Demo Session

Service Health Analyzer
Ira Cohen, HP Labs
HP Service Health Analyzer (SHA) is the industry's first predictive analytics tool built on top of a real-time, dynamic service model providing customers with a configuration free system that proactively detects and decodes IT performances problems. We will provide a demo of SHA and describe the research behind it involving machine learning applied to IT data.

Agurim: Multi-dimensional Flow Re-aggregation for Traffic Monitoring
Midori Kato, Keio University
A promising way to capture the characteristics of changing traffic is to extract significant flow clusters in traffic. However, clustering flows by 5-tuple requires flow matching in huge flow attribute spaces, and thus, is difficult to perform on the fly. We propose an efficient yet flexible flow aggregation technique for monitoring the dynamics of network traffic. In the demonstration, we present Agurim, our resulting software.

FDiag A Failure Diagnostics Toolkit based on the Analysis of Cluster System Logs
Edward Chuah, University of Texas at Austin
A goal for the analysis of cluster system logs is to determine the sources and causes of system failures.  Large cluster systems are composed of many hardware and software components, and they are used to execute jobs that require the immense computational power provided by these systems.  When nodes of a large cluster system crash or when jobs hang, the root-causes of these failures must be identified.  However, cluster system logs are huge, incomplete and contain considerable ambiguity so that direct discovery of the complete causal trace path of events leading to the failure is difficult.  In this presentation, we will demonstrate how FDiag can be used to process the logs of the Ranger supercomputer and to generate the diagnostics reports from which the systems administrators can use to determine where (the nodes and jobs), when (the times) and why (the causes) the system failure occurred.