ABSTRACT
CONTENTS 8.1 Introduction................................................................................................149 8.2 Online Control Concepts ..........................................................................152 8.3 Case Study: Processor Power Management ..........................................156 8.4 Hierarchical Control ..................................................................................159 8.5 Case Study: Signal Detection System .....................................................161 8.6 Conclusions ................................................................................................165 References.............................................................................................................165
This chapter describes a model-based control and optimization framework to design autonomic or self-managing computing systems that continually optimize their performance in response to changing workload demands and operating conditions. The performance management problems of interest are posed as one of sequential optimization under uncertainty, and a lookahead control approach is used to optimize the forecast system behavior over a limited prediction horizon. The basic control concepts are then extended to tackle distributed computing systems where multiple controllers must interact with each other to ensure overall performance goals. Two case studies, dealing with processor power management and distributed signal classification, demonstrate the applicability of the control framework.
