Tim Bass
11-30-2008 06:18 AM
A complex system is defined as a system composed of related components that as a whole exhibit one or more properties not obvious from the easily observed properties of the individual parts.* This is certainly true of the CEP notion of the “event cloud” in network systems. * A modern energy or telecommunications network is composed of many systems and, to the casual observer, the properties of the individual events seem disconnected and unrelated.* Making sense of the myriad seemingly disconnected components and causal relationship in the “event cloud” is the purpose of complex event processing.
The complexity of a networked system may be of one of two forms:
disorganized complexity and organized complexity. In a nutshell, disorganized complexity is the situation of a very large number of disparate, but possibly related, components.* Organized complexity is the situation of systems exhibiting emergent properties.** Examples of complex systems include biological systems such as ant colonies cells, living things, human beings and nervous systems.* In fact, many systems of interest to humans are complex systems, both natural (climate, for example), biological systems (ant colonies, social networks) and man-made complex systems (telecommunications networks).
Complex event processing is the machine-machine and machine-human process of trying to make sense out of difficult to observe network-centric situations, both opportunties and threats, inherent in networked systems, as discussed in our eight part series,
What is Complex Event Processing?
CEP is, therefore, like a type of network-centric microscope.* For example, when we look at a complex system such as a lake with our naked eyes, we are only able to observe a fraction of the system properties.* We see the surface of the lake, surface plant life and fish, and perhaps a boat with a fisherman.* We see the clouds and the sun and other related complex systems.* However, with our naked eyes we can not understand the complexity of the vast majority of activity in the lake.* Nor can we understand, with our naked eyes, the relationships between seeminly disconnected complex systems.
The same is true of man-made network systems.** With our naked eyes we see routers, hubs, switches,* and servers.* We see log files and computer screens.** We see performance graphs and visual results of constructed, but limited, queries.* However, we cannot see the myriad causal relationships in the network with our naked eyes.* What is required is the capability to processs the events within the complex system.** That capability is not a single technology.* That capability is what we call “complex event processing”,* or CEP.
Earlier I posed the question,
What Defines Complexity in Rule Processing? and received no responses from the many “rules experts” who are kind enough to frequent this corner of cyberspace.* In addition, I asked the rhetorical question,
Should We Simply Rename CEP BRMS? because the capability of the self-described CEP vendors tend to mirror the capabilities of business rule management systems, relatively speaking.
What is required in the evolution of our critical understanding of telecommuncations and data networks as a complex system is far beyond what we are seeing in the self-described CEP commercial marketplace.** What customers require is the capability to process complex events in complex systems.* This capability is what we call “CEP”, decribed in
The Genesis of Complex Event Processing: Asymmetric Capabilities, CEP, Event Noise and Asymmetric Event Processing and
The Motivation Behind Adaptive Analytics and CEP.
Source...
