Turning Privacy Constraints into Syslog Analysis Advantage

Nowadays, failures in high performance computers (HPC) became the norm rather than the exception [10]. In the near future, the mean time between failures (MTBF) of HPC systems is expected to be too short, and that current failure recovery mechanisms e.g., checkpoint-restart, will no longer be able to recover the systems from failures [1]. Early fail- ure detection is a new class of failure recovery methods that can be beneficial for HPC systems with short MTBF. De- tecting failures in their early stage can reduce their negative effects by preventing their propagation to other parts of the system [3]. The goal of the current work is to contribute to the foun- dation of failure detection techniques via sharing an ongo- ing research with the community. Herein we consider user privacy as the main priority, and then turning the applied constraint for protecting users privacy into an advantage for analyzing the system behavior. We use De-identification, constantification, and hashing to reach this goal. Our ap- proach also contributes to the reproducibility and openness of future research in the field. Via this approach, system ad- ministrators can unwarily share their syslogs with the public domain.