Tim Bass
12-05-2008 05:19 AM
Continuing with our discussion on
Complex Systems and CEP, let’s turn our attention, monentarily, to more scientific, or perhaps philosophical, discussions. Let’s review a bit of
chaos theory via the Lorenz effect and talk about the “event soup”, a phrase I shamelessly coined in
On the Maturity of CEP.
Edward Lorenz was using a computer model to simulate weather when he rounded .506127 to .506 and the result was a vastly different weather scenario.* This motivated Lorenz and others to discuss, mathematically, how small, seemingly insignificant events can have profound effects over time.* Most of you have heard of “The Butterfly Effect”, the popular notion that a butterfly flapping their cute little wings can have a dramatic effect on weather patterns far away (in space and time); this is a metaphor for the Lorenz effect.* Let’s ground this concept with a similar metaphor, The Story of Amaldo.
The Story of Amaldo.
A man promised his wife that he would stop smoking.** Worried about his wife’s nagging, the man sneaks out the back door and walks down an alley behind his urban home.* When he stops to light his cigarette his eyes meet the eyes of a small alley cat.* The frightened alley cat runs under the building into a coven of alley cats who scatter in many different directions.* We follow one particular cat who jumps up on a nearly ledge of a neighbor’s bedroom window.** At the same time, a man and a woman are making love and they woman sees the cat in the window, she turns, and accidentally knocks over a lamp.** Her lover, overly aroused and a bit drunk, get angry and complains how she always ruins his mood.** They argue and there is no love making that night.** It so happens that because they did not make love and conceive a child, their son was not born, who fathered another son who turned out to be a great scientist, Amaldo, who (would have) discovered a cure for a devastating disease. Many people died because Amaldo was not born into this world.
For a lack of a more precise definition, let’s call Amaldo’s story
causality, or simply cause-and-effect.*** One of the factors that makes causality complex is that causality is vast and deeply inter-related. * For example, in our event-scenario above, we only followed one cat and a bit of causality of one cat’s journey and how it effected Alamdo.* It is easy to see that the cause-and-effect of events increases exponentially over time in most circumstances. Unfortunately, I don’t have a formal mathematical model at hand provide support to this claim.
Yet, without formal proof, I think most readers will agree that the universal set of events grows exponentially with each microsecond or nanosecond.** Every decision you make, every choice you pick, every mouse click, every stock transaction, is an event which can (and does) effect many lives.** Each decision you make follows the same principles as the Lorenz effect.* Each event also follows the same principles.* Naturally, some events are more “influential” or “consequential” than others and therefore have a larger effect.* My apologies for the lack of formality in this part of the discussion.* I hope this truth is self-evident without formal proof.
Events in computer networks have similar non-linear consequences.* The Lorenz effect was based on a simple computer rounding error.* Professor Luckham attempted to describe this in his formal CEP model as the “event cloud.”* It turns out Professor Luckham’s formal model of his event cloud, based on POSET theory, was in the right direction, albeit too simplistic.**** For event processing, and particular complex event processing, we might be better off if we think of the “event soup” versus the “event cloud”.
When we think about events in electronic networks,* some events have an obvious profound effect on our world.** Some events, similar to the Lorenz effect, are seemingly insignificant and have a great effect as well.** No matter how we view this, the shear volume of events is increasing exponentially over time, today.** This trend will continue indefinitely unless something very dramatic happens and our planet freezes over or some other global disaster happens!
Note: An interesting question not addressed here, but obviously important; is will the exponentially increasing event soup effect a future global disaster?
There is some interesting research to be done on applying complexity theory to events and event processing.* The very nature of the expanding event soup demands that other researchers look at complexity, causality, Lorenz effects and more, in the field of event processing.
We need more formal models in this area, and I suggest to you in this post that there is a future Nobel Prize winner out there who will build formal [complex event] processing models that will help us get our minds around the every growing complexity and space and time causality of the exponentially expanding electronic event soup.
Source...
