Improving Robustness of Network Fault Diagnosis to Uncertainty in Observations

Abstract

Abstract-Performing decentralized network fault diagnosis based on network traffic is challenging. Besides inherent stochastic behaviour of observations, measurements may be subject to errors degrading diagnosis timeliness and accuracy. In this paper we present a novel approach in which we aim to mitigate issues of measurement errors by quantifying uncertainty. The uncertainty information is applied in the diagnostic component to improve its robustness. Three diagnosis components have been proposed based on the Hidden Markov Model formalism: (H0) representing a classical approach, (H1) a static compensation of (H0) to uncertainties and (H2) dynamically adapting diagnosis to uncertainty information. From uncertainty injection scenarios of added measurement noise we demonstrate how using uncertainty information can provide a structured approach of improving diagnosis.

Towards a Framework for Self-Adaptive Reliable Network Services in Highly-Uncertain Environments

“…Next, we introduce an example scenario and finally present intermediate results based on a detailed system level simulation analysis. Further details can be found in [13].…”

“…(Hu) is the same as (H) but uses the information on measurement uncertainty to update B for each observation provided by the OPP. In practice, this is commenced by a dynamic discretization approach: details may be found in [13]. Applying a HMM for diagnosis, it can be used to derive the probability δ FAULT of being in the fault state given an observation sequence.…”