Advanced Studies Mobile Research Center Bremen

Edited byO. Herzog, C. Görg,M. Lawo,Bremen, Germany

Das Mobile Research Center Bremen (MRC) im Technologie-Zentrum Informatik

erprobt in enger Zusammenarbeit mit der Wirtschaft mobile Informatik-, Infor-mations- und Kommunikationstechnologien. Als Forschungs- und Transferinsti-tut des Landes Bremen vernetzt und koordiniert das MRC hochschulübergreifend eine Vielzahl von interdisziplinären Arbeitsgruppen, die sich mit der Entwicklung und Anwendung mobiler Lösungen beschäft igen. Die Reihe „Advanced Studies“ präsentiert ausgewählte hervorragende Arbeitsergebnisse aus der Forschungstätig-keit der Mitglieder des MRC.

In close collaboration with the industry, the Mobile Research Center Bremen (MRC), a division of the Center for Computing and Communication Technologies (TZI) of the University of Bremen, investigates, develops and tests mobile comput-ing, information and communication technologies. Th is research cluster of the state of Bremen links and coordinates interdisciplinary research teams from diff erent universities and institutions, which are concerned with the development and ap-plication of mobile solutions. Th e series “Advanced Studies” presents a selection of outstanding results of MRC’s research projects.

Edited byProf. Dr. Otthein HerzogProf. Dr. Carmelita GörgProf. Dr. Michael LawoMobile Research Center, Bremen, Germany

und Informationstechnik (TZI) der Universität Bremen erforscht, entwickelt und

Yasir Zaki

Future Mobile Communications

LTE Optimization and Mobile Network Virtualization

RESEARCH

Yasir ZakiBremen, Germany

Dissertation University of Bremen, 2012

ISBN 978-3-658-00807-9 ISBN 978-3-658-00808-6 (eBook) DOI 10.1007/978-3-658-00808-6

Th e Deutsche Nationalbibliothek lists this publication in the Deutsche Nationalbibliografi e;detailed bibliographic data are available in the Internet at http://dnb.d-nb.de.

Library of Congress Control Number: 2012953090

Springer Vieweg© Springer Fachmedien Wiesbaden 2013 Th is work is subject to copyright. All rights are reserved by the Publisher, whether the whole or part of the material is concerned, specifi cally the rights of translation, reprinting, reuse of illustrations, recitation, broadcasting, reproduction on microfi lms or in any other physical way, and transmission or information storage and retrieval, electronic adaptation, compu-ter soft ware, or by similar or dissimilar methodology now known or hereaft er developed. Exempted from this legal reservation are brief excerpts in connection with reviews or scho-larly analysis or material supplied specifi cally for the purpose of being entered and executed on a computer system, for exclusive use by the purchaser of the work. Duplication of this publication or parts thereof is permitted only under the provisions of the Copyright Law of the Publisher’s location, in its current version, and permission for use must always be obtai-ned from Springer. Permissions for use may be obtained through RightsLink at the Copyright Clearance Center. Violations are liable to prosecution under the respective Copyright Law. Th e use of general descriptive names, registered names, trademarks, service marks, etc. in this publication does not imply, even in the absence of a specifi c statement, that such names are exempt from the relevant protective laws and regulations and therefore free for general use. While the advice and information in this book are believed to be true and accurate at the date of publication, neither the authors nor the editors nor the publisher can accept any legal responsibility for any errors or omissions that may be made. Th e publisher makes no warranty, express or implied, with respect to the material contained herein.

Printed on acid-free paper

Springer Vieweg is a brand of Springer DE. Springer DE is part of Springer Science+Business Media.www.springer-vieweg.de

Printed with friendly support of MRC Mobile Research Center, University of Bremen

To my beloved wife Tamara Hamed and to my parents.

Acknowledgments

It would not have been possible to finish this work and write the thesis withoutthe help and support of all the kind people around me, to only some of whom it ispossible to give particular mention here. I would like to thank my beloved wife Ta-mara Hamed, for her unconditional support and great patience. I will forever oweher a debt of gratitude for always believing in me, even in the times when I didn’t.I would also like to thank my parents, who have given me their un-equivalent sup-port, for which my mere expression of thanks likewise does not suffice. And to mylife joy, my kids, Tanya and Yezin.

