Abstract

Network management is a complex task involving various challenges, such as the heterogeneity of the infrastructure or the information flood caused by billions of log messages from different systems and operated by different organizational units. All of these messages and systems may contain information relevant to other operational units. For example, in order to ensure reliable DSL connections for IPTV customers, optimal customer traffic path assignments for the current network state and traffic demands need to be evaluated. Currently reassignments are only manually performed during routine maintenance or as a response to reported problems. In this paper we present a decision support system for this task. In addition, the system predicts future possible demands and allows reconfigurations of a DSL access network before congestions may occur.

@INPROCEEDINGS{bsufka:2010:dsl,
  author = {Karsten Bsufka and Rainer Bye and Jo{"e}l Chinnow and Stephan Schmidt
	and Leonid Batyuk},
  title = {{Agent-Based Decision Support: A Case-Study on DSL Access Networks}},
  booktitle = {Proceedings of the Twenty-Second Innovative Applications of Artificial
	Intelligence Conference (IAAI-10)},
  year = {2010},
  pages = {1757-1762},
  publisher = {Association for the Advancement of Artificial Intelligence},
isbn = {978-1-57735-463-5}, 
}