CDC2000 Challenges in the Application of Control to Computer Systems

Challenges in the Application of Control to Computer Systems
Home Full List of Titles 1: Proceedings of CDC2000 Discrete Event Systems Control in Communication Systems Optimal Control and Applications I Optimisation Approaches and Methods Model Predictive Control Advances in Linear Estimation Stochastic and Uncertain Systems Nonlinear Control and Applications Nonlinear Estimation and Filtering Formation Control and its Applications New Approaches to Fuzzy Control Manufacturing Systems Automotive Applications Stability Issues in Hybrid Control Recent Advances in Stochastic Networks Optimal Control and Applications II Robust Controller Design - mu, L1 and H2 Constrained and Receding Horizon Control Identification and Control around the World Markov Decision Processes Nonlinear Optimisation Observers for Nonlinear Systems Motion Planning Neural / Fuzzy Stability and Control Motor Control Control of Quantum Phenomena I Hybrid Systems Methods Control in Communication Networks Robustness and Optimisation Bumpless Transfer, Antiwindup and Saturation Adaptive Control: Linear Systems Estimation and Closed Loop Identification Control of Markov Processes Nonlinear Filtering and Control Modelling, Identification and Validation of Nonlinear Systems Differential Geometric Control Theory for Mechanical Systems Nonlinear Output Feedback Control Pneumatics and Compression Systems Control of Quantum Phenomena II Stability of Hybrid Systems Performance Analysis in Communication Networks Adaptive Control of Nonlinear Systems LMI Methods in Design Robust Control of Time Delay Systems Subspace Identification Methods Nonlinear Stochastic Filtering and Estimation Bifurcations, Chaos and Control I New Progress in Synthesis of Nonlinear Systems I Implementation Issues of Sliding Mode Control Theory Control of Mixing in Shear Flows Novel Neural Network Control Techniques for Industrial Motion Control Systems Physiological Control Systems Optimal Control of Hybrid Systems Stochastic Models for Communication Networks Control and Stabilisation of Nonlinear Systems New Directions in Robust Control Linear Systems Theory Advanced Topics in Systems Theory Estimation in Action Bifurcations, Chaos and Control II New Progress in Synthesis of Nonlinear Systems II Numerical Design and Analysis Techniques for Nonlinear Systems Analysis and Control of Underactuated Systems Sliding Mode Control I Challenges in the Application of Control to Computer Systems Estimation and Diagnosis of Discrete Event Systems Communications and Games Optimal Control Stochastic Systems Model Reduction Methodologies Identification and Subspace Methods Applications of Nonlinear Adaptive Control Advances in Nonlinear Output Feedback Design The Behavioural Approach to Systems and Control Vision Based Estimation and Control: Recent Advances and Open Problems Agile Control of Military Operations Sliding Mode Control II Model-based Fault Diagnosis of Industrial Processes Discrete Event Systems / Petri Nets System Identification and Confidence Estimation New Approaches to H-Infinity Control I Probabilistic Approaches to Robust Control Time Delay System Stabilisation Identification Methods Controlled Stochastic Processes Output Feedback of Nonlinear Systems Topics in Nonlinear Stabilisation Mobile Robots: Tracking Control Robust Control of Nonlinear Systems Power Systems Stabilisation and Control Disk Drive Control Hybrid Control Applications Discrete Time Systems New Approaches to H-Infinity Control II Linear Systems with Saturating Actuators New Theories in Distributed Parameter Systems Applications of Estimation and Identification Stochastic Control and Tuning Methodologies Control of Nonlinear Systems Iterative Learning and Control Coordinating Robot Systems Nonlinear Time Varying Systems Novel Applications of Neural Networks Aerospace Applications Switched Systems Implicit and Descriptor Systems LQG Periodic Systems and Disturbances New Horizons for Distributed Parameter Systems State Estimation Learning and Neuro-Control Nonlinear Control and Stabilisation I Tracking Vision Servoing Controllability of Nonlinear Systems Control of Flexible Systems Electro-Mechanical Systems Robust Control Methods and Applications Fault Detection and Diagnosis Optimisation and Applications Robust Stability Analysis Numerical Methods in Control Filtering in Continuous Time Stochastic Systems Interplay between Control and Signal Processing Fault Detection and Analysis Nonlinear Dynamical Systems Nonlinear Time Delay Systems Computational Issues in Nonlinear Control Disturbance Rejection Process Control Industry Applications Linear Parameter Varying Systems Linear Control Systems Dynamic and Nonlinear Programming Model Reduction Applications New Techniques for Control and Systems: Numerical Linear Algebra Estimation and Identification using Hidden Markov Models Applications of Stochastic Control Topics in Linear Design Nonlinear Control and Stabilisation II Ambulatory Robot Systems Chaotic and Oscillatory Systems Biomedical System Control Integrated Control and CPU Scheduling Linear Design Techniques Adaptive Disturbance / Noise Compensation Nonlinear Model Predictive Control Sensitivity Design, Analysis and Limitations Analysis of Linear Systems Linear Matrix Inequalities in Design Lyapunov's 2nd Method Robotics: Tracking Control Lagrangian and Hamiltonian Theory Variable Structure Control Machine Vision Signal Processing Methods in Control Applied Nonlinear Control Author Index A B C D E F G H I J K L M N O P Q R S T U V W X Y Z	Control and Modeling Issues in Computer Operating Systems: Resource Management