A Distributed Challenge Detection System for Resilient Networks

Abstract

The network has become essential to our daily life. With the increase in dependence, challenges to the normal operation of the network bear ever more severe consequences. Challenges include malicious attacks, misconfigurations, faults, and operational overloads. Understanding challenges is needed to build resilience mechanism. A crucial part of resilience strategy involves real-time detection of challenges, followed by identification to initiate appropriate remediation. We observe that the state-of-art to challenge detection is insufficient. Our goal is to advocate a new autonomic, distributed challenge detection approach. In this paper, we present a resilient distributed system to identify the challenges that have severe impact on the wired and wireless mesh network (WMN). Our design shows how a challenge (malicious attack) is handled initially by lightweight network monitoring, then progressively applying more heavyweight analysis in order to identify the challenge. Non-malicious challenges could also be simulated by our network failure module. Furthermore, WMNs are an interesting domain to consider network resilience. Automatic detection and mitigation is a desirable property of a resilient WMN. We present guidelines to address the challenge of channel interferences in the WMN. The feasibility of our framework is demonstrated through experiment. We conclude that our proof-of-concept case study has provided valuable insight into resilient networks, which will be useful for further research.