Assessing Anonymized System Logs Usefulness for Behavioral Analysis in RNN Models

TL;DR

This paper evaluates the effectiveness of anonymized system logs, processed with PaRS, for behavioral analysis using RNN models, addressing privacy concerns while maintaining analytical usefulness.

Contribution

It introduces an assessment of anonymized logs' usefulness for RNN-based behavioral analysis, highlighting the impact of content-aware anonymization methods like PaRS.

Findings

Anonymized logs retain significant behavioral information for RNN analysis.

Content-aware anonymization preserves log correlations better than traditional methods.

The study demonstrates the feasibility of privacy-preserving behavioral analysis in HPC systems.

Abstract

System logs are a common source of monitoring data for analyzing computing systems' behavior. Due to the complexity of modern computing systems and the large size of collected monitoring data, automated analysis mechanisms are required. Numerous machine learning and deep learning methods are proposed to address this challenge. However, due to the existence of sensitive data in system logs their analysis and storage raise serious privacy concerns. Anonymization methods could be used to clean the monitoring data before analysis. However, anonymized system logs, in general, do not provide adequate usefulness for the majority of behavioral analysis. Content-aware anonymization mechanisms such as PaRS preserve the correlation of system logs even after anonymization. This work evaluates the usefulness of anonymized system logs taken from the Taurus HPC cluster anonymized using PaRS, for behavioral analysis via recurrent neural network models.