for Real-Rate Computer Applications Authors: David C. Steere, Molly H. Shor, Ashvin Goel, Jonathan Walpole, Calton Pu, Volume: 1, Page 2212 Paper number 5701 Abstract: Commonplace computer applications on general-purpose computers increasingly are expected to meet "real-rate" requirements, processing or displaying data or images at an externally driven "rate". We describe a feedback-control-based resource manager design approach, allowing the computer system to allocate resources such as CPU and network bandwidth based on the measured "progress" of the applications. Progress is measured by separating a complex application into a number of simpler applications separated by buffers. The resource scheduler measures the buffer fill levels to determine whether the rates of data coming in and going out of each buffer are matched. Feedback controllers keep the buffer levels around a certain fill level. We have developed prototype systems in the Linux environment that demonstrate that (classical) feedback control can be used to match the real rates. However, more formal methods, such as those that can be developed by the control theory community, are needed to help with the analysis and design of such systems to make them commercially viable. This paper presents the computer system problems, results from the prototype designs showing feasibility, some preliminary modeling results, and demonstrations and discussions of which control modeling, analysis and design results and techniques appear to be relevant to this computer system problem, and why. CD005701.PDF (From Author) TOP Feedback Control Resource Management Using A Posteriori Workload Characterizations Authors: David R. Alexander, Douglas A. Lawrence, Lonnie R. Welch, Volume: 1, Page 2222 Paper number 5702 Abstract: Certain real-time applications must operate in highly dynamic environments, thereby precluding accurate characterization of the applications' workloads by static models. Thus, guarantees of real-time performance based on a priori characterizations are not possible. However, potential benefits of a posteriori approaches are significant, including the ability to function correctly in dynamic environments (through adaptability to unforeseen conditions), and higher actual utilization of computing resources. In this paper, we consider a control theoretic framework that is appropriate for systems which experience large variations in workload. The goal is to manage a distributed collection of computing resources by continuously computing and assessing QoS and resource utilization metrics that are determined a posteriori. Potential benefits of developing a control theoretic framework for resource management include: reuse of the large body of results in control theory, focus on problem characteristics instead of control infrastructures, and stability analysis. CD005702.PDF (From Author) TOP Control Challenges in Multi-Level Adaptive Video Streaming Authors: Dylan McNamee, Charles Krasic, Kang Li, Ashvin Goel, Erik Walthinsen, David C. Steere, Jonathan Walpole, Volume: 1, Page 2228 Paper number 5704 Abstract: Streaming video is one of the fastest-growing applications of the Internet. The Internet's diversity and dynamism demands that video streams adapt to ensure maximum quality at all times. This paper describes the control challenges we have encountered in the Quasar project's "multi-level" adaptive streaming video player. We first describe the framework and environment of the player. This frame-work uses software feedback to control resource allocation as well as the quality of media delivery. We present the control challenges raised by our framework, which include horizontal and vertical feedback composition, difficult to model systems, and unpredictable, non-linear actuators. We describe some of the approaches we are taking to address these challenges, related work, and future application areas and the challenges they will raise. CD005704.PDF (From Author) TOP Modeling and Performance Control of Internet Servers Authors: Tarek F. Abdelzaher, Chenyang Lu, Volume: 1, Page 2234 Paper number 5705 Abstract: The paper describes modeling and performance control of an Internet server using classical feedback control theory. We show that classical feedback control can leverage on well-known real-time scheduling results to resolve one of the fundamental problems in Internet-servers today; namely, achieving overload protection and performance guarantees in the presence of load unpredictability. The research is motivated by the increasing proliferation of a new category of Web-based services, such as online trading, banking, and business transactions, where performance guarantees are required in the face of unpredictable server load. Failure to meet desired performance levels may result in loss of customers, financial damage or liability violations. State-of-the-art Web servers are not designed to offer such performance guarantees. We show that control theory offers a robust solution to the server performance control problem. We demonstrate that a general web server may be modeled as a linear time-varying system, describe the equivalents of sensors and actuators in that system, formulate a simple feedback loop, describe how it can leverage on real-time scheduling theory to achieve timing guarantees, and evaluate the efficacy of the scheme on an experimental testbed using a real web server (Apache), which is the most popular Internet server today. Experimental results indicate that control-theoretical techniques offer a promising way of achieving desired performance in emerging critical Internet applications. CD005705.PDF (From Author) TOP