Design, analysis and simultion for optical access and wide-area networks.

(1)

KTH Information and Communication Technology

Design, Analysis and Simulation of Optical Access and Wide-area Networks

JIAJIA CHEN

Doctoral Thesis in Microelectronics and Applied Physics Stockholm, Sweden 2009

(2)

Akademisk avhandling som med tillstånd av Kungl Tekniska Högskolan framlägges till offentlig granskning för avläggande av teknologie doktorsexamen fredag den 8 maj 2009 klockan 10:00 i C1, Electrum 1, Isafjordsgatan 20, Kista, Stockholm

©

Jiajia Chen, May 2009

Tryck: Universitetsservice US AB TRITA-ICT/MAP AVH Report 2009:1 ISSN 1653-7610

ISRN KTH/ICT-MAP/AVH-2009:1-SE

KTH School of Information and Communication Technology SE-164 40 Kista Sweden

(3)

iii

Abstract

Due to the tremendous growth of traffic volume caused by both exponential increase of number of Internet users and continual emergence of new bandwidth demanding applications, high capacity networks are required in order to satisfactorily handle the extremely large amount of traffic. Hence, optical fiber communication is the key technology for the network infrastructure. This thesis addresses design, analysis and simulation of access and core networks targeting important research problems, which need to be tackled for the effective realization of next generation optical networks.

Among different fiber access architectures, passive optical network (PON) is considered as the most promising alternative for the last mile connection due to its relatively low cost and resource efficiency. The inherent bursty nature of the user generated traffic results in dynamically changing bandwidth demand on per subscriber basis. In addition, access networks are required to support differentiated quality of service and accommodate multiple service providers. To address these problems we proposed three novel scheduling algorithms to efficiently realize dynamic bandwidth allocation in PON, along with guaranteeing both the priority and fairness of the differentiated services among multiple users and/or service providers. Meanwhile, because of the increasing significance of reliable access to network services, an efficient fault management mechanism needs to be provided in PON. In addition, access networks are very cost sensitive and the cost of protection should be kept as low as possible. Therefore, we proposed three novel cost-effective protection architectures keeping in mind that reliability requirement in access networks should be satisfied at the minimal cost.

Regarding the optical core networks, replacing electronic routers with all-optical switching nodes can offer significant advantages in realizing high capacity networks. Because of the technological limitations for realizing all-optical nodes, the focus is put on the ingenious architecture design. Therefore, we contributed on novel switching node architectures for optical circuit and packet switching networks. Furthermore, we addressed different aspects of routing and wavelength assignment (RWA) problem, which is an important and hard task to be solved in wavelength routed networks. First, we proposed an approach based on the information summary protocol to reduce the large amount of control overhead needed for dissemination of the link state information in the case of adaptive routing. In addition, transparency in optical networks may cause vulnerability to physical layer attacks. To target this critical security related issue, we proposed an RWA solution to minimize the possible reachability of a jamming attack.

Finally, in order to evaluate our ideas we developed two tailor-made simulators based on discrete event driven system for the detailed studies of PON and switched optical networks.

Moreover, the proposed tabu search heuristic for our RWA solution was implemented in C++.

Key words: fiber access networks, passive optical network, dynamic bandwidth allocation, reliability, switched optical networks, switching node, optical circuit switching, optical packet switching, routing and wavelength assignment, security.

(4)

(5)

v

Acknowledgements

First and foremost, I would like to thank my supervisor Prof. Lena Wosinska, under whose kind supervision my research work proceeded in the right direction. Besides, she is a precious mentor and enthusiastic friend. Her support was often beyond my doctoral work, e.g. she was the chief witness of my wedding ceremony.

I would like to express my grateful thanks to Prof. Sailing He, my supervisor in China, who led me to the research world and provided a valuable opportunity of my study at The Royal Institute of Technology (KTH). I would also like to thank Prof. Lars Thylén for making this work possible. In addition, I want to thank China Scholarship Council (CSC) for providing partial financial support for my studies abroad, as well as Networks of Excellence BONE and e-Photon/ONe+ for funds of the cooperation with other European research groups.

Especially, I would like to thank Prof. Biao Chen, Dr. Paolo Monti, and Amornrat Jirattigalachote, for their insightful discussion on my research work. Special thanks to Jawwad Ahmed, who helped me a lot on the work of this thesis.

