Muutke küpsiste eelistusi

E-raamat: Network Tomography: Identifiability, Measurement Design, and Network State Inference

, , (Pennsylvania State University), (University of Massachusetts, Amherst)
  • Formaat: PDF+DRM
  • Ilmumisaeg: 27-May-2021
  • Kirjastus: Cambridge University Press
  • Keel: eng
  • ISBN-13: 9781108383783
Teised raamatud teemal:
  • Formaat - PDF+DRM
  • Hind: 97,55 €*
  • * hind on lõplik, st. muud allahindlused enam ei rakendu
  • Lisa ostukorvi
  • Lisa soovinimekirja
  • See e-raamat on mõeldud ainult isiklikuks kasutamiseks. E-raamatuid ei saa tagastada.
Network Tomography: Identifiability, Measurement Design, and Network State InferenceSuurem pilt
  • Formaat: PDF+DRM
  • Ilmumisaeg: 27-May-2021
  • Kirjastus: Cambridge University Press
  • Keel: eng
  • ISBN-13: 9781108383783
Teised raamatud teemal:

DRM piirangud

  • Kopeerimine (copy/paste):

    ei ole lubatud

  • Printimine:

    ei ole lubatud

  • Kasutamine:

    Digitaalõiguste kaitse (DRM)
    Kirjastus on väljastanud selle e-raamatu krüpteeritud kujul, mis tähendab, et selle lugemiseks peate installeerima spetsiaalse tarkvara. Samuti peate looma endale  Adobe ID Rohkem infot siin. E-raamatut saab lugeda 1 kasutaja ning alla laadida kuni 6'de seadmesse (kõik autoriseeritud sama Adobe ID-ga).

    Vajalik tarkvara
    Mobiilsetes seadmetes (telefon või tahvelarvuti) lugemiseks peate installeerima selle tasuta rakenduse: PocketBook Reader (iOS / Android)

    PC või Mac seadmes lugemiseks peate installima Adobe Digital Editionsi (Seeon tasuta rakendus spetsiaalselt e-raamatute lugemiseks. Seda ei tohi segamini ajada Adober Reader'iga, mis tõenäoliselt on juba teie arvutisse installeeritud )

    Seda e-raamatut ei saa lugeda Amazon Kindle's. 

"Providing the first truly comprehensive overview of network tomography - a novel network monitoring approach that makes use of inference techniques to reconstruct the internal network state from external vantage points - this rigorous yet accessible treatment of the fundamental theory and algorithms of network tomography covers the most prominent results demonstrated on real-world data, including identifiability conditions, measurement design algorithms, and network state inference algorithms, alongsidepractical tools for applying these techniques to real-world network management. It describes the main types of mathematical problems, along with their solutions and properties, and emphasizes the actions that can be taken to improve the accuracy of network tomography. With proofs and derivations introduced in an accessible language for easy understanding, this is an essential resource for professional engineers, academic researchers, and graduate students in network management and network science"--

