Radu Marculescu, Carnegie Mellon University (radum[at]ece.cmu.edu)
This half-day tutorial will be held on the morning of Monday, April 11. The complete schedule can be found here.
Abstract: We live in a world were computation, communication, and control are continuously and increasingly interwoven to produce functionally rich and energy efficient cyber-physical systems (CPS). As such, the focus of CPS design is not only to establish a solid and reliable communication infrastructure between such computational elements, but also to include time and feedback-based control as intrinsic components of the programming model; this allows generalizing the computational paradigm such that more direct interaction between the system and physical world becomes possible. It is obvious that such complex requirements call not only for a new science of networks, but also for a multi-disciplinary approach towards CPS design. Indeed, such a new science cannot rely on classical approaches for workload modeling and linear control paradigms.
In this tutorial we argue that precise and complex workload characterization should be one of the main drivers in CPS design and optimization. Consequently, we plan to introduce a new framework for workload characterization based on statistical physics and then show how this new paradigm can change the design of future CPS for the better. Towards this end, we plan to discuss various approaches inspired from statistical physics (e.g., master equation, renormalization group theory, fractional derivatives) that can greatly improve CPS modeling and optimization. Indeed, adopting such a theoretical approach has many implications in the design of new control strategies, topology selection and resource allocation, scheduling, routing protocols, and, last but not least, security.
Keywords: Cyber-physical systems, embedded systems, multiscale and multifractal behavior, multicore platforms, networks-on-chip, low-power, variability, control.
1. Characteristics of physical systems
2. Multiscale and multifractal modeling
3. Statistical physics approaches to CPS workload modeling
4. Non-stationary traffic analysis and its implications on multicore systems
5. Centralized vs. distributed control of multicore systems
6. Fundamental implications in future CPS design
Intended Audience: This tutorial is intended for an audience relatively new to CPS design with minimal background in algorithms, optimization, networking, and statistical physics. The presentation will introduce all the relevant background material, give an overview of the current state-of-the-art results in CPS design, and finally talk about dynamic and statistical resource optimization in the presence of parameter and workload variations. The material discussed in this tutorial is highly relevant to system designers and software developers interested in the future of cyber-physical systems.
Radu Marculescu is a Professor in the Dept. of Electrical and Computer Engineering at Carnegie Mellon University, USA. He received his Ph.D. in Electrical Engineering from the University of Southern California in 1998. He has received the Best Paper Award of IEEE Transactions on VLSI Systems in 2005, as well as several best paper awards in major conferences in the area of design automation. Dr. Marculescu is currently an Associate Editor of IEEE Trans. on Computers, IEEE Trans. on Computer-Aided Design of Circuits and Integrated Systems, and ACM Trans. on Embedded Computing Systems. He has been involved in organizing several international symposia, conferences, workshops, and tutorials, as well as guest editor of special issues in archival journals and magazines. His research focuses on design methodologies and software tools for embedded systems design, on-chip communication, and cyber-physical systems. Radu Marculescu is an ACM Distinguished Speaker.