Analyzing System Performance Dynamics
Overview
This project aims to detect problems and anomalies in dynamics of large system performance data. It uses time-series modeling and feature selection methods to model normal activity and check for unexpected adverse conditions. It focuses on the zVM operating system and the performance measurements that it collects from its hypervisor and virtual machines. This is a joint program with zSeries development in STG.
Contact: Lavi Shpigelman (lavi@il.ibm.com)