vincent
07-15-2008 06:44 PM
… was the title chosen by OMG for the CEP round table [*1] discussion at their Workshop on Distributed Object Computing for Real-time and Embedded Systems this week. “CEP vs hype” has been somewhat an over-discussed (but ultimately uninteresting) topic on various blogs of late; the point here at OMG’s workshop was to raise awareness and solicit new feedback from the real-time community.

Many thanks to the round table participants, of whom:

• most were from the defense industry in Europe and the US
• most had not come across CEP before (which may indicate an unprecedented - for the software business - lack of hype in the CEP business)
• some commented that commercial CEP tooling was an opportunity for the defense community to more easily, and cost-effectively, deliver “situation awareness” solutions
• some had insights on positive factors for the CEP market, including:
  • the 2 main CEP (pattern-) processing mechanisms: rule-based and continuous query-based
  • … and the ability to easily extend CEP tools to handle new 3rd party or own-build event processing algorithms
  • the availability of a wide range of event adapters / middleware plug-ins
• … and also observed negative factors for the CEP market, including:
  • no standard definition of “CEP capability/capabilities” across CEP tools
• expressed some surprise at the lack of “embedded real-time” CEP tooling (for embedded control systems [*2]).

And the result? The overriding vote was for CEP being “useful”; no-one thought it was mere “hype”. My vote? I prefer croissants, so CEP is certainly a next best thing…

Notes:

[1] The round table format works really well, and should be considered for contentious issues at the next EPTS meeting.

[2] Of course, if any embedded control system provider would like to discuss the deployment of distributed CEP capabilities to their favorite real-time OS, please drop us a line!



Source...