Self-Configuring Dynamic Networks

Self-configuring dynamic networks automate configuration management. This concept will allow self-configuring nodes to adapt their configurations based on the requirements of the environment. The goal of self-configuring networks is to create programmable configuration changes based on organizational configuration policies. This topic has been the subject of many research papers and university projects.

Normally, networks are configured and maintained by skilled administrators. Automation will increase configuration efficiency and productivity and minimize errors and costs. However, some problems can arise when automating tasks:

The change propagation problem
  • Different elements, their relationships, and configuration states must be recognized
  • The knowledge of the sequence of changes in these elements must be represented
  • Each element must be able to be changed through varying proprietary instrumentations, configuration tools, and operations procedures
  • Recovery and undoing of changes in case of failures must be enabled
  • Network topology discovery
  • Change propagation rules
  • Element heterogeneity
  • Recoverability in case of failures
The configuration policy problem
  • Configuration states are inconsistent resulting in operational failures and inefficiencies
  • Understand policy and configuration relationships
  • Enforce policy constraints to assure consistent configurations
  • Enable organizations to program policy constraints to effect operational policies
The composition problem
  • Adaption to change propagation rules and policy constraints needs to occur
  • The self-configuration management software propagates changes to the configuration model. For example, upon discovery of a new source, the software will indicate the relationship to the active nodes and configuration changes will occur.
Table derived from Konstantinou, A. V., Florissi, D., & Yemini, Y. (2002). Toward self-configuring networks.