Muu info

A rigorous yet accessible treatment of the fundamental theory and algorithms of network tomography.
Notation ix
Introduction 1(2)
1 Preliminaries
3(7)
1.1 Background on Graphs
3(3)
1.1.1 Graph-Theoretic Definitions
3(1)
1.1.2 Graph Algorithms
4(2)
1.2 Background on Linear Algebra
6(1)
1.3 Background on Parameter Estimation
6(2)
1.3.1 Fisher Information Matrix and the Cramer-Rao Bound
7(1)
1.3.2 Maximum Likelihood Estimator
7(1)
1.4 Background on Routing Mechanisms
8(1)
References
8(2)
2 Fundamental Conditions for Additive Network Tomography
10(33)
2.1 Problem Setting
10(1)
2.2 Algebraic Condition for General Measurements
11(2)
2.2.1 Illustrative Example
12(1)
2.2.2 Limitations of Algebraic Conditions
13(1)
2.3 Topological Condition under Arbitrarily Controllable Routing
13(1)
2.4 Topological Condition under Controllable Cycle-Based Routing
14(6)
2.4.1 Identifiability with One Monitor
14(2)
2.4.2 Identifiability with Multiple Monitors
16(1)
2.4.3 Robust Identifiability
17(3)
2.5 Topological Condition under Controllable Cycle-Free Routing
20(21)
2.5.1 Unidentifiability with Two Monitors
20(3)
2.5.2 Identifiability of Interior Links with Two Monitors
23(6)
2.5.3 Complete Identifiability Condition
29(4)
2.5.4 Determination of Partial Identifiability
33(8)
References
41(2)
3 Monitor Placement for Additive Network Tomography
43(35)
3.1 Monitor Placement under Controllable Cycle-Based Routing
43(7)
3.1.1 Monitor Placement for Complete Identifiability
43(2)
3.1.2 Robust Monitor Placement
45(5)
3.2 Monitor Placements under Controllable Cycle-Free Routing
50(26)
3.2.1 Monitor Placement for Complete Identifiability
50(6)
3.2.2 Monitor Placement for Maximal Identifiability
56(8)
3.2.3 Preferential Monitor Placement
64(8)
3.2.4 Robust Monitor Placement
72(4)
References
76(2)
4 Measurement Path Construction for Additive Network Tomography
78(24)
4.1 Path Construction for Cycle-Based Measurements
78(3)
4.2 Path Construction for Cycle-Free Measurements
81(12)
4.2.1 Algorithm Design
81(2)
4.2.2 Spanning Tree-Based Path Construction
83(3)
4.2.3 Spanning Tree-Based Link Identification
86(1)
4.2.4 Complexity Analysis
87(1)
4.2.5 An Example
88(1)
4.2.6 Performance Evaluation
89(4)
4.3 Robust Path Construction
93(7)
4.3.1 Problem Formulation
94(1)
4.3.2 Properties and Algorithm
95(2)
4.3.3 Conditions for Optimality of RoMe
97(1)
4.3.4 Approximating the Expected Rank
97(1)
4.3.5 Performance Evaluation
98(2)
4.4 Conclusion
100(1)
References
101(1)
5 Fundamental Conditions for Boolean Network Tomography
102(36)
5.1 Problem Setting
103(2)
5.2 Conditions for k-Identifiability
105(11)
5.2.1 Abstract Conditions for k-Identifiability
105(1)
5.2.2 Verifiable Conditions for k-Identifiability
106(10)
5.3 Measure of Maximum Identifiability
116(4)
5.3.1 Abstract Bounds on Maximum Identifiability
116(1)
5.3.2 Computable Bounds on Maximum Identifiability
117(2)
5.3.3 Evaluation of Maximum Identifiability
119(1)
5.4 Generalization to Preferential Boolean Network Tomography
120(12)
5.4.1 Generalized Identifiability Measures
120(3)
5.4.2 Generalized k-Identifiability Conditions
123(5)
5.4.3 Generalized Maximum Identifiability Bounds
128(3)
5.4.4 Evaluation of Generalized Maximum Identihability
131(1)
5.5 Other Types of Failures
132(5)
5.5.1 Node and Link Failures
132(1)
5.5.2 Only Link Failures
133(4)
References
137(1)
6 Measurement Design for Boolean Network Tomography
138(36)
6.1 Monitor Placement Problem
138(11)
6.1.1 Monitor Placement for Link Failure Localization
