Saturday, November 1, 2008

On the head and the tail of EDA

Somehow the discussion around "EDA" and "CEP" continues in the Blogsphere - Giles Nilson from Apama has published seven points , out of which I quite agree to the first five. Jack van Hoof, who started this whole thread of discussion, argues that "CEP is not the beginning but finishing of EDA". So what is the answer? head or tail ? the right answer is --- it depends. Some customers work with methodic architecture way, in which first the EDA architecture is being set up, and then CEP tools are invoked on top of this architecture -- this is the case that Jack is talking about. However, in some cases, customers don't apply architectural thinking, but just acquiring some application that is implemented with CEP tool, and thus it introduces a kind of EDA ad-hoc for a specific application, this is the case that Jiles is talking about.

My guess is that most early adopters have applied CEP in ad-hoc way, so it serves as a "head" for more EDA in the future, while recently we see more customers looking at EDA first, since they are taking it from enterprise architecture perspective.

If we'll take the CEP functionality -- finding patterns on multiple events, it may not be implemented on EDA at all -- since the same technique may fit to "non event" environment.

More - Later.


Jack van Hoof said...

Hi Opher,

Thanks for joining the debate.

You say that "in some cases, customers don't apply architectural thinking". That's exactly what the discussion is about: applying CEP without any architectural context (which is possible) is not EDA. Mind that EDA is the acronym for Event-driven ARCHITECTURE.


Opher Etzion said...

Hi Jack. What I have tried to point out is that introduction of CEP application may entail introduction of event driven architecture that will be used for other purposs, but this is kind of bottom-up approach.



Jack van Hoof said...

I see what you mean, Opher. That sounds fair. It's even the approach we are using in my practice ;-)


harvey said...


We are encountering this "two sides" of the coin in our contribution to an EPTS workgroup ("Challenges"). The CEP portion and the EDA portion are just two ways to look at the very same thing in a fully realized implementation. So I agree they are just two different starting points.

The interesting difference is:
-- In CEP the organizational "owner" is likely to be a single entity (or tight coalition) for example a "fusion center" using government terminology.
-- In EDA the feeds if you will for the CEP can cross many organizational boundaries (national, provincial/state, local, tribal, etc.)

Thus the act of building and managing the CEP portion and EDA portion will feel different, but they really are part of the same thing.

Take care,