What was Stefan Petri doing previously?

From 1992 to 1997, I have worked at the Institute for Operating Systems and Networking (Institut für Betriebssysteme und Rechnerverbund) of the TU Braunschweig. (A description of my (former) work can currently still be found there.) In January 1997, I finished my dissertation there.

From April to December 1997, I was working at the Lehrstuhl für Rechnertechnik und Rechnerorganisation (LRR-TUM, Prof. Bode) of the Technische Universität München, funded by the Sonderforschungsbereich 342 ``Werkzeuge und Methoden zur Nutzung paralleler Rechnerarchitekturen'' (roughly: special research area 342 ``Tools and Methods for Exploiting Parallel Computer Architectures''), in the project part A1 ``Integration von Werkzeugumgebung und Rechnerarchitektur zur Parallelisierung'' (``Integration of Tool Environment and Computer Architecture for Parallelization'').

On 1. January 1998 I migrated to the Institut für Technische Informatik of the Medizinische Universität zu Lübeck (Institute of Computer Engineering of the Medical University at Lübeck). There I worked on clusters of workstations and PCs connected by high speed networks (SCI, HIC), on fault tolerance and load balancing.
I was involved in the ESPRIT project 20693 ``ARCHES''.
I co-organized the 3rd Workshop on System-Integrated Load and Resource Management, SVLR'99.
And also subsequent SVLR workshops (at some point renamed to SILR).

In Braunschweig, I worked on , which provides transparent checkpointing and migration for distributed applications in workstation clusters.
In Munich, I have ``inherited'' the CoCheck co-ordinated checkpointer for parallel and distributed applications of Georg Stellner.
Part of my work was integration of CoCheck into the Online Monitoring Interface Specification (OMIS) environment.
My general ``favourite'' research interest is Load Balancing and Fault Tolerance in Distributed Systems (a link back to the old project home page at TU Braunschweig). That comprises the following subjects:
- Tools for Monitoring, Visualisation and Performance Analysis of Parallel and Distributed Architectures and Applications.
- Combination of System and Application Integrated Load Balancing Techniques.
- The system for Migration of Communicating Processes (i.e. distributed applications) in Workstation Clusters.
- Scheduling Strategies for Clusters of Configurational Heterogenious Workstations, especially learning parameters and rules based on Neural Networks and/or Fuzzy Logic.
Here is a shortcut to my page with References to Online Documents about Load Balancing and Process Migration.
And a shortcut to the BibTeX file from my PhD thesis about load balancing and fault tolerance in workstation clusters.
In Braunschweig, I have also worked on CURE - Computer Assisted Rehabilitation.