This thesis would not have seen the light without the help and support of mysupervisor, Prof. Dr. Carmelita Görg. Her wide knowledge, encouragement andpersonal guidance have been of great value to me. I would also like to express mygratitude to Prof. Dr.-Ing. Andreas Timm-Giel, for all his valuable input.

The good advice, support and friendship of my colleague, Dr. Thushara Weera-wardane has been invaluable on both academic and personal level, for which I amextremely grateful. During this work I have collaborated with many colleagues forwhom I have great respect, in particular I wish to extend my warmest thanks toDr. Koojana Kuladinithi and Asanga Udugama for their professional and personalsupport. I would also like to express my gratitude to Dr. Hadeer Hamed and Mr.Khalis Mahmoud Khalis for their efforts in revising the thesis.

I would like to acknowledge all of my friends and colleagues within the Com-Nets department of the University of Bremen. Dr. Xi Li for her help in proofreading the thesis, Liang Zhao for being a great colleague in all the projects weworked together. Lots of gratitude to Dr. Andreas Könsgen, Markus Becker, UmarToseef, Muhammed Mutakin Siddique, Aman Singh, Dr. Bernd-Ludwig Wenning,Martina Kamman and Karl-Heinz Volk for their support. In addition, I would alsolike to thank my students Nikola Zahariev and Safdar Nawaz Khan Marwat fortheir support and good work.

Finally, a special acknowledgment for the DAAD (German Academic ExchangeService), for giving me the chance to come to Germany to do my Master studies,which led eventually to the finishing of this doctoral thesis.

Yasir Zaki

Abstract

Providing QoS while optimizing the LTE network in a cost efficient manner isvery challenging. Thus, radio scheduling is one of the most important functionsin mobile broadband networks. The design of a mobile network radio schedulerholds several objectives that need to be satisfied, for example: the scheduler needsto maximize the radio performance by efficiently distributing the limited radio re-sources, since the operator’s revenue depends on it. In addition, the scheduler hasto guarantee the user’s demands in terms of their Quality of Service (QoS). Thus,the design of an effective scheduler is rather a complex task. In this thesis, the au-thor proposes the design of a radio scheduler that is optimized towards QoS guar-antees and system performance optimization. The proposed scheduler is called“Optimized Service Aware Scheduler” (OSA). The OSA scheduler is tested andanalyzed in several scenarios, and is compared against other known schedulers.

A novel wireless network virtualization framework is also proposed in this the-sis. The framework targets the concepts of wireless virtualization applied withinthe 3GPP Long Term Evolution (LTE) system. LTE represents one of the new mo-bile communication systems that is just entering the market. Therefore, LTE waschosen as a case study to demonstrate the proposed wireless virtualization frame-work. The framework is implemented in the LTE network simulator and analyzed,highlighting the many advantages and potential gain that the virtualization processcan achieve. Two potential gain scenarios that can result from using network virtu-alization in LTE systems are analyzed: Multiplexing gain coming from spectrumsharing, and multi-user diversity gain.

Several LTE radio analytical models, based on Continuous Time Markov Chainsare designed and developed in this thesis. These models target the modelingof three different time domain radio schedulers: Maximum Throughput (MaxT),Blind Equal Throughput (BET), and Optimized Service Aware Scheduler (OSA).The models are used to obtain faster results (i.e., in a very short time period in theorder of seconds to minutes), compared to the simulation results that can take con-siderably longer periods, such as hours or sometimes even days. The model resultsare also compared against the simulation results, and it is shown that it provides agood match. Thus, it can be used for fast radio dimensioning purposes.

X Abstract

Overall, the concepts, investigations, and the analytical models presented in thisthesis can help mobile network operators to optimize their radio network and pro-vide the necessary means to support services QoS differentiations and guarantees.In addition, the network virtualization concepts provides an excellent tool that canenable the operators to share their resources and reduce their cost, as well as pro-vide good chances for smaller operators to enter the market.

