Systems Modeling: Methodologies and Tools (eBook)

Antonio Puliafito is a full professor of computer engineering at the University of Messina, Italy. His interests include parallel and distributed systems, networking, IoT, Cloud computing, advanced analytical modeling techniques. He regularly acts as a referee for the European Community since 1998. He contributed to the development of the software tools WebSPN, ArgoPerformance and Stack4Things. He co-authored the text entitled "Performance and Reliability Analysis of Computer Systems", edited by Kluwer. He leads the Center for Information Technologies at University of Messina (CIAM). From 2006 to 2008 he acted as the technical director of the Project 901, winner of the CISCO innovation award. He actively contributed to the success of the TriGrid VL and PI2S2 projects. He has been working in several EU funded projects such as: Reservoir, Vision Cloud, CloudWave, Beacon, Frontier Cities. He was also the main investigator of the Italian PRIN2008 research project Cloud@Home,  to combine cloud and volunteer computing. He acted as scientific coordinator of the PON 2007-2013 SIGMA project on using cloud computing to manage severe risk phenomena. He is coordinating the #SmartME crowdfunding initiative to develop a smart city framework in the city of Messina. He is the co-founder of SmatMe.io, a startup working on the integration of cloud and IoT in smart cities contexts.    Kishor Trivedi holds the Fitzgerald Hudson Chair in the Department of Electrical and Computer Engineering at Duke University, Durham, NC. He has a B.Tech. (EE) from IIT Mumbai and MS and PhD (CS) from the University of Illinois at Urbana-Champaign. He has been on the Duke faculty since 1975. He is the author of a well-known text entitled, Probability and Statistics with Reliability, Queuing and Computer Science Applications, originally published by Prentice-Hall; a thoroughly revised second edition (including its Indian edition) of this book has been published by John Wiley. The book is recently translated into Chinese. His most recent book, Reliability and Availability Engineering, is published by Cambridge University Press in 2017. He has also published two other books entitled, Performance and Reliability Analysis of Computer Systems, published by Kluwer Academic Publishers and Queueing Networks and Markov Chains, John Wiley. He is a Life Fellow of the Institute of Electrical and Electronics Engineers and a Golden Core Member of IEEE Computer Society. He has published over 600 articles and has supervised 46 Ph.D. dissertations. He is the recipient of IEEE Computer Society’s Technical Achievement Award for his research on Software Aging and Rejuvenation. His research interests are in reliability, availability, performance and survivability of computer and communication systems and in software dependability. His h-index is 96. He works closely with industry in carrying our reliability/availability analysis, providing short courses on reliability, availability, and in the development and dissemination of software packages such as HARP, SHARPE, SREPT and SPNP.

Contents 6
1 Systems Modelling: Methodologies and Tools 9
References 14
Part I Modelling Theory 16
2 SMVA: A Stable Mean Value Analysis Algorithm for Closed Systems with Load-Dependent Queues 17
2.1 Introduction 17
2.2 Background 18
2.3 Related Work 18
2.4 Stable Mean Value Analysis 19
2.5 Experimental Results 22
2.5.1 One Load-Dependent Queue 24
2.5.2 Two Load-Dependent Queues 27
2.6 Multi-class Extension 30
2.7 Conclusions 31
Appendix 32
References 34
3 Dispatching Discrete-Size Jobs with Multiple Deadlines to Parallel Heterogeneous Servers 35
3.1 Introduction 35
3.2 M/G/1 FCFS Queue with Deadlines 36
3.3 M/D/1 FCFS Queue with Deadlines 38
3.3.1 M/D/1 FCFS with a Single Deadline 38
3.3.2 M/D/1 FCFS with Multiple Job Classes 40
3.4 M/iD/1 FCFS Queue with Deadlines 41
3.4.1 Value Function for the M/iD/1 FCFS Queue 42
3.4.2 Special Case: Systems J2 and B2 with Two Sizes 44
3.4.3 Steady State Performance with J2 and B2 45
3.4.4 Value Functions for J2 and B2 46
3.5 Parallel Servers 48
3.6 Summary 50
References 51
4 Modelling and Efficient Solution of Multiple-Phased Systems 53
4.1 Introduction and Paper Contribution 53
4.2 Extended Phased Petri Nets 55
