On Intelligent, Autonomous and Collaborative Agents to Manage Internet Routing Domains

This article describes an environment for knowledge acquisition, learning, use and collaboration inter agents over Internet Infrastructure. Four agent types are used in a previously applied fourtier model, such as the use case on the Internet Routing Registry. This model, which can be implemented in each Autonomous System domain of the Internet infrastructure, is integrated into an environment with (a) capturing information from unstructured databases, (b) creating and updating training bases appropriate to machine learning algorithms and (c) creation and feeding of a knowledge base. Such resources become readily available to agents in each domain and to agents in all other domains with the aim of making them autonomous. The agents collaborate and interact with each other, through individual blockchain structures that also take care of operational security and integration aspects. In addition, a test bed to validate the entire model, including the functionalities of the agents, is also proposed and characterized.