NADIA—Network Analysis Using Distributed Intelligent Agents

Architecture Technology Corporation (ATCorp) developed the architecture of NADIA and established its effectiveness using simulations. Subsequently, ATCorp demonstrated three powerful capabilities of NADIA during the Phase II prototyping effort:

1. a simulation-based tool for automatic composition of intelligent agents for network diagnosis;
2. cooperative, consensus-driven diagnosis of network faults using a collection of diverse and autonomous intelligent agents; and
3. a framework for implementing a distributed community of diagnostic agents that not only allows new types of agents to be easily integrated into the system, but also conserves network resources by allowing agents to share data and computing resources.

A NADIA-based fault-diagnosis system has the benefit of informing a network administrator not only that a potentially debilitating event has occurred on the network but can also tell the network administrator the cause of the problem. This is a significant deviation from current network-management systems that detect network anomalies, but not their causes. The benefit of such a system is that less-sophisticated network administrators can resolve problems more quickly.