Project: Using In-Memory Computing for Proactive Cloud Operations


  • Dr. Felix Salfner (SAP Innovation Center Potsdam)
  • Dr. Peter Tröger (Hasso-Plattner-Institute, Potsdam University)
  • Adam Oliner (AMP Lab, University of California, Berkeley)

Abstract: Our proposed FutureSOC lab project focuses on one specific aspect of cloud operation: Anomaly detection. When something goes wrong in the system, it is essential to trigger preventive maintenance activities quickly, such as alarming of administrators or automated load migration. Additionally, the administrators must be enabled to quickly identify the root cause of a problem, since this reduces time-to- repair and hence improves availability as well. In order to advance the state-of-the-art in this area, we want to apply a novel anomaly detection approach developed at Stanford University to the specific case of SAP’s Business ByDesign? (ByD) cloud offering. We want to investigate how well anomaly detection performs in this specific case, as well as how it might have to be adapted or refined. Additionally, we want to explore intuitive ways to represent the results in the user interface of a SAP ByDesign system management console. For these experiments, we plan to utilize the SAP Business ByDesign installation in the FutureSOC lab.

Last modified 9 years ago Last modified on May 3, 2012 12:21:23 PM