Network and system administration usually refers to the skill of keeping computers and networks running properly. But in truth, the skill needed is that of managing complexity. This book describes the science behind these complex systems, independent of the actual operating systems they work on. It provides a theoretical approach to systems administration that:
- saves time in performing common system administration tasks.
- allows safe utilization of untrained and trained help in maintaining mission-critical systems.
- allows efficient and safe centralized network administration.
Managing Human-Computer Networks:
- Will show how to make informed analyses and decisions about systems, how to diagnose faults and weaknesses
- Gives advice/guidance as to how to determine optimal policies for system management
- Includes exercises that illustrate the key points of the book
The book provides a unique approach to an old problem and will become a classic for researchers and graduate students in Networking and Computer Science, as well as practicing system managers and system administrators.
Table of Contents:
Foreword xi
Preface xiv
1 Introduction 1
1.1 What is system administration? 1
1.2 What is a system? 2
1.3 What is administration? 2
1.4 Studying systems 3
1.5 What’s in a theory? 6
1.6 How to use the text 10
1.7 Some notation used 10
2 Science and its methods 13
2.1 The aim of science 13
2.2 Causality, superposition and dependency 16
2.3 Controversies and philosophies of science 17
2.4 Technology 20
2.5 Hypotheses 20
2.6 The science of technology 21
2.7 Evaluating a system—dependencies 22
2.8 Abuses of science 22
3 Experiment and observation 25
3.1 Data plots and time series 26
3.2 Constancy of environment during measurement 27
3.3 Experimental design 28
3.4 Stochastic (random) variables 29
3.5 Actual values or characteristic values 30
3.6 Observational errors 30
3.7 The mean and standard deviation 31
3.8 Probability distributions and measurement 32
3.8.1 Scatter and jitter 35
3.8.2 The ‘normal’ distribution 35
3.8.3 Standard error of the mean 36
3.8.4 Other distributions 37
3.9 Uncertainty in general formulae 38
3.10 Fourier analysis and periodic behaviour 39
3.11 Local averaging procedures 41
3.12 Reminder 43
4 Simple systems 45
4.1 The concept of a system 45
4.2 Data structures and processes 46
4.3 Representation of variables 47
4.4 The simplest dynamical systems 48
4.5 More complex systems 49
4.6 Freedoms and constraints 50
4.7 Symmetries 51
4.8 Algorithms, protocols and standard ‘methods’ 52
4.9 Currencies and value systems 53
4.9.1 Energy and power 53
4.9.2 Money 54
4.9.3 Social currency and the notion of responsibility 54
4.10 Open and closed systems: the environment 56
4.11 Reliable and unreliable systems 58
5 Sets, states and logic 59
5.1 Sets 59
5.2 A system as a set of sets 61
5.3 Addresses and mappings 61
5.4 Chains and states 62
5.5 Configurations and macrostates 64
5.6 Continuum approximation 65
5.7 Theory of computation and machine language 65
5.7.1 Automata or State Machines 66
5.7.2 Operators and operands 68
5.7.3 Pattern matching and operational grammars 69
5.7.4 Pathway analysis and distributed algorithms 70
5.8 A policy-defined state 71
6 Diagrammatical representations 73
6.1 Diagrams as systems 73
6.2 The concept of a graph 74
6.3 Connectivity 77
6.4 Centrality: maxima and minima in graphs 77
6.5 Ranking in directed graphs 80
6.6 Applied diagrammatical methods 84
7 System variables 91
7.1 Information systems 91
7.2 Addresses, labels, keys and other resource locators 92
7.3 Continuous relationships 94
7.4 Digital comparison 94
8 Change in systems 97
8.1 Renditions of change 97
8.2 Determinism and predictability 98
8.3 Oscillations and fluctuations 99
8.4 Rate of change 102
8.5 Applications of the continuum approximation 103
8.6 Uncertainty in the continuum approximation 105
9 Information 109
9.1 What is information? 109
9.2 Transmission 110
9.3 Informationandcontrol 111
9.4 Classification and resolution 111
9.5 Statistical uncertainty and entropy 114
9.6 Propertiesoftheentropy 118
9.7 Uncertainty in communication 119
9.8 A geometrical interpretation of information 123
9.9 Compressibility and size of information 127
9.10 Information and state 128
9.11 Maximum entropy principle 129
9.12 Fluctuation spectra. 133
10 Stability 135
10.1 Basic notions 135
10.2 Types of stability 135
