Comparative analysis of different autonomic management architectures
Sławomir Kukliński · International Teletraffic Congress · 2012
The concept of Autonomic Network Management (ANM) was born several years ago. It comes with the promise of automated network management that provides proper network configuration, fault management (self-healing) and network optimization. In opposite to existing management approaches ANMs is based on real-time or near-real-time operations. Moreover, in order to provide scalability and fast reaction time, a distributed management approach is preferred over a classical centralized one. The ANM concept is of premium importance for network operators due to the elimination of human assisted management (human configuration errors probability is significantly minimized), OPEX reduction and ability to cope with continuously growing number of networking devices that have to be managed (including those installed at user premises). The existing Autonomic Network Management (ANM) approaches are typically based on the IBM Autonomic Management model. Such a feedback loop system consists of sensors (monitoring part), knowledge base (algorithms) and actuators (control knobs). In the cognitive approach the algorithms have learning ability.