|
|
article
page | 1 | 2 | 3 | 4 | 5 | 6 | 7 | 8 | 9 |
-
Most often cited as the source archetype is the autonomic nervous system of the higher mammals – hence the most common name – Autonomic Networks. This model assumes a distributed hierarchal control system of sensors and patterned responses. It is the most comfortable pattern that we derive from nature because it is already familiar. Most networks today are hierarchal in their organization and their management. The notion of sensor/responder is parallel to our existing network management pattern of agent/manager.
|
|
To truly understand and manage our existing and future networks, we need to see them as complex, organized, evolving systems. |
|
Whatever pattern you pick from nature’s colors, it can be applied to the organization and architecture that a group uses in the design of an autonomic network. Each pattern has strengths and weaknesses. Each is a better or worse fit - for after all, networks are not natural systems. Fundamentally, natural systems reach stability via the interaction of evolutionary principles over long time spans. Despite our overloading of the meaning of the word ‘evolution’, there is no |
|
|
|
-
Social insects also provide a pattern. Here swarms and colonies of often related individuals produce quite complex and seemingly intelligent decisions based on the interaction of a much simpler set of behaviors. This pattern is used by humans seeking to build policy-controlled systems that are comprised on mostly homogeneous elements.
- Predator-prey, producer-consumer ecosystems are another biological example from which patterns can be abstracted and applied to network and service designs. The give and take, up and down flows of these systems are well matched to peer-to-peer systems; and many people think they can provide insight into the governing mechanisms for P2P systems - allowing the rules of engagement in the environment to provide indirect controls for the autonomous rules of interacting agents. This type of pattern bears a relationship to today’s economic systems and may model the interaction of commercial networks. It is the most complex pattern and relies heavily upon the identification and reinforcement of often unexpected emergent behaviors. Herding behavior of cows and sheep, or schooling of fish are such example patterns.
|
|
actual Evolution going on in networks and devices. They do not have or propagate DNA. Still, with caution and remembering that “ the map is not the territory”:
-
The autonomic nervous system is a good pattern for active management and control systems.
-
The Social insects are a good pattern for designing stable, well mannered, homogeneous networks of similar devices, for example, optical networks or routed IP networks.
-
Producer-consumer ecosystems can give us insights into how the larger, heterogeneous commercial network environments behave.
But the sheer complexity of producer-consumer ecosystems systems do not readily allow for ‘design’ of behavior. Instead behavior stabilities and abnormalities are observed and cataloged, using methods developed via Complexity Science, for use in predicting future behavior and indirectly influencing behavior by the tweaking of parameters in the environment.
Do autonomic networks exist today?
There are three voices to this. Most today say that architecting Autonomic Networks (AN) is
article
page | 1 | 2 | 3 | 4 | 5 | 6 | 7 | 8 | 9 |
|
|
|
|
|