4.3 X-PPN Solution 57
4.4 An Example of MPS Modelling and Evaluation with X-PPN 60
4.5 GreatSPN vs DEEM on PPN 63
4.6 Conclusions and Future Work 65
References 66
5 Deterministic Network Calculus Analysis of Multicast Flows 68
5.1 Introduction 68
5.2 Deterministic Network Calculus Background 69
5.2.1 Network Analysis 71
5.2.1.1 Tandems of Servers 71
5.2.1.2 Feed-Forward Networks 72
5.2.2 Multicast Flows 72
5.3 Related Work 73
5.3.1 unicastFFA Transformation: A Set of Unicast Flows 73
5.3.2 Multicast TFA 74
5.3.3 Explicit Intermediate Bounds (EIB) 74
5.3.4 Non-network Calculus Approaches 75
5.4 A Multicast Feed-Forward Analysis Procedure 75
5.4.1 Analysis of the Running Example 77
5.4.1.1 mcastFFA Step 1 77
5.4.1.2 mcastFFA Step 2 77
5.4.2 Theoretical Evaluation 78
5.5 Numerical Evaluation 79
5.5.1 Comparison to (Non-)Network Calculus Approaches 79
5.5.2 An Industry-Scale AFDX Data Network 80
5.6 Conclusion and Outlook 81
References 82
6 Modeling Techniques for Pool Depletion Systems 84
6.1 Introduction 84
6.2 Related Work 85
6.3 Scenario 87
6.4 Models Analysis 89
6.4.1 Markov Analysis 89
6.4.2 Discrete Event Simulation 90
6.4.3 Fluid Approximation 92
6.4.4 Techniques Comparison 94
6.5 Conclusions 98
References 98
7 Performance of a Single Server Queue Supported by an Intermittent Server 100
7.1 Introduction 100
7.2 Hypotheses and Model 102
7.3 Steady State Probability Distribution, Mean Number of Customers 103
7.3.1 Steady State Probability Distribution 103
7.3.2 Mean Number of Customers, Mean Waiting Time 105
7.4 Pseudo-Idle and Busy Periods of the Intermittent Server 107
7.4.1 Mean Time of a Passage in the Back Office 107
7.4.2 Mean Time of a Passage in the Front Office 109
7.5 Cost Function 111
7.6 Conclusions 113
Appendix 1: Determination of Eq. (7.11) 114
Appendix 2: Determination of the Mean Number of Customers 115
References 117
8 Simulation from the Tail of the Univariate and Multivariate Normal Distribution 119
8.1 Introduction 119
8.2 Simulation from the Tail of the Univariate Normal 121
8.2.1 Inversion Far in the Right Tail 121
8.2.2 Rejection Methods 123
8.2.3 Speed Comparisons 127
8.3 Simulation from the Tail of the Multivariate Normal 129
8.3.1 Preliminaries and Notation 130
8.3.2 The Rejection Algorithm 130
8.3.3 Asymptotic Efficiency 132
8.4 Conclusion 134
References 135
Part II Applications to Communication Systemsand Infrastructures 137
9 A Comparison of Markov Reward Based Resource-Latency Aware Heuristics for the Virtual Network Embedding Problem 138
9.1 Introduction 138
9.2 Related Work 140
9.3 The VNE Problem 141
9.3.1 Substrate Network 141
9.3.2 Virtual Network Request 141
9.3.3 Virtual Network Embedding 142
9.3.4 Objective 143
9.4 Markov Chain Rewards Based Latency Aware Node Ranking Algorithm 144
9.4.1 MCRR-LA Node Ranking Metric 145
9.4.2 MCRR-LA2 and MCRR-LA3 147
9.5 Markov Chain Rewards Metrics (MCRM) Ranking Algorithm 148
9.5.1 MCRM-I-KSP 148
9.5.2 MCRM-B 149
9.6 Experimental Evaluation 149
9.7 Conclusions 151
References 152
10 Delay Efficient Load Balancing Scheme for Component Carrier Selection in Carrier Aggregation in LTE-A 153
10.1 Introduction 153
10.2 Related Work 154
10.3 Proposed Scheme and Its Performance Model 155
10.3.1 Radio Resource Management Framework for Carrier Aggregation 155
10.3.1.1 CC Selection and Management 155
10.3.1.2 Packet Scheduling 156
10.3.2 Performance Model 157
10.3.3 Feedback Fluid Queueing Model 158
10.4 Performance Analysis 161
10.4.1 Average Buffer Content 162
10.4.2 Mean Throughput 162
10.4.3 Mean Delay 163
10.5 Conclusions and Future Work 164
References 165
11 Modeling Security Requirements for VNE Algorithms:A Practical Approach 166
11.1 Introduction 166
11.2 The Virtual Network Embedding Problem 167
11.3 Problem Description 168
11.3.1 Overview 168
11.3.2 Motivational Example 169
11.3.3 Classification of Requirements 170
11.4 Modeling Security Requirements with Resource/Demand Pairs 171
11.5 Implementation of Security Requirements 172
11.5.1 Implementation of Resource/Demand Pairs 173
11.5.2 Realization of the Motivational Scenario 174
11.6 Evaluation 175
11.6.1 CPU vs. TH 175
11.6.2 CPU vs. Firewall 176
11.6.3 Test Results 177
11.7 Related Work 178
11.8 Conclusion and Future Work 178
References 179
