translation or Standards.

If components are free to use their own formats, API’s, and languages, then translation must be provided between all pairs of components.  Often, the current state of the art requires interposing specialized adapters or translation mechanisms at each interaction point.  Technology to automate or facilitate adapter construction is one approach to solving this problem.  Nonetheless, translation can still be very expensive.  A better solution is to agree on a standard that is adopted by all parties.  While standards eliminate the expense of translation it can be very difficult to reach consensus when a technology is not mature.

This session will explore the space of interoperability options for CEP.  It will go into some depth for those areas that seem most promising and for which the lack of a solution is seen as an impediment to wide-scale adoption.

REQUEST FOR INPUT

1)     Area(s) for focus: In which of the four areas named above do you feel the absence of standards is inhibiting wide-scale CEP technology adoption—and where additional effort to advance standards would be important for the growth of the CEP marketplace?   Are there additional areas not listed where you feel standards would be important?

2)     In the area(s) for focus identified in your response to (1), what are the needs and requirements?   Please be as specific as possible regarding protocols or standards you feel should be in place which do not yet exist today.  (e.g. If you feel standard data formats are needed, you might say XML is or is not the right interchange standard; or in the area of languages, you might say the community needs to reconcile differences in how time is modeled and in how this impacts pattern-matching primitives).

Input is preferred via email, Word, or Powerpoint documents.   Please send submissions to CEP-interoperability@yahoogroups.com.  You can also subscribe to the newsgroup by sending email to subscribe-CEP-interoperability@yahoogroups.com.  The results will be compiled and reported back at the November session. Thank you for your consideration and for taking the time to provide input towards this community effort.

Regards,

Bill Hobbib (on behalf of the working group)

bill.hobbib@...