10.3 Constancy 136
10.4 Convergence of behaviour 137
10.5 Maxima and minima 138
10.6 Regions of stability in a graph 139
10.7 Graph stability under random node removal 141
10.8 Dynamical equilibria: compromise 142
10.9 Statistical stability 143
10.10 Scaling stability 145
10.11 Maximum entropy distributions 148
10.12 Eigenstates 148
10.13 Fixed points of maps 151
10.14 Metastable alternatives and adaptability 155
10.15 Final remarks 156
11 Resource networks 159
11.1 What is a system resource? 159
11.2 Representation of resources 160
11.3 Resource currency relationships 161
11.4 Resource allocation, consumption and conservation 162
11.5 Where to attach resources? 163
11.6 Access to resources 165
11.7 Methods of resource allocation 167
11.7.1 Logical regions of systems 167
11.7.2 Using centrality to identify resource bottlenecks 168
11.8 Directed resources: flow asymmetries 170
12 Task management and services 173
12.1 Task list scheduling 173
12.2 Deterministic and non-deterministic schedules 174
12.3 Human–computer scheduling 176
12.4 Service provision and policy 176
12.5 Queue processing 177
12.6 Models 178
12.7 The prototype queue M/M/ 1 179
12.8 Queue relationships or basic ‘laws’ 181
12.9 Expediting tasks with multiple servers M/M/k 186
12.10 Maximum entropy input events in periodic systems 188
12.11 Miscellaneous issues in scheduling 189
13 System architectures 191
13.1 Policy for organization 191
13.2 Informative and procedural flows 192
13.3 Structured systems and ad hoc systems 193
13.4 Dependence policy 193
13.5 System design strategy 195
13.6 Event-driven systems and functional systems 200
13.7 The organization of human resources 201
13.8 Principle of minimal dependency 202
13.9 Decision-making within a system 202
13.9.1 Layered systems: Managers and workers 202
13.9.2 Efficiency 203
13.10 Prediction, verification and their limitations 204
13.11 Graphical methods 205
14 System normalization 207
14.1 Dependency 207
14.2 The database model 209
14.3 Normalized forms 210
15 System integrity 215
15.1 System administration as communication? 215
15.2 Extensive or strategic instruction 219
15.3 Stochastic semi-groups and martingales 223
15.4 Characterizing probable or average error 224
15.5 Correcting errors of propagation 226
15.6 Gaussian continuum approximation formula 228
16 Policy and maintenance 231
16.1 What is maintenance? 231
16.2 Average changes in configuration 231
16.3 The reason for random fluctuations 234
16.4 Huge fluctuations 235
16.5 Equivalent configurations and policy 236
16.6 Policy 237
16.7 Convergent maintenance 237
16.8 The maintenance theorem 240
16.9 Theory of back-up and error correction 241
17 Knowledge, learning and training 249
17.1 Information and knowledge 250
17.2 Knowledgeasclassification 250
17.3 Bayes’ theorem 252
17.4 Belief versus truth 254
17.5 Decisions based on expert knowledge 255
17.6 Knowledge out of date 259
17.7 Convergence of the learning process 260
18 Policy transgressions and fault modelling 263
18.1 Faults and failures 263
18.2 Deterministic system approximation 265
18.3 Stochasticsystemmodels 269
18.4 Approximate information flow reliability 273
18.5 Fault correction by monitoring and instruction 275
18.6 Policy maintenance architectures 279
18.7 Diagnostic cause trees 286
18.8 Probabilistic fault trees 290
18.8.1 Faults 290
18.8.2 Conditions and set logic 291
18.8.3 Construction 293
19 Decision and strategy 295
19.1 Causal analysis 295
19.2 Decision-making 296
19.3 Game theory 297
19.4 The strategic form of a game 301
19.5 The extensive form of a game 302
19.6 Solving zero-sum games 303
19.7 Dominated strategies 304
19.8 Nash equilibria 305
19.9 A security game 309
19.9.1 Zero-sum approximation 310
19.9.2 Non-zero sum approximation 313
19.10 The garbage collection game 315
19.11 A social engineering game 321
19.12 Human elements of policy decision 328
19.13 Coda: extensive versus strategic configuration management 328
20 Conclusions 331
A Some Boolean formulae 335
A.1 Conditional probability 335
A.2 Boolean algebra and logic 336
B Statistical and scaling properties of time-series data 339
B. 1 Local averaging procedure 339
B. 2 Scaling and self-similarity 343
B. 3 Scaling of continuous functions 344
C Percolation conditions 347
C. 1 Random graph condition 347
C. 2 Bi-partite form 350
C. 3 Small-graph corrections 351
Bibliography 353
Index 359
