Distributed Self Management for Distributed Security Systems

TL;DR

This paper explores distributed self-management in security systems, introducing and analyzing two neighbor-based information protocols, and testing their effectiveness within an artificial immune system for network protection.

Contribution

It proposes and evaluates two novel neighbor-information protocols to enhance distributed system performance without central control.

Findings

Neighbor information improves system efficiency

Protocols outperform centralized approaches in certain scenarios

Artificial immune system SANA effectively uses these protocols

Abstract

Distributed system as e.g. artificial immune systems, complex adaptive systems, or multi-agent systems are widely used in Computer Science, e.g. for network security, optimisations, or simulations. In these systems, small entities move through the network and perform certain tasks. At some time, the entities move to another place and require therefore information where to move is most profitable. Common used systems do not provide any information or use a centralised approach where a center delegates the entities. This article discusses whether small information about the neighbours enhances the performance of the overall system or not. Therefore, two information-protocols are introduced and analysed. In addition, the protocols are implemented and tested using the artificial immune system SANA that protects a network against intrusions.