I want to thank all the co-authors of each paper of this thesis for our nice collaborations. I would also like to thank all my colleagues in Photonics and Microwave Engineering (FMI) group for the friendly atmosphere in the lab. Thanks to all my Chinese friends, as they gave me a lot of happy memories during these years.

Finally, I would like to thank my parents for supporting and encouraging me throughout my whole life. Thanks to my dear husband Clark for his everlasting love.

Jiajia Chen

Stockholm, March 2009

(6)

(7)

vii

List of Papers

List of papers included in the thesis:

I. Jiajia Chen, Biao Chen, and Sailing He, “A Novel Algorithm for Intra-ONU Bandwidth Allocation in Ethernet Passive Optical Networks”, IEEE Communications Letters, vol. 9, pp. 850-852, Sep. 2005.

II. Biao Chen, Jiajia Chen, and Sailing He, “Efficient and Fine Scheduling Algorithm for Bandwidth Allocation in Ethernet Passive Optical Networks”, IEEE J. Selected Topics in Quantum Electronics, vol. 12, pp. 653 – 660, Jul-Aug. 2006.

III. Jiajia Chen, Biao Chen, and Lena Wosinska, “A Novel Joint Scheduling Algorithm for Multiple Services in 10G EPON”, SPIE APOC Asia-Pacific Optical Communication, 2008. (Best student paper award)

IV. Jiajia Chen, Biao Chen, and Sailing He, “Self-protection Scheme against Failures of Distributed Fiber Links in An Ethernet Passive Optical Network”, OSA Journal of Optical Networking, vol. 5, pp. 662-666, Sep. 2006.

V. Jiajia Chen, and Lena Wosinska, “Protection Schemes in PON Compatible with Smooth Migration from TDM-PON to Hybrid WDM/TDM PON”, OSA Journal of Optical Networking, vol. 6, pp. 514-526, May 2007.

VI. Jiajia Chen, Lena Wosinska, and Sailing He, “High Utilization of Wavelengths and Simple Interconnection between Users in A Protection Scheme for Passive Optical Networks”, IEEE Photonics Technology Letters, vol. 20, pp. 389-391, Mar. 2008.

VII. Lena Wosinska and Jiajia Chen, "Reliability Performance Analysis vs. Deployment Cost of Fiber Access Networks", 7th International Conference on Optical Internet, COIN’08, 2008

VIII. Jiajia Chen, Lena Wosinska, Lars Thylén and Sailing He, “Novel Architectures of Asynchronous Optical Packet Switch”, 33rd European Conference and Exhibition on Optical Communication ECOC’07, 2007.

IX. Jiajia Chen, Amornrat Jirattigalachote, Lena Wosinska, and Lars Thylén, “Novel Node Architectures for Wavelength-Routed WDM Networks with Wavelength Conversion Capability” 34th European Conference and Exhibition on Optical Communication ECOC’08, 2008.

(10)

x

X. Jiajia Chen, Lena Wosinska, Marco Tacca, and Andrea Fumagalli, “Dynamic Routing Based on Information Summary-LSA in WDM Networks with Wavelength Conversion”, Transparent Optical Networks 10th International Conference on, ICTON '08, 2008.

XI. Nina Skorin-Kapov, Jiajia Chen and Lena Wosinska, “A Tabu Search Algorithm for Attack-Aware Lightpath Routing”, Transparent Optical Networks 10th International Conference on, ICTON '08, 2008.

List of related papers not included in the thesis:

1 Changjian Guo, Jiajia Chen, Dawei Wang, Meng Jiang, and Biao Chen, "Experimental Demonstration of A Hybrid 1/2-dimentional En/Decoding Optical Code Division Multiple Access System", SPIE APOC Asia-Pacific Optical Communication, Oct. 2008.

2 Jiajia Chen, Lena Wosinska, Miroslaw Kantor, and Lars Thylén, “Comparison of Hybrid WDM/TDM Passive Optical Networks (PONs) with Protection”, 34th European Conference and Exhibition on Optical Communication ECOC’08, 2008.

3 Lena Wosinska, Jiajia Chen and Mas Machuca, “Techno-economical Evaluation of Selected Passive Optical Network Architectures” Transparent Optical Networks 10th International Conference on, ICTON '08, 2008.

4 Miroslaw Kantor, Jiajia Chen, Lena Wosinska and Krzysztof Wajda, “Techno-economic Analysis of PON Protection Schemes” BroadBand Europe, 2007.

