You are here

Event Details

MP Associates, Inc.
MONDAY October 14, 3:30pm - 5:00pm | KC 405
The Information Processing Factory – A Paradigm for Life Cycle Management of Dependable Systems
Sankar Basu - National Science Foundation (NSF)
Rolf Ernst - Technische Univ. Braunschweig
Nikil Dutt - Univ. of California, Irvine
Fadi Kurdahi - Univ. of California, Irvine
Andreas Herkersdorf - Technische Univ. München
The number and complexity of embedded system platforms used in mixed-criticality applications are rapidly growing. They run large and evolving applications on heterogeneous multi-/many-core processing platforms requiring long term dependable operation. Examples include automated driving, smart buildings, industry 4.0, or personal medical devices. A 2016 ESWEEK special session gave an overview of state of the art in self-aware HW/SW system technology which could become an important basis for mastering complex dependable systems. The session presentations explained types and implementations of self-awareness and compared different approaches, including on-the-fly computing, self-aware platform control, and cross-layer self-awareness. One of the talks introduced the paradigm of an "Information Processing Factory" (IPF). An abstract concept at that time, IPF has been further elaborated and became the foundation of a US-German research initiative for research into detailed solutions and applications. IPF applies principles inspired by factory management to the continuous operation and optimization of highly-integrated embedded systems. A general objective is identifying a sweet spot between a maximum of autonomy among IPF constituent components and a minimum of centralized control in order to ensure guaranteed service even under strict safety and availability requirements. Emphasis is on intensive self- diagnosis for early detection of degradation and imminent failures combined with unsupervised platform self-adaptation to meet performance and safety targets. The initiative developed into a research cluster that is jointly funded by the NSF and DFG. The cluster exploits a variety of technologies including proactive reconfiguration to mitigate the risk of failures, self-optimization, and self-identification using learning classifiers, and chip-level operation with flexible boundaries between critical and best effort regions, all guided by a self-aware planning component. A large many-core multi-OS simulation platform provides the means for in-depth cross-layer experiments.

3A.1Introduction to the IPF platform and research cluster
 Speaker: Rolf Ernst - TU Braunschweig
 Author: Rolf Ernst - TU Braunschweig
3A.2Proactive self-diagnosis and task migration for safety-critical applications
 Speaker: Rolf Ernst - TU Braunschweig
 Authors: Eberle A. Rambo - TU Braunschweig
Thawra Kadeed - TU Braunschweig
Rolf Ernst - TU Braunschweig
3A.3IPF runtime verification
 Speaker: Fadi Kurdahi - UC Irvine
 Authors: Minjun Seo - UC Irvine
Fadi Kurdahi - UC Irvine
3A.4Reflective supervisory control in hierarchical machine learning
 Speaker: Nikil Dutt - UC Irvine
 Authors: Bryan Donyanavard - UC Irvine
Caio Batista de Melo - UC Irvine
Biswadip Maity - UC Irvine
Kasra Moazzemi - UC Irvine
Kenneth Stewart - UC Irvine
Saehanseul Yi - UC Irvine
Amir M. Rahmani - UC Irvine
Nikil Dutt - UC Irvine
3A.5Hardware-based learning classifiers for mixed-critical environments
 Speaker: Andreas Herkersdorf - TU Munich
 Authors: Nguyen Anh Vu Doan - TU Munich
Florian Maurer - TU Munich
Anmol Surhonne - TU Munich
Thomas Wild - TU Munich
Andreas Herkersdorf - TU Munich