12 Performance Analysis of Data Traffic in Small Cells Networks with User Mobility 181
12.1 Introduction 181
12.2 Generic Queueing Model 182
12.2.1 A PS Queue with Impatience 182
12.2.2 Regularity Properties of the Empty-System Probability 185
12.3 Network with Inter-Cell Mobility 188
12.3.1 A Closed Network of Queues 188
12.3.2 The Case of a Homogeneous Network 190
12.4 Numerical Results 192
12.4.1 Impatience Model 192
12.4.2 Mobility Model 194
12.5 Conclusion 196
References 196
Part III Optimization and Quantitative Evaluation Techniques Applied to Cloud Computing and the Internet of Things 198
13 A Technique to Identify Data Exchange Between Cloud Virtual Machines 199
13.1 Introduction 199
13.2 Reference Scenario 200
13.3 Methodology Description 202
13.3.1 Traces Interpolation and Synchronization 202
13.3.2 Correlation Matrix Creation 203
13.3.3 Identification of Interacting VMs 205
13.3.4 Scalability and Computational Complexity 206
13.4 Implementation 206
13.4.1 Gossip-Based Aggregation 207
13.4.2 Network Agent 207
13.5 Experimental Results 209
13.5.1 Setup Description 209
13.5.2 Correlation Coefficients Analysis 210
13.5.3 Identification of Communicating VMs 211
13.5.4 VMs Clustering 212
13.5.5 Comparison of Time Sampling Intervals 213
13.6 Related Work 214
13.7 Conclusions 215
References 216
14 Container Orchestration: A Survey 218
14.1 Introduction 218
14.2 System Container, Application Container, and Container Manager 220
14.3 Container Orchestration 222
14.4 Reference Architecture 223
14.5 State of the Art 225
14.5.1 Monitoring and Analysis 225
14.5.2 Self-optimization 227
14.5.3 Self-healing 228
14.6 Final Remarks 228
References 229
15 A Cloud-Based Overlay Networking for the Internet of Things: Quantitative Evaluation 233
15.1 Introduction 233
15.2 IoT Infrastructure as a Service 234
15.3 Overlay Networking for IoT 236
15.3.1 Tunneling 237
15.3.2 Layering 238
15.3.3 Server-Less Mesh Implementation 240
15.4 Quantitative Evaluation 241
15.4.1 Scenario 241
15.4.2 Experimental Results 242
15.5 Conclusions 244
References 245
Part IV Tools Development for the Analysis of Specific Areas of Interests 247
16 Markovian Performance Evaluation with BuTools 248
16.1 Introduction 248
16.2 Related Work 249
16.3 Installation, Basic Concepts 249
16.4 Working with PH Distributions 250
16.5 Tools for MAPs 253
16.6 Fitting Tools 254
16.6.1 The trace Package 254
16.6.2 Likelihood Based Fitting 255
16.6.3 Application Example 255
16.7 Analysis of Queues 256
16.7.1 Support for Matrix-Analytic Methods 256
16.7.2 Queueing Models 257
16.7.3 Application Examples 259
16.8 Some Further, Small Packages 261
16.8.1 The moments Package 261
16.8.2 The mc Package 261
16.9 Conclusion 262
References 262
17 J2CBROKER as a Service: A Service Broker Simulation Tool Integrated in OpenStack Environment 264
17.1 Introduction 264
17.2 Motivations 265
17.3 Related Work 266
17.4 The J2CBROKER Simulation Tool 267
17.4.1 SaaS Deployment Model 269
17.4.2 SaaS Benefits 269
17.4.3 OpenStack Integration 269
17.4.4 J2CBROKER Description 270
17.4.4.1 The Client 270
17.4.4.2 The Data Set Simulator 271
17.4.4.3 The Client/Server Communication 271
17.4.4.4 The Server 272
17.4.4.5 The Brokerage Engine 273
17.5 Case Study: Sustainability-Cost Model 273
17.5.1 Scenario 274
17.5.1.1 The Sustainability-Cost Model Simulator 275
17.5.1.2 The Sustainability-Cost Model Engine 275
17.5.2 Experimental Results 277
17.6 Conclusions and Future Work 279
References 279
18 A Software Tool for the Evaluation of Transient Removal Methods in Discrete Event Stochastic Simulations 281
18.1 Introduction 281
18.2 Methods for Transient Removal 283
18.3 fDRIT: A Framework to Detect and Remove Initial Transients 287
18.3.1 Evaluation Parameters for the Quality of the Algorithms 290
18.3.1.1 Accuracy 291
18.3.1.2 Precision 291
18.3.1.3 Computational Cost 292
18.3.1.4 Parameter Estimation 292
18.3.1.5 Overall Benefit Analysis 292
18.4 Example and Results 293
18.5 Conclusion 294
References 294
19 A House Appliances-Level Co-simulation Framework for Smart Grid Applications 296
19.1 Introduction 296
19.2 Related Work 297
19.3 SGsim-Home 299
19.4 Case Study: Integrated Privacy Protection and Demand Response 302
19.4.1 Smooth Consumption 303
19.4.2 Evaluation 305
19.5 Conclusion 306
References 308
Index 311