Contents

Abstract IX

List of Figures XV

List of Tables XIX

List of Abbreviations XXI

List of Symbols XXV

1 Introduction 1

2 Mobile Communication Systems 5

2.1 Global System for Mobile Communication (GSM) . . . . . . . . 62.2 Universal Mobile Telecommunication System (UMTS) . . . . . . 9

3 Long Term Evolution (LTE) 13

3.1 Motivation and Targets . . . . . . . . . . . . . . . . . . . . . . . 133.2 LTE Multiple Access Schemes . . . . . . . . . . . . . . . . . . . 14

3.2.1 OFDM . . . . . . . . . . . . . . . . . . . . . . . . . . . 143.2.2 OFDMA . . . . . . . . . . . . . . . . . . . . . . . . . . 153.2.3 SC-FDMA . . . . . . . . . . . . . . . . . . . . . . . . . 16

3.3 LTE Network Architecture . . . . . . . . . . . . . . . . . . . . . 173.3.1 User Equipment (UE) . . . . . . . . . . . . . . . . . . . 183.3.2 Evolved UTRAN (E-UTRAN) . . . . . . . . . . . . . . . 183.3.3 Evolved Packet Core (EPC) . . . . . . . . . . . . . . . . 19

3.4 E-UTRAN Protocol Architecture . . . . . . . . . . . . . . . . . . 203.4.1 Radio Link Control (RLC) . . . . . . . . . . . . . . . . . 233.4.2 Medium Access Control (MAC) . . . . . . . . . . . . . . 24

3.4.2.1 Logical and Transport Channels . . . . . . . . . 243.4.2.2 HARQ . . . . . . . . . . . . . . . . . . . . . . 253.4.2.3 Scheduling . . . . . . . . . . . . . . . . . . . . 26

XII Contents

3.4.3 LTE Frame and Physical Resource Structure . . . . . . . 273.5 LTE Quality of Service Bearers . . . . . . . . . . . . . . . . . . . 283.6 Beyond LTE . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30

3.6.1 Wider Bandwidth for Transmission . . . . . . . . . . . . 313.6.2 Advanced MIMO Solutions . . . . . . . . . . . . . . . . 313.6.3 CoMP . . . . . . . . . . . . . . . . . . . . . . . . . . . . 323.6.4 Relays and Repeaters . . . . . . . . . . . . . . . . . . . . 32

4 LTE Network Simulator 35

4.1 Simulation Environment . . . . . . . . . . . . . . . . . . . . . . 354.2 Simulation Framework . . . . . . . . . . . . . . . . . . . . . . . 364.3 Simulation Model . . . . . . . . . . . . . . . . . . . . . . . . . . 37

4.3.1 UE Node Model . . . . . . . . . . . . . . . . . . . . . . 384.3.2 eNodeB Node Model . . . . . . . . . . . . . . . . . . . . 394.3.3 Access Gateway Node Model . . . . . . . . . . . . . . . 404.3.4 Packet Data Network Gateway Node Model . . . . . . . . 404.3.5 Mobility and Channel Node . . . . . . . . . . . . . . . . 414.3.6 Global User Database . . . . . . . . . . . . . . . . . . . 444.3.7 Application Configuration . . . . . . . . . . . . . . . . . 444.3.8 Profile Configuration . . . . . . . . . . . . . . . . . . . . 45

4.4 Traffic Models . . . . . . . . . . . . . . . . . . . . . . . . . . . . 454.4.1 Voice over IP Model (VoIP) . . . . . . . . . . . . . . . . 464.4.2 Web Browsing Model . . . . . . . . . . . . . . . . . . . 474.4.3 Video Streaming Model . . . . . . . . . . . . . . . . . . 484.4.4 File Transfer Model . . . . . . . . . . . . . . . . . . . . 49

4.5 Statistical Evaluation . . . . . . . . . . . . . . . . . . . . . . . . 494.5.1 Confidence Interval Estimation . . . . . . . . . . . . . . . 494.5.2 Independent Replications Method . . . . . . . . . . . . . 50

5 LTE Virtualization 53

5.1 Virtualization . . . . . . . . . . . . . . . . . . . . . . . . . . . . 535.1.1 Server Virtualization . . . . . . . . . . . . . . . . . . . . 545.1.2 Network Virtualization . . . . . . . . . . . . . . . . . . . 56

5.2 4WARD Project . . . . . . . . . . . . . . . . . . . . . . . . . . . 575.2.1 4WARD Virtualization Paradigm . . . . . . . . . . . . . . 57

5.3 Wireless Virtualization in Mobile Communication . . . . . . . . . 595.3.1 Motivation behind Mobile Network Virtualization . . . . . 615.3.2 LTE Virtualization Framework . . . . . . . . . . . . . . . 61

5.3.2.1 Framework Architecture . . . . . . . . . . . . . 62

Contents XIII

5.3.2.2 LTE Hypervisor Algorithm . . . . . . . . . . . 645.3.2.3 Operator Bandwidth Estimation . . . . . . . . . 645.3.2.4 Contract-Based Framework . . . . . . . . . . . 65

5.4 LTE Virtualization Evaluation . . . . . . . . . . . . . . . . . . . 685.4.1 Multiplexing Gain-Based Analysis . . . . . . . . . . . . . 695.4.2 Multi-User Diversity Gain-Based Analysis . . . . . . . . 735.4.3 Contract-Based Framework Analysis . . . . . . . . . . . 76

5.5 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 83

6 LTE Radio Scheduler 85

6.1 LTE Dynamic Packet Scheduling . . . . . . . . . . . . . . . . . . 866.2 LTE MAC Schedulers State of the Art . . . . . . . . . . . . . . . 87

6.2.1 Classical Scheduling Algorithms . . . . . . . . . . . . . . 876.3 Downlink MAC Scheduler Design . . . . . . . . . . . . . . . . . 89

6.3.1 QCI Classification . . . . . . . . . . . . . . . . . . . . . 906.3.2 Time Domain Scheduler (TDS) . . . . . . . . . . . . . . 916.3.3 Frequency Domain Scheduler (FDS) . . . . . . . . . . . . 946.3.4 Link-to-System Mapping (L2S) . . . . . . . . . . . . . . 966.3.5 HARQ Modeling . . . . . . . . . . . . . . . . . . . . . . 98

6.4 Downlink MAC Scheduler Analysis . . . . . . . . . . . . . . . . 996.4.1 OSA vs. Classical Schedulers . . . . . . . . . . . . . . . 99

6.4.1.1 FTP Only Scenario . . . . . . . . . . . . . . . 1006.4.1.2 Mixed Traffic Scenario . . . . . . . . . . . . . 104

6.4.2 GBR Delay Exploitation . . . . . . . . . . . . . . . . . . 1086.5 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 112

7 Analytical Modeling of the LTE Radio Scheduler 113

7.1 General Analytical Model . . . . . . . . . . . . . . . . . . . . . . 1147.1.1 Performance Evaluation . . . . . . . . . . . . . . . . . . 1157.1.2 Generic Departure Rate . . . . . . . . . . . . . . . . . . . 117

7.2 LTE Downlink Scheduler Modeling . . . . . . . . . . . . . . . . 1207.2.1 Single Class Model . . . . . . . . . . . . . . . . . . . . . 1207.2.2 Single Class Model Results . . . . . . . . . . . . . . . . 124

7.2.2.1 Analysis 1 - MaxT . . . . . . . . . . . . . . . . 1257.2.2.2 Analysis 2 - BET . . . . . . . . . . . . . . . . 1297.2.2.3 Analysis 3 - Sensitivity Analysis . . . . . . . . 1327.2.2.4 Analysis 4 - OSA . . . . . . . . . . . . . . . . 1337.2.2.5 Analysis 5 - BET with Random Direction (RD) 134

XIV Contents

7.2.3 Two-Dimensional Model . . . . . . . . . . . . . . . . . . 1367.2.3.1 2D TDS Modeling . . . . . . . . . . . . . . . . 1387.2.3.2 2D Performance Analysis . . . . . . . . . . . . 139

7.2.4 Two-Dimensional Model Results . . . . . . . . . . . . . . 1417.2.4.1 Analysis - w-MaxT . . . . . . . . . . . . . . . 1427.2.4.2 Analysis - w-BET . . . . . . . . . . . . . . . . 1467.2.4.3 Analysis - OSA . . . . . . . . . . . . . . . . . 150

7.3 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 151

8 Conclusions and Outlook 153

A Appendix 159

A.1 Mobility Models . . . . . . . . . . . . . . . . . . . . . . . . . . 159A.2 3GPP Transport Block Size . . . . . . . . . . . . . . . . . . . . . 161A.3 Simulation Results Confidence Interval . . . . . . . . . . . . . . 162

Bibliography 165

List of Figures

2.1 3GPP releases overview [HT09] [Zah11] . . . . . . . . . . . . . . 62.2 GSM network architecture 1 . . . . . . . . . . . . . . . . . . . . 72.3 GSM system evolution . . . . . . . . . . . . . . . . . . . . . . . 92.4 UMTS network architecture 1 . . . . . . . . . . . . . . . . . . . . 10

3.1 LTE EPS network architecture . . . . . . . . . . . . . . . . . . . 133.2 OFDM signal in frequency and time domain [Hoa05] . . . . . . . 153.3 An example of channel dependent scheduling between two users

[ADF+09] . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 163.4 An example comparing SC-FDMA to OFDMA [Agi09] . . . . . . 173.5 LTE E-UTRAN architecture . . . . . . . . . . . . . . . . . . . . 193.6 LTE E-UTRAN user plane protocol stack . . . . . . . . . . . . . 203.7 Detailed LTE downlink protocol architecture [Dah07] . . . . . . . 223.8 Multiple parallel HARQ processes example [Dah07] . . . . . . . 263.9 LTE FDD frame structure [Anr09] . . . . . . . . . . . . . . . . . 273.10 Relationship between slot, symbol and resource blocks [Anr09] . . 283.11 SAE bearer model [HT09] . . . . . . . . . . . . . . . . . . . . . 293.12 LTE-advanced CoMP . . . . . . . . . . . . . . . . . . . . . . . . 323.13 LTE-advanced in-band relay and backhaul [Agi11] . . . . . . . . 33

4.1 OPNET modeler© hierarchical editors . . . . . . . . . . . . . . . 354.2 LTE reference model . . . . . . . . . . . . . . . . . . . . . . . . 364.3 LTE OPNET simulation model . . . . . . . . . . . . . . . . . . . 374.4 UE node model . . . . . . . . . . . . . . . . . . . . . . . . . . . 384.5 eNodeB node model . . . . . . . . . . . . . . . . . . . . . . . . 394.6 aGW node model . . . . . . . . . . . . . . . . . . . . . . . . . . 404.7 PDN-GW node model . . . . . . . . . . . . . . . . . . . . . . . . 414.8 An example result of the fast fading model . . . . . . . . . . . . . 434.9 Sample OPNET application configuration . . . . . . . . . . . . . 444.10 Sample OPNET profile configuration . . . . . . . . . . . . . . . . 454.11 VoIP traffic model . . . . . . . . . . . . . . . . . . . . . . . . . . 46

XVI List of Figures

4.12 VoIP MOS values [IT09] . . . . . . . . . . . . . . . . . . . . . . 474.13 Web traffic model . . . . . . . . . . . . . . . . . . . . . . . . . . 47

5.1 Full virtualization environment [Cha09] . . . . . . . . . . . . . . 555.2 Para virtualization environment [Cha09] . . . . . . . . . . . . . . 555.3 OS virtualization environment [Cha09] . . . . . . . . . . . . . . . 565.4 Network virtualization proposed business model . . . . . . . . . . 585.5 Multiple access schemes . . . . . . . . . . . . . . . . . . . . . . 605.6 LTE virtualization framework architecture . . . . . . . . . . . . . 635.7 The general hypervisor algorithm framework . . . . . . . . . . . 675.8 LTE virtualization simulation model in OPNET . . . . . . . . . . 695.9 Virtual operators allocated bandwidth over simulation time . . . . 715.10 Virtual operator 1 VoIP air interface throughput . . . . . . . . . . 725.11 Virtual operator 1 VoIP application end-to-end delay . . . . . . . 725.12 Virtual operator 1 video application end-to-end delay . . . . . . . 735.13 Virtual operator 1 cell throughput with and without virtualization . 755.14 Virtual operator 1 cell throughput gain due to virtualization . . . . 765.15 Virtual operator allocated bandwidth/PRBs . . . . . . . . . . . . 785.16 Virtual operator 1 video downlink application end-to-end delay . . 795.17 Virtual operator 2 video downlink application end-to-end delay . . 805.18 Virtual operator 2 downlink allocated bandwidth . . . . . . . . . . 805.19 Virtual operator 3 video downlink application end-to-end delay . . 815.20 Virtual operator 3 average FTP download time . . . . . . . . . . . 825.21 Virtual operator 3 average number of FTP Files downloaded . . . 825.22 Virtual operator 4 VoIP downlink application end-to-end delay . . 83

6.1 General packet scheduling framework [HT09] . . . . . . . . . . . 866.2 OSA general scheduler framework . . . . . . . . . . . . . . . . . 906.3 FDS general flow chart . . . . . . . . . . . . . . . . . . . . . . . 956.4 Reference BLER versus SINR AWGN curves . . . . . . . . . . . 966.5 Average user FTP download time . . . . . . . . . . . . . . . . . . 1016.6 Unfairness between users FTP download time . . . . . . . . . . . 1026.7 Average cell throughput comparison . . . . . . . . . . . . . . . . 1036.8 40 UE scenario - scheduler comparison . . . . . . . . . . . . . . 1046.9 Application delay performance comparison between schedulers . . 1066.10 Fairness and cell throughput comparison between schedulers . . . 1076.11 Average cell throughput . . . . . . . . . . . . . . . . . . . . . . . 1096.12 Average VoIP application end-to-end delay . . . . . . . . . . . . 1106.13 Average VoIP MOS value . . . . . . . . . . . . . . . . . . . . . . 110

List of Figures XVII

6.14 non-GBR services performance comparison spider chart . . . . . 111

7.1 General Continuous Time Markov Chain (CTMC) . . . . . . . . . 1147.2 MaxT scheduling of users (example with 7 MCSs) . . . . . . . . 1227.3 BET scheduling of users (example with 7 MCSs) . . . . . . . . . 1227.4 MCSs static probability obtained from simulations . . . . . . . . 1237.5 OSA scheduling of users (example with 7 MCSs) . . . . . . . . . 1247.6 MaxT 10UEs scenario - average FTP download time . . . . . . . 1267.7 MaxT 20UEs scenario - average FTP download time . . . . . . . 1277.8 MaxT Markov chain state probability . . . . . . . . . . . . . . . 1287.9 BET 10UEs scenario - average FTP download time . . . . . . . . 1297.10 BET 20UEs scenario - average FTP download time . . . . . . . . 1307.11 BET Markov chain state probability . . . . . . . . . . . . . . . . 1317.12 MaxT sensitivity analysis results . . . . . . . . . . . . . . . . . . 1327.13 OSA 10UEs scenario - average FTP download time . . . . . . . . 1337.14 OSA Markov chain state probability . . . . . . . . . . . . . . . . 1347.15 BET RD 10UEs scenario - average FTP download time . . . . . . 1357.16 Two-dimensional Markov chain . . . . . . . . . . . . . . . . . . 1367.17 2D Markov chain represented in a single chain . . . . . . . . . . . 1397.18 Example Q mapping with N1=2 and N2=4 . . . . . . . . . . . . . 1397.19 Analysis1 w-MaxT average FTP download time . . . . . . . . . . 1427.20 Analysis2 w-MaxT average FTP download time . . . . . . . . . . 1437.21 Analysis1 w-MaxT Markov chain state probability . . . . . . . . 1447.22 Analysis2 w-MaxT Markov chain state probability . . . . . . . . 1457.23 Analysis3 w-BET average FTP download time . . . . . . . . . . . 1467.24 Analysis3 w-BET Markov chain state probability . . . . . . . . . 1477.25 Analysis4 w-BET average FTP download time . . . . . . . . . . . 1487.26 Analysis4 w-BET Markov chain state probability . . . . . . . . . 1497.27 Analysis5 OSA average FTP download time . . . . . . . . . . . . 150

A.1 Random Way Point (RWP) mobility model . . . . . . . . . . . . 159A.2 Random Direction (RD) mobility model . . . . . . . . . . . . . . 160

List of Tables

3.1 LTE MAC logical channels [36.11b] . . . . . . . . . . . . . . . . 253.2 LTE MAC transport channels [36.11b] . . . . . . . . . . . . . . . 253.3 LTE standardized QCIs and their parameters [SBT09] . . . . . . . 303.4 System performance comparison [Nak09] . . . . . . . . . . . . . 31

4.1 VoIP traffic model parameters [Li10] . . . . . . . . . . . . . . . . 464.2 Web browsing traffic model parameters [Wee11] . . . . . . . . . . 48

5.1 Scenario I simulation configurations . . . . . . . . . . . . . . . . 705.2 Scenario II simulation configurations . . . . . . . . . . . . . . . . 745.3 Scenario III simulation configurations . . . . . . . . . . . . . . . 77

6.1 DSCP/QCI to MAC-QoS-Class mapping example . . . . . . . . . 916.2 An example of QoS weight values for different non-GBR services 936.3 β values for each MCS [KSW+08][LV08a][Val06] . . . . . . . . 986.4 BLER and HARQ transmissions . . . . . . . . . . . . . . . . . . 996.5 Simulation configurations . . . . . . . . . . . . . . . . . . . . . . 1006.6 Simulation configurations . . . . . . . . . . . . . . . . . . . . . . 1056.7 Simulation configurations . . . . . . . . . . . . . . . . . . . . . . 108

7.1 n=2, K=2 combinations example . . . . . . . . . . . . . . . . . . 1197.2 Single class validation general parameters . . . . . . . . . . . . . 1247.3 Single class validation scenarios . . . . . . . . . . . . . . . . . . 1257.4 2-D model validation general parameters . . . . . . . . . . . . . . 1417.5 2-D model validation scenarios . . . . . . . . . . . . . . . . . . . 141

A.1 3GPP transport block size table (subset) . . . . . . . . . . . . . . 161A.2 Analysis 1 - MaxT 10 UEs simulation results confidence interval . 162A.3 Analysis 1 - MaxT 20 UEs simulation results confidence interval . 162A.4 Analysis 2 - BET 10 UEs simulation results confidence interval . . 162A.5 Analysis 2 - BET 20 UEs simulation results confidence interval . . 162

XX List of Tables

A.6 Analysis 3 - BET sensitivity analysis simulation results confidenceinterval . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 162

A.7 Analysis 4 - OSA simulation results confidence interval . . . . . . 163A.8 Analysis 5 - BET RD simulation results confidence interval . . . . 163A.9 Analysis 1 - w-MaxT simulation results confidence interval . . . . 163A.10 Analysis 2 - w-MaxT simulation results confidence interval . . . . 163A.11 Analysis 3 - w-BET simulation results confidence interval . . . . 163A.12 Analysis 4 - w-BET simulation results confidence interval . . . . 163

List of Abbreviations

3GPP 3rd Generation PartnershipProject

AMC Adaptive Modulation andCoding

AODV Ad-hoc On-demand DistanceVector

ARP Allocation and RetentionPriority

ARQ Automatic Repeat Request

AUC Authentication Center

AWGN Additive White GaussianNoise

BET Blind Equal Throughput

BLER Block Error Rate

BSC Base Station Controller

BSS Base Station Subsystem

BTS Base Transceiver Station

CA Carrier Aggregation

CDMA Code Division MultipleAccess

CN Core Network

CoMP Coordinated Multi-Point

CQI Channel Quality Indicator

CTMC Continuous Time MarkovChain

DeNB Donor eNodeB

DL Downlink

DSCP Differentiated Services CodePoint

DVB Digital Video Broadcasting

eNodeB enhanced NodeB

E-UTRAN Evolved Universal TerrestrialRadio Access Network

EDGE Enhanced Data for GSMEvolution

EESM Exponential Effective SINRMapping

EFR Enhanced Full Rate

EMA Exponential Moving Average

EPC Evolved Packet Core

EPS Evolved Packet System

FDD Frequency Division Duplex

FDM Frequency DomainMultiplexing

FDMA Frequency Division MultipleAccess

FDS Frequency Domain Scheduler

FTP File Transfer Protocol

GBR Guaranteed Bit Rate

GMSK Gaussian Minimum ShiftKeying

GPRS General Packet Radio Service

GSM Global System for MobileCommunication

GTP Gateway Tunneling Protocol

XXII List of Abbreviations

HARQ Hybrid Automatic RepeatRequest

HLR Home Location Registry

HSDPA High Speed Downlink PacketAccess

HSPA High Speed Packet Access

HSS Home Subscriber Server

HSUPA High Speed Uplink PacketAccess

HTTP Hypertext Transfer Protocol

IMU International MobileTelecommunication

IP Internet Protocol

ISP Internet Service Provider

IT Information Technology

ITU InternationalTelecommunication Union

LTE Long Term Evolution

MAC Medium Access Channel

MaxT Maximum Throughput

MCS Modulation and CodingScheme

MIESM Mutual Information EffectiveSINR Mapping

MIMO Multi Input Multi Output

MME Mobility Management Entity

MOS Mean Opinion Score

MS Mobile Station

MSC Mobile Switching Center

non-GBR non-Guaranteed Bit Rate

OFDMA Orthogonal FrequencyDomain Multiple Access

OS Operating System

OSA Optimized Service Aware

PCRF Policy and Charging RulesFunction

PDA Personal Digital Assistant

PDCP Packet Data ConvergenceProtocol

PDN Packet Data Network

PDN-GW Packet Data Network Gateway

PDU Protocol Data Unit

PHY Physical Layer

PRB Physical Resource Block

PRBs Physical Resource Blocks

PSK Phase Shift Keying

PSTN Public Switched TelephoneNetwork

QCI QoS Class Identifier

QoS Quality of Service

RD Random Direction

RLC Radio Link Control

RN Relay Node

RNC Radio Network Controller

RRC Radio Resource Control

RWP Random Way Point

S-GW Serving Gateway

SAE System Architecture Evolution

SAN Storage Area Network

SC-FDMA Single Carrier FrequencyDomain Multiple Access

SDMA Space Division MultipleAccess

SDR Software Defined Radio

SIM Subscriber Identity Module

SINR Signal to Interference NoiseRatio

TBS Transport Block Size

TDS Time Domain Scheduler

List of Abbreviations XXIII

TDMA Time Division MultipleAccess

TE Terminal Equipment

TTI Transmission Time Interval

UE User Equipment

UL Uplink

UML User Mode Linux

UMTS Universal MobileTelecommunication System

USIM User Service Identity Module

UTRAN UMTS Terrestrial RadioAccess Network

VLR Visitor Location Registry

VM Virtual Machine

VMM Virtual Machine Monitor

VNet Virtual Network

VNOs Virtual Network Operators

VoIP Voice over Internet Protocol

WCDMA Wideband Code DivisionMultiple Access

WLAN Wireless Local Area Network

List of Symbols

Symbol Meaning

α smoothing factorβ MCS scaling factor

γk[t] normalized average channel condition of bearer kδ 2 varianceη actual number of users served per TTI

θmax maximum achieved throughput if all PRBs are used underperfect channel conditions

θk[t] instantaneous achieved throughput for bearer kθk[t] normalized average throughout of bearer k

λ arrival rate (file inter-arrival time)μ(n) generic departure rate of state nπ(n) state n steady state probability

πππ Markov chain steady state probability vectorτ smoothing factorψ maximum number of users served per TTI

BLEP([γk]) instantaneous Block Error Probability for channel state γk

BLEP([γe f f ]) instantaneous Block Error Probability for channel state γe f f

D mean number of departures by unit timeEtotal total BE PRB estimate over all BE operatorsE(N) average required PRBs at the Nth TTI

Fi operator i fairness factorHOLdelayk head-of-line packet delay for bearer k

K Number of MCSsMCSk kth modulation and coding scheme

n Number of active users per TTInk number of users in MCSk

N Number of users in the systemN0 thermal noise (dB)NF noise floor (dBm)

XXVI List of Symbols

Symbol Meaning

Pk MCSk static probabilityPL path loss (dB)Ptx eNodeB transmission power per PRB (dBm)

PBETk (t) BET scheduler time domain priority factor for bearer k

PGBRk (t) time domain GBR priority metric of bearer k

PMaxTk (t) MaxT scheduler time domain priority factor for user k

PnonGBRk (t) time domain non-GBR priority metric of bearer k

Pw−BETk (t) weighted BET scheduler time domain priority factor for user k

Pw−MaxTk (t) weighted MaxT scheduler time domain priority factor for user k

PRBsAlloci operator i allocated number of PRBsPRBsT T I(N) instantaneous PRB count at the Nth TTI

Q Markov chain infinitesimal generator matrixQ mean number of usersR distance between UE and eNodeB (km)

S(nδ ) slow fading at point nδ (dB)SINRe f f effective SINR mappingSINRk[t] instantaneous SINR value of bearer kSINRi, j Signal to Interference Noise Ratio on PRBi) for user j (dB)

SINRmax scaling factor (maximum achieved SINR)t(α/2,N−1) upper critical value of the t-distribution with N-1 degrees of freedom

to f f traffic model average OFF durationton average ON duration (file download time)

Tavg average download time of all usersTi per-user average download time

T BSk(η) number of bits that can be transmitted by a served UE using MCSk

T BS(n) state n average number of bits transmitted within a TTIT BS(n0, ...,nk) total bits sent by all served users under combination (n0, ...,nk)

UF% Unfairness factor (%)Vi i.i.d. normal random variable

WQoSj QoS weight of the jth MAC QoS classx sample mean

Xc de-correlation distance (m)Xk,i scheduler decision whether a UE is served or not (1 or 0)Xon traffic model average file size

Zα/2 upper α/2 critical value of the standard normal distribution