5 Lena Wosinska, JiaJia Chen, Miroslaw Kantor and Krzysztof Wajda, “Reliability and Cost Analysis of Passive Optical Networks”, ICTON-MW'07 RTON, 2007.

6 Jiajia Chen, Lena Wosinska and Sailing He, “A Novel Protection Scheme for Hybrid WDM/TDM PONs”, SPIE APOC Asia-Pacific Optical Communication, 2007 (Best student paper award).

7 Biao Chen, Changjian Guo, Jiajia Chen, Lingjian Zhang, Meng Jiang and Sailing He,

“Add/drop Multiplexing and TDM Signal Transmission in An Optical CDMA Ring Network”, OSA Journal of Optical Networking, vol. 6, pp. 969-974, Jul. 2007.

8 Lena Wosinska and Jiajia Chen, “Reliability Performance of Passive Optical Networks”, Transparent Optical Networks 9th International Conference on, ICTON '07, 2007.

9 Jiajia Chen, and Lena Wosinska, “Performance Analysis of Protection Schemes Compatible with Smooth Migration from TDM-PON to Hybrid WDM/TDM-PON”, Optical Fiber Communication conference OFC’07, 2007.

(11)

xi

10 Lena Wosinska, and Jiajia Chen, “Contention Resolution in An Asynchronous All- optical Packet Switch”, International conference on photonics in switching, 2006.

11 Jiajia Chen, Biao Chen, and Sailing He, “A Novel Hierarchical Algorithm for Intra-ONU Scheduling in An Ethernet Passive Optical Network”, SPIE APOC Asia-Pacific Optical Communication, 2005.

12 Xiang Lu, Jiajia Chen and Sailing He: “Wavelength Assignment Method of WDM Network of Star Topology”, Electronics Letters, vol. 40, pp. 625- 626, May 2004.

(12)

(13)

xiii

Acronyms

AGC Automatic Gain Control AON Active Optical Network ATM Asynchronous Transfer Mode AWG Arrayed Waveguide Gratings CAPEX CAPital EXpenditure

CDR Clock-and-Data Recovery

CO Central Office

DBA Dynamic Bandwidth Allocation

DF Distributed Fiber

DFG Different Frequency Generation DPM Differential-Phase-Modulation DSL Digital Subscriber Loop EAM Electro-Absorption Modulator

EIT Electromagnetically Induced Transparency EPON Ethernet Passive Optical Network

FCFS First Come First Serve

FF Feeder Fiber

FQSE Fair Queuing with Service Envelop

FSR Free Spectrum Range

FTTH Fiber To The Home

FWM Four-Wave Mixing

GEM General Encapsulation Method GPON Gigabit Passive Optical Network HDTV High-Definition TeleVision

IEEE Institute of Electrical and Electronics Engineers

IF Interconnection Fiber

ILP Integer Linear Program

IP Internet Protocol

IPACT Interleaved Polling with Adaptive Cycle Time

IS Information Summary

ITU-T International Telecommunication Union-Telecommunication standardization sector

JAR Jamming Attack Radius

LRD Long-Range Dependence

LSA Link State Advertisement

MAC Media Access Control

MEMS MicroElectroMechanical System MPCP Multiple-Point Control Protocol MSFQ Modified Start-time Fair Queuing

MTB Modified Token Bucket

(14)

xiv

NP Neighboring Protection

O/E/O Optical/Electrical/Optical

OAM Operation, Administration and Maintenance OBS Optical Burst Switching

OCS Optical Circuit Switching ODN Optical Distribution Network OLT Optical Line Terminal

ONU Optical Network Unit

OPEX OPerational EXpenditure OPS Optical Packet Switching

OS Optical Switch

OXC Optical CrossConnect

P2P Point-to-Point

PON Passive Optical Network

QoS Quality of Service

RAM Random Access Memory

RBD Reliability Block Diagram

RN Remote Node

RWA Routing and Wavelength Assignment SBA Static Bandwidth Allocation

SLA Service Level Agreement

SOA Semiconductor Optical Amplifier SONET Synchronous Optical NETwork

SP Shortest Path

SRD Short-Range Dependence

T/R Transceiver

T-CONT Transport CONTainer TDM Time-Division Multiplexing TWC Tunable Wavelength Converter

WA Wavelength Assignment

WC Wavelength Converter

WDM Wavelength-Division Multiplexing WPON WDM broadcast and select PON

WRPON Wavelength Routed PON

XAM Cross-Absorption-Modulation

XGM Cross-Gain-Modulation

XPM Cross-Phase-Modulation

(15)

1

Chapter 1 Introduction

1.1 High capacity networks

We are witnessing a rapid growth in Internet traffic volume on a yearly basis, and it is expected that this trend will continue in the future. Driving force behind this uphill trend can be attributed to advances in personal computers, Internet and telecommunication technologies.