139(4)
6.1.2 Monitor Placement for Node Failure Localization
143(6)
6.2 Beacon Placement Problem
149(5)
6.2.1 Routing Assumptions
149(1)
6.2.2 Beacon Placement under Single-Link Failures
150(3)
6.2.3 Beacon Placement under Arbitrary-Link Failures
153(1)
6.3 Path Construction Problem
154(11)
6.3.1 Path Selection under Uncontrollable Routing
154(4)
6.3.2 Path Construction under Controllable Routing
158(7)
6.3.3 Bounds on Number of Paths
165(1)
6.4 Service Placement Problem
165(7)
6.4.1 Monitoring-Aware Service Placement
166(2)
6.4.2 Hardness and Approximation Algorithms
168(3)
6.4.3 Performance Evaluation
171(1)
References
172(2)
7 Stochastic Network Tomography Using Unicast Measurements
174(27)
7.1 Problem Formulation
175(2)
7.1.1 Network Model
175(1)
7.1.2 Stochastic Link Metric Tomography
175(1)
7.1.3 Loss Tomography Using Unicast Measurements
175(1)
7.1.4 Packet Delay Variation Tomography
176(1)
7.1.5 Measurement Design
176(1)
7.2 Identihability and Invertibility of Fisher Information Matrix
177(1)
7.3 Link Parameter Estimation
178(2)
7.3.1 Maximum Likelihood Estimator for Packet Loss Tomography
178(1)
7.3.2 Maximum Likelihood Estimator for Packet Delay Variation Tomography
179(1)
7.4 Measurement Design
180(7)
7.4.1 D-Optimal Design
180(3)
7.4.2 A-Optimal Design
183(1)
7.4.3 Weighted A-Optimal Design
184(1)
7.4.4 Application to Loss/PDV Tomography
185(2)
7.5 Experiment Design Algorithms
187(4)
7.5.1 Closed-Form Solutions for |P| = |L|
188(1)
7.5.2 Heuristic Solution for |P| > |L|
188(1)
7.5.3 Iterative Design Algorithm
189(2)
7.6 Performance Evaluation
191(8)
7.6.1 Dataset for Evaluation
192(1)
7.6.2 Evaluation of Loss Tomography
192(5)
7.6.3 Evaluation of Packet Delay Variation Tomography
197(2)
7.7 Conclusion
199(1)
References
200(1)
8 Stochastic Network Tomography Using Multicast Measurements
201(17)
8.1 Loss Tomography Using Multicast Measurements
202(2)
8.1.1 Loss Model
203(1)
8.1.2 Data, Likelihood, and Inference
203(1)
8.2 The Maximum Likelihood Estimator
204(2)
8.2.1 Evaluation
205(1)
8.3 Multicast versus Unicast for Link Loss Tomography
206(4)
8.3.1 Measurement Design
208(1)
8.3.2 Evaluation
208(2)
8.3.3 Summary
210(1)
8.4 Inference of Multicast Trees
210(6)
8.4.1 Evaluation
214(2)
8.5 Discussion
216(1)
References
217(1)
9 Other Applications and Miscellaneous Techniques
218(8)
9.1 Network Topology Tomography
218(2)
9.1.1 Techniques Based on Multicast Probing
218(1)
9.1.2 Techniques Based on Unicast Probing
219(1)
9.1.3 Techniques for Non-tree Topologies
219(1)
9.1.4 Future Directions
219(1)
9.2 Traffic Matrix Tomography
220(1)
9.3 Miscellaneous Techniques
221(1)
9.3.1 Range-Based or Bound-Based Network Tomography
221(1)
9.3.2 Network Coding-Based Network Tomography
221(1)
9.4 Practical Issues and Future Directions
222(1)
References
223(3)
Appendix Datasets for Evaluations 226(4)
Index 230
Ting He is an Associate Professor in the School of Electrical Engineering and Computer Science at The Pennsylvania State University. She is a Senior Member of the IEEE. Liang Ma is a Research Scientist in the AI Group at Dataminr Inc. Ananthram Swami is the Senior Research Scientist for Network Science at the US Army's CCDC Army Research Laboratory. He is an ARL Fellow and a Fellow of the IEEE. Don Towsley is a Distinguished Professor of Computer Science at the University of Massachusetts, Amherst. He is a Fellow of the IEEE and ACM.
Lisainfo e-raamatute kohta Lisainfo e-raamatute kohta
Püsilink: https://www.kriso.ee/db/97811083837832e.html
Märksõnad: