Control-Based Operating System Design

✍ Scribed by Alberto Leva, Martina Maggio, Alessandro Vittirio Papadopoulos, Federico Terraneo

• Publisher: The Institution of Engineering and Technology
• Year: 2013
• Tongue: English
• Leaves: 218
• Series: IET Control Engineering Series 89
• Category: Library

✦ Synopsis

Control-Based Operating System Design describes the application of system- and control-theoretical methods to the design of computer operating system components. It argues that computer operating system components should not be first “designed" and then “endowed with control", but rather conceived from the outset as controllers, synthesized and assessed in the system-theoretical world of dynamic models, and then realized as control algorithms. The book includes both a theoretical treatment of the usefulness of the approach, and the description of a complete implementation in the form of a microcontroller kernel, made available as free software. Topics covered include modelling and control design paradigms, task scheduling, resource allocation, application performance control, sensing and actuating, and the implementation and assessment of Miosix, a control-based kernel.

✦ Table of Contents

Control-Based Operating System Design......Page 4
Contents......Page 10
List of trademarks......Page 15
Preface......Page 16
Acknowledgements......Page 20
1 Introduction......Page 22
1.1 Βίοι παράλληλοι (two parallel stories)......Page 24
1.2 Control-based design as a means for convergence......Page 27
2.1 Dynamic systems......Page 32
2.1.1 State-space representation......Page 34
2.1.2 Motion, equilibrium, stability......Page 36
2.2.1 Motion and equilibrium......Page 37
2.2.2 Stability......Page 38
2.3.1 The Z transform......Page 40
2.3.2 The transfer function......Page 41
2.3.3 Block diagrams......Page 42
2.3.3.2 Parallel connection......Page 43
2.3.3.3 Feedback (loop) connection......Page 44
2.4.1 Definition......Page 45
2.4.2 Interpretation and use......Page 46
2.5 Time domain responses......Page 48
2.5.2 Impulse response of IIR systems......Page 49
2.7 Problems......Page 51
3 Modelling for computing systems......Page 54
3.1 The quest for a computer physics......Page 55
3.2 Modelling and simulation......Page 57
3.3.1 Core allocation......Page 58
3.3.2 Producer and consumer......Page 59
3.3.4 Communication bandwidth partitioning......Page 60
3.4 Concluding remarks......Page 62
4.1 Specifications......Page 64
4.2 Main control schemes......Page 65
4.3 Feedback and its power......Page 66
4.4.1 Synthesis by transfer function assignment (set point tracking)......Page 68
4.4.2 Synthesis by transfer function assignment (disturbance rejection)......Page 70
4.5 Some typical feedback controllers......Page 71
4.5.1 Proportional control......Page 72
4.5.3 Proportional-integral (PI) control......Page 73
4.6 Standard controllers on simple plants......Page 74
4.7 From controller model to control law......Page 75
4.7.1 Managing control saturations......Page 76
4.8 Problems......Page 77
5 Scheduling......Page 78
5.1 Modelling......Page 79
5.1.1 The core phenomenon......Page 82
5.2 Control synthesis......Page 83
5.2.1 Inner loop......Page 84
5.2.2 Outer loop......Page 85
5.2.3 Complexity comparison with existing policies......Page 86
5.3 Set point generation for (soft) real-time systems......Page 89
5.3.1 Overload detection and rescaling......Page 91
5.3.2 Reinitialisation and feedforward......Page 92
5.4.1 MiBench benchmark......Page 94
5.4.2 Hartstone benchmark......Page 95
5.4.4 Summary of results......Page 98
5.5 Set point generation for general purpose systems......Page 99
5.5.1 Tasks with periodic deadlines......Page 100
5.5.3 Tasks without deadlines......Page 101
5.5.5 Parameter setting......Page 102
5.5.6 Simulation examples......Page 103
5.5.7 Concluding remarks......Page 105
6 Memory management......Page 106
6.1 Problem statement......Page 109
6.2 The plant model......Page 110
6.2.1 Requirements......Page 113
6.3 Control synthesis......Page 114
6.5 Implementation-related considerations......Page 117
6.6 Concluding remarks......Page 119
7.1.1 Predictive control......Page 122
7.1.3 State-space models......Page 123
7.1.4 Predictive control within a single optimisation window......Page 124
7.1.4.1 Prediction of state and output variables......Page 125
7.1.4.2 Optimisation......Page 126
7.1.5 Receding-horizon predictive control......Page 127
7.1.5.1 Closed-loop control system......Page 128
7.2.1 Least squares......Page 129
7.2.2 Persistent excitation......Page 130
7.2.3 Recursive least squares......Page 132
7.3.1 Online identification and adaptive control......Page 135
7.3.2.1 Adaptive identification algorithms with forgetting factor......Page 136
7.4 Problems......Page 137
8 Resource allocation......Page 140
8.1 Literature review......Page 142
8.2.1 Sensing......Page 143
8.2.3.1 Heuristic......Page 144
8.2.3.2 Basic control......Page 145
8.2.3.3 Adaptive control......Page 148
8.2.4 Modelling for advanced control......Page 149
8.2.5 Regulating with tuning......Page 151
8.3.1 Swaptions......Page 152
8.3.2 Vips......Page 154
8.4 Concluding remarks......Page 156
9 Power-awareness......Page 158
9.1.1 Step 1: Analysis, sensors and actuators......Page 159
9.1.2 Step 2a: Data collection......Page 161
9.1.3 Step 2b: Control design......Page 163
9.2 Experimental results......Page 164
9.3.2 Step 2a......Page 169
9.3.4 Step 3......Page 170
9.4 Concluding remarks......Page 171
10.1 Motivations......Page 172
10.2 Requirements and design decisions......Page 173
10.3 Architecture of Miosix......Page 174
10.4 The Miosix scheduler......Page 175
10.4.1 Pluggable schedulers in Miosix......Page 177
10.4.2.1 Sensors and actuators......Page 178
10.4.2.2 Context switch implementation......Page 179
10.5 Future directions......Page 180
11.1 Control-related concepts coverage......Page 182
11.2.2 Bandwidth scheduling......Page 184
11.3.1 Multi-core and multi-CPU scheduling......Page 185
11.4 A cyber-physical perspective......Page 186
A.1.1 Bandwidth allocation simulation......Page 192
A.1.3 Full memory management simulator......Page 193
A.1.4 Scheduler simulator......Page 196
A.2 An implementation example......Page 201
References......Page 206
Index......Page 212