|
|
article
page | 1 | 2 | 3 | 4 | 5 | 6 | 7 | 8 | 9 |
Engaging the full power of Complex Systems Dynamics
Dr. Daniele Miorandi of CREATE-NET & BIONETS writes:
We make “a case to go for biology as the "source of inspiration" for breaking such complexity ceiling. The basic idea is to move from the conventional top-down engineering design approach (from use-cases to specifications, then to architecture, implementation and testing, all based on the "divide et impera" paradigm) to a bottom-up approach, in which the "intelligence" lies in the design of an artificial ecosystem, embedding the principle underpinning evolution in simple components, the expected behavior arising from a complex web of interactions.” [Proc. of IEEE SMC (http://www.smc2007.org/)]
Further he states,
“Concerning self-healing networks, my opinion is that - again - nature represents one of the most promising directions to look at. The basic idea shall be here to design reactive systems, where the reaction patterns is not defined at the system design phase, but shall be able to adapt and evolve (again, autonomically) over time.”
Fundamentally, study of complex communications has two different solution approaches. For example, CASCADAS, seeking the maximum level of autonomous behavior, postulates that each component unit must embody within itself the self* characteristics. They seek systems with autonomous components. The Fine Grain approach, shared by developing Grid standards and open source projects like RIO, is to embed the self-* characteristics in the matrix of the system interactions. No component is really autonomous; instead the system as a whole is autonomous, as the structures and DNA of the services enforce collaboration to achieve practical self-* applications. This style, we know how to do today.
|
|
The futuristic, fully autonomous component is still beyond practical application, but this is no reason to wait for self-adaptive, semantic, policy-driven systems. The first Java application servers were barely able to run a single service; systems improve over time because bright developers and architects see better approaches. Different projects and companies cross fertilize each other. Operations groups identify issues and flag them for priority solution, thereby focusing resources toward generating new behaviors for the whole system. The evolution of software is itself a complex system – one that is adaptive, but it takes leadership to turn that adaptation toward deploying self-* systems.
Leadership for Survival
Rudy Puryear pegs the central issue: “IT reflects the complexity that's been built into business. Unnecessary complexity in the business has resulted in a lot of unnecessary complexity in IT. That's slowing down the cycle times in IT. The cycle times in IT are increasingly much slower and out of sync with the cycle times required by the business to stay competitive.” [CIO]
However, there is good complexity and bad complexity – based on how an organization harnesses its complex systems. If you are only reactive, you are bound into system eddies, facing the same problems over and over, or worse you become victim to Darwinian-like negative selection. If you look again at the world and your place in the market as an “ordered system” reflecting complexity, you can begin to develop strategies for success. Every time a new network paradigm or a new class of service product is created, this becomes an opportunity to explore and deploy self-healing systems. To date, our telecommunications ecosystem has not made this commitment to be truly healing. Supporting self-* companies by buying and deploying their products is a step toward reducing costs and controlling the ship-of-market-state. Then networks and support systems will make an evolutionary leap to a new plateau.
article
page | 1 | 2 | 3 | 4 | 5 | 6 | 7 | 8 | 9 |
|
|
|
|
