Wednesday, May 26, 2010

On event processing standards


One of the Dagstuhl follow-up will be to have action items in advancing standards in the event processing world. One of the topics discussed in Dagstuhl was standards, The team working on it was moderated by Paul Vincent, who already blogged about it. While I'll write more about it when the final report will be ready, here are some initial thoughts:

  1. It seems that in the era where the vendor community is now dominated by bigger companies, the atmosphere for standards become more friendly.
  2. For other communities - standards have been critical success factor, e.g. web services.
  3. Somebody mentioned the immortal Stonbraker's phrase about SQL being "intergalactic data speak", we need the "intergalactic event speak" - and it is not an extension to SQL.
  4. There are different standardization issues -- event representation, meta-modelling, event processing language; as well as extensions to many existing standards possible.
  5. The language standardization will be trickiest - due to the variety of languages styles exist, here I think that we'll better start with a language at the PIM (platform independent model) level. In the EPIA book we provided one that can serve as a starting point.

I believe that standards is one of the important ways to go forward, and will write more about it when we'll have the Dagstuhl final report.

No comments: