Analytical Network and System Administration

Mark Burgess is an Associate Professor at University College in Oslo, Norway,

• Foreword xiPreface xiv1 Introduction 11.1 What is system administration? 11.2 What is a system? 21.3 What is administration? 21.4 Studying systems 31.5 What’s in a theory? 61.6 How to use the text 101.7 Some notation used 102 Science and its methods 132.1 The aim of science 132.2 Causality, superposition and dependency 162.3 Controversies and philosophies of science 172.4 Technology 202.5 Hypotheses 202.6 The science of technology 212.7 Evaluating a system—dependencies 222.8 Abuses of science 223 Experiment and observation 253.1 Data plots and time series 263.2 Constancy of environment during measurement 273.3 Experimental design 283.4 Stochastic (random) variables 293.5 Actual values or characteristic values 303.6 Observational errors 303.7 The mean and standard deviation 313.8 Probability distributions and measurement 323.8.1 Scatter and jitter 353.8.2 The ‘normal’ distribution 353.8.3 Standard error of the mean 363.8.4 Other distributions 373.9 Uncertainty in general formulae 383.10 Fourier analysis and periodic behaviour 393.11 Local averaging procedures 413.12 Reminder 434 Simple systems 454.1 The concept of a system 454.2 Data structures and processes 464.3 Representation of variables 474.4 The simplest dynamical systems 484.5 More complex systems 494.6 Freedoms and constraints 504.7 Symmetries 514.8 Algorithms, protocols and standard ‘methods’ 524.9 Currencies and value systems 534.9.1 Energy and power 534.9.2 Money 544.9.3 Social currency and the notion of responsibility 544.10 Open and closed systems: the environment 564.11 Reliable and unreliable systems 585 Sets, states and logic 595.1 Sets 595.2 A system as a set of sets 615.3 Addresses and mappings 615.4 Chains and states 625.5 Configurations and macrostates 645.6 Continuum approximation 655.7 Theory of computation and machine language 655.7.1 Automata or State Machines 665.7.2 Operators and operands 685.7.3 Pattern matching and operational grammars 695.7.4 Pathway analysis and distributed algorithms 705.8 A policy-defined state 716 Diagrammatical representations 736.1 Diagrams as systems 736.2 The concept of a graph 746.3 Connectivity 776.4 Centrality: maxima and minima in graphs 776.5 Ranking in directed graphs 806.6 Applied diagrammatical methods 847 System variables 917.1 Information systems 917.2 Addresses, labels, keys and other resource locators 927.3 Continuous relationships 947.4 Digital comparison 948 Change in systems 978.1 Renditions of change 978.2 Determinism and predictability 988.3 Oscillations and fluctuations 998.4 Rate of change 1028.5 Applications of the continuum approximation 1038.6 Uncertainty in the continuum approximation 1059 Information 1099.1 What is information? 1099.2 Transmission 1109.3 Informationandcontrol 1119.4 Classification and resolution 1119.5 Statistical uncertainty and entropy 1149.6 Propertiesoftheentropy 1189.7 Uncertainty in communication 1199.8 A geometrical interpretation of information 1239.9 Compressibility and size of information 1279.10 Information and state 1289.11 Maximum entropy principle 1299.12 Fluctuation spectra. 13310 Stability 13510.1 Basic notions 13510.2 Types of stability 13510.3 Constancy 13610.4 Convergence of behaviour 13710.5 Maxima and minima 13810.6 Regions of stability in a graph 13910.7 Graph stability under random node removal 14110.8 Dynamical equilibria: compromise 14210.9 Statistical stability 14310.10 Scaling stability 14510.11 Maximum entropy distributions 14810.12 Eigenstates 14810.13 Fixed points of maps 15110.14 Metastable alternatives and adaptability 15510.15 Final remarks 15611 Resource networks 15911.1 What is a system resource? 15911.2 Representation of resources 16011.3 Resource currency relationships 16111.4 Resource allocation, consumption and conservation 16211.5 Where to attach resources? 16311.6 Access to resources 16511.7 Methods of resource allocation 16711.7.1 Logical regions of systems 16711.7.2 Using centrality to identify resource bottlenecks 16811.8 Directed resources: flow asymmetries 17012 Task management and services 17312.1 Task list scheduling 17312.2 Deterministic and non-deterministic schedules 17412.3 Human–computer scheduling 17612.4 Service provision and policy 17612.5 Queue processing 17712.6 Models 17812.7 The prototype queue M/M/ 1 17912.8 Queue relationships or basic ‘laws’ 18112.9 Expediting tasks with multiple servers M/M/k 18612.10 Maximum entropy input events in periodic systems 18812.11 Miscellaneous issues in scheduling 18913 System architectures 19113.1 Policy for organization 19113.2 Informative and procedural flows 19213.3 Structured systems and ad hoc systems 19313.4 Dependence policy 19313.5 System design strategy 19513.6 Event-driven systems and functional systems 20013.7 The organization of human resources 20113.8 Principle of minimal dependency 20213.9 Decision-making within a system 20213.9.1 Layered systems: Managers and workers 20213.9.2 Efficiency 20313.10 Prediction, verification and their limitations 20413.11 Graphical methods 20514 System normalization 20714.1 Dependency 20714.2 The database model 20914.3 Normalized forms 21015 System integrity 21515.1 System administration as communication? 21515.2 Extensive or strategic instruction 21915.3 Stochastic semi-groups and martingales 22315.4 Characterizing probable or average error 22415.5 Correcting errors of propagation 22615.6 Gaussian continuum approximation formula 22816 Policy and maintenance 23116.1 What is maintenance? 23116.2 Average changes in configuration 23116.3 The reason for random fluctuations 23416.4 Huge fluctuations 23516.5 Equivalent configurations and policy 23616.6 Policy 23716.7 Convergent maintenance 23716.8 The maintenance theorem 24016.9 Theory of back-up and error correction 24117 Knowledge, learning and training 24917.1 Information and knowledge 25017.2 Knowledgeasclassification 25017.3 Bayes’ theorem 25217.4 Belief versus truth 25417.5 Decisions based on expert knowledge 25517.6 Knowledge out of date 25917.7 Convergence of the learning process 26018 Policy transgressions and fault modelling 26318.1 Faults and failures 26318.2 Deterministic system approximation 26518.3 Stochasticsystemmodels 26918.4 Approximate information flow reliability 27318.5 Fault correction by monitoring and instruction 27518.6 Policy maintenance architectures 27918.7 Diagnostic cause trees 28618.8 Probabilistic fault trees 29018.8.1 Faults 29018.8.2 Conditions and set logic 29118.8.3 Construction 29319 Decision and strategy 29519.1 Causal analysis 29519.2 Decision-making 29619.3 Game theory 29719.4 The strategic form of a game 30119.5 The extensive form of a game 30219.6 Solving zero-sum games 30319.7 Dominated strategies 30419.8 Nash equilibria 30519.9 A security game 30919.9.1 Zero-sum approximation 31019.9.2 Non-zero sum approximation 31319.10 The garbage collection game 31519.11 A social engineering game 32119.12 Human elements of policy decision 32819.13 Coda: extensive versus strategic configuration management 32820 Conclusions 331A Some Boolean formulae 335A.1 Conditional probability 335A.2 Boolean algebra and logic 336B Statistical and scaling properties of time-series data 339B. 1 Local averaging procedure 339B. 2 Scaling and self-similarity 343B. 3 Scaling of continuous functions 344C Percolation conditions 347C. 1 Random graph condition 347C. 2 Bi-partite form 350C. 3 Small-graph corrections 351Bibliography 353Index 359