Our routine activities these days involve a frequent use of many bandwidth demanding applications. Therefore, high capacity networks are needed in order to satisfy the tremendous growth in bandwidth requirements, both in terms of the increased number of online users and continual emergence of new online applications. Optical fiber communication is the key technology to realize these high capacity network infrastructures, which are required to “keep pace” with this global trend of exponential increase in traffic volume.

A traditional telecommunication network follows a hierarchical structure, and can be divided into the access, metro and core (wide-area) parts. For the last mile bottleneck, the fiber access network is a proven solution that can offer significantly high bandwidth and long reach. So far the metro networks form an intermediate part that connects the access and core networks. Due to the trend towards convergence in the network, metro infrastructure could be replaced by the long reach fiber access solution that can extend directly from end users to the core network.

As for the wide-area part, the switched optical network is a promising solution to realize huge bandwidth demands. With this in mind, this thesis focuses on some specific research problems in the context of the fiber access and switched optical networks.

1.1.1 Fiber access networks

Due to both economical and practical reasons the significance of broadband communication for the community is growing rapidly triggering an explosion of fiber access network deployments, and consequently providing great business opportunities for both system and network providers.

On the other hand, bandwidth demanding applications, such as high-definition television (HDTV), real-time interactive gaming, telemedicine, broadband Internet service, etc as well as user behavior (always on) are creating a new challenge of efficiently and flexibly providing ultra-high bandwidth in the access networks.

Several broadband access technologies exist today, such as copper based digital subscriber loop (DSL), wireless and fiber access. However, the fiber access is the only viable technology for the future access networks [1-5]. Fiber-to-the-home (FTTH) is the future-proof technology offering ultra-high bandwidth and long reach. Several fiber access network architectures have been developed, e.g., point-to-point (P2P), active optical network (AON) [6] and passive optical network (PON) [7-11]. However, PON is considered as the most promising solution

(16)

2 Chapter 1. Introduction due to the relatively low deployment cost and resource efficiency. Furthermore, based on two resource sharing technologies of time-division multiplexing (TDM) and wavelength-division multiplexing (WDM), there are three main types of PONs, namely TDM PON, WDM PON and hybrid WDM/TDM PON.

The common traffic characteristic for access network is not uniform. The load of diverse channels is variable not only in distinct time periods but also at different geographic locations.

In addition, differentiated quality of service (QoS) provisioning requirements are demanded by various customers. Therefore, the scheduling algorithm design for flexible bandwidth allocation as well as service level guarantees becomes a crucial issue in PONs. In Papers I, II and III of this thesis, some novel algorithms are proposed to realize fair bandwidth scheduling between end users along with priority guarantee for different traffic classes.

Meanwhile, fault management is also important for the reliable service delivery and business continuance. Network operators need to guarantee the level of connection availability specified in the service level agreement (SLA). Furthermore, access networks are very cost- sensitive due to the low resource sharing factor. Therefore, it is important in PON deployment to minimize the cost of protection while maintaining the connection availability at an acceptable level. With this in mind, cost efficient protection schemes are proposed in Papers IV-VI for TDM PON and hybrid WDM/TDM PON. In Paper VII of this thesis, the availability and deployment cost analysis of different fiber access networks is done in order to select the most efficient solution.

1.1.2 Switched optical networks

The concept of optical transparency is widely discussed, e.g. in [12, 13]. Transparency refers to the property of an optical network to show independence with respect to a number of characteristics, such as bit rate, protocol, and modulation format. Optical transparent networks, based on WDM technology, seem to be the most promising candidates for future high capacity long distance communication. In such networks, switching functions will be carried out directly in the optical domain so that high speed optical signals can travel through the network without any optical-to-electrical conversion. Different switching paradigms [14] can be applied to exploit the optical technology in terms of different switching granularities.

These are:

• Optical circuit switching

• Optical burst switching

• Optical packet switching

1.1.2.1 Optical circuit switching

Circuit switching has been used in telephone networks for a long time. In this classical approach, a physical circuit is established for the complete duration of the connection from the source to the destination and the reserved resources cannot be shared. The traditional (electronic) circuit switching can be performed by space switching, time switching or usually a combination of both. The optical circuit switching (OCS) paradigm (mostly at wavelength level) is a technique to offer huge bandwidth in the backbone part of the network [12]. This approach provides access to bandwidth with a coarse granularity. An OCS network can also be referred to as wavelength routed network. It provides end-to-end optical channels

(17)

1.2. Discrete event driven simulation 3 (lightpaths) between source and destination nodes. Lightpath can be set up and torn down on

request. One of the most important challenges is solving routing and wavelength assignment (RWA) problem, which consists of finding a suitable physical route for each lightpath request, and assigning an available wavelength to that route. Demands to set up lightpaths may be known in advance and set up semi-permanently (static or off-line), or can arrive in a stochastic manner with random holding times (dynamic or on-line). In the static case, the common objective of RWA is to minimize the resources (such as number of wavelengths or number of fibers) that will be needed to support all the lightpaths in the network, while in the dynamic scenario lightpath blocking probability is a major performance characteristic. A suitable OCS node (referred to as optical crossconnect OXC) architecture can significantly improve the blocking performance. To address this issue, a novel OXC architecture is proposed and evaluated in Paper VIII of this thesis. Furthermore, we present some related work on RWA in Paper X and XI of this thesis.

1.1.2.2 Optical burst switching

In contrast to OCS, optical burst switching (OBS) [15] is based on statistical multiplexing, which can increase the efficiency of network resource utilization. OBS networks mainly consist of two types of switching nodes, namely edge and core nodes. The edge node can aggregate client data, e.g., Internet protocol (IP) packets into bursts. Each burst has an associated control packet. Usually, a burst is separated from the control packet by the interval of offset time. The main functions of the edge nodes are assembly/disassembly of optical burst, and decision of offset time and burst size. The OBS core nodes perform control packet lookup, optical crossconnecting and data burst monitoring. Compared with the edge nodes, the core nodes can have relatively simple structure.

1.1.2.3 Optical packet switching

In optical packet switching (OPS), packets are buffered and routed in the optical domain. OPS may become a competitive solution in the future for the high capacity wide-area networks [12]. In contrast to OCS and OBS, OPS networks have the switching granularity on the packets level, and can realize most flexible and efficient bandwidth management. The functionality of OPS node should include: decoding packet header, (can be electronic if the packet header is encoded at lower bit rates), configuring a switch fabric (the reconfiguration needs to be performed very fast in nanosecond range), synchronization (for synchronous OPS nodes), multiplexing, and contention resolution. A lot of existing research has focused on the architecture design of OPS nodes with efficient contention resolutions, e.g. in [16-17].

Regarding this, in Paper IX of this thesis, we propose and evaluate two novel architectures of OPS nodes based on a few controllable buffers and dedicated or shared wavelength converters.

1.2 Discrete event driven simulation

In many cases it is hard to develop an analytical model for performance evaluation of optical networks, due to the complicated system structures and complex traffic patterns. Therefore, discrete event driven simulation can be a feasible and efficient way to evaluate network and system performance. Two simulation tools based on discrete event system are developed in the framework of this thesis. One of them is for evaluation of scheduling algorithms in TDM PON and is meant to support our research on fiber access networks. The second one is for our

(18)

4 Chapter 1. Introduction study of switched optical networks to test performance of novel switching nodes for OCS and OPS along with our proposed link state advertisement protocol in OCS networks.

In discrete event simulation the operation of a system is represented as a chronological sequence of events. Each event occurs at an instant in time and marks a change of state in the system [18]. For example, for the simulation of RWA algorithm, an event could be “the arrival of a lightpath request”, with the resulting system state of “RWA algorithm being implemented to find an available route and wavelength for this lightpath request”, and eventually (unless a failure needs to be simulated in an OCS network) “route and wavelength being assigned or the lightpath request being blocked if no available resource can be found”.

Typically, the discrete event simulation consists of five key components, namely, clock, event-lists, random number generator, statistics and end condition.

Clock

In contrast to the continuous time simulations, time “hops” (i.e. the clock skips to the next event start time directly during the simulation proceeding), because events are always instantaneous.

Event-lists

It needs to maintain at least one list of simulation events, i.e., pending event set. Each event in the list is described by the time at which it occurs and a type, indicating the code that will be used to simulate that event.

Random number generator

The simulation needs to generate random variables, depending on the system model. Usually, this can be accomplished by one or more pseudorandom number generators. Different from the true random numbers, the use of pseudorandom numbers is convenient for the simulation that needs a rerun with exactly the same traffic by using the same seed.

Statistics

The simulation needs to keep track of the system’s statistics, which quantify the aspects of interest. For instance, in the scheduling algorithm simulation in EPON, the interesting performance parameters are throughput, delay, jitter, etc.

End condition

Theoretically a discrete event simulation could run forever. So end condition is used to decide when the simulation will be terminated. Typical choices are “at time t” or “after processing n number of events” or, “some important statistical measure reaching the desired confidence level”.

1.3 Organization of the thesis

This thesis contributes on the design, analysis and simulation of optical access and wide-area networks. It is based on the research papers that have been published in international research journals and conferences. The covered topics target some specific research problems in the fiber access and switched optical networks. The introduction to the subject area covered by this thesis has already been presented in this chapter. Remainder of this thesis are divided in

(19)

1.3. Organization of the thesis 5 two parts, namely, fiber access networks and switched optical networks along with the related

evaluation methodology.

Part I addresses fiber access networks and consists of four chapters. Description of P2P, AON and PON architectures is given in Chapter 2. The remainder of Part I focus on PON. Chapter 3 provides a brief review of PON topologies and different multiplexing techniques including TDM PON, WDM PON and hybrid WDM/TDM PON. In Chapter 4, first the related work on single-level and hierarchical scheduling algorithms for dynamic bandwidth allocation is presented. Then, the main advantages of the algorithms proposed in Papers I, II and III of this thesis are reviewed. In the end of this chapter, a general discussion on dynamic resource allocation in next generation PONs is provided. Finally, Chapter 5 focuses on the cost- effective protection aiming to compare cost and reliability performance of some representative PON architectures. The evolution of the PON protection schemes including our contribution in Papers IV-VI of this thesis is reviewed, and then the reliability and cost performance are evaluated by using the method proposed in Paper VII of this thesis.

There are four chapters in Part II which is related to switched optical networks. First, Chapter 6 reviews three switching paradigms, namely, OCS, OBS and OPS. Chapter 7 focuses on switching domains in optical layer and describes the corresponding component technologies for space, time and wavelength switching. Then, the OCS, OBS and OPS nodes are described in Chapter 8. Furthermore, our contributions on novel node architectures for OCS and OPS are reviewed referring to the work published in Papers VIII and IX of this thesis. Finally, Chapter 9 concentrates on solving RWA problems in OCS networks. Our work related to the RWA presented in Papers X and XI of this thesis is also included.

In Chapter 10, the methodology for our simulation work is provided. First, two simulators based on discrete event driven systems are described. One of them is for evaluation of scheduling algorithms in PON. It is related to the work in Part I of this thesis. The second one is developed for the study in Part II of this thesis. Then, for the study of security issue in transparent optical networks our implementation for tabu search heuristic proposed in Paper XI of this thesis is also presented.

Finally, the conclusion of the research work included in this thesis is presented along with the suggestions for the future research. A brief summary of each paper of this thesis and the author’s contributions are also provided.

(20)

6 Chapter 1. Introduction

References

[1] Y. K. M. Lin, D. R. Spears, and M. Yin, “Fiber-based Local Access Network Architectures”, IEEE Communications Magazine, vol 27, pp.64 -73, Oct. 1989.

[2] L. A. Ims, B. T. Olsen, D. Myhre, M. Lahteenoja, J. Mononen, U. Ferrero, and A.

Zaganlaris, “Multiservice Access Network Upgrading in Europe: A Techno-economic Analysis”, IEEE Communications Magazine, vol. 34, pp.124 -134, Dec. 1996.

[3] I. Yamashita, “The Latest FTTH Technologies for Full Service Access Networks”, Circuits and Systems IEEE Asia Pacific Conference on, 1996

[4] R. Luo, T. Ning, L. Cai, F. Qiu, S. Jian, and J. Xu, “FTTH - A Promising Broadband Technology” Communications, Circuits and Systems International Conference on, 2005.

[5] C. Lin, J. Chen, P. Peng, C. Peng, W. Peng, B. Chiou and S. Chi, “Hybrid Optical Access Network Integrating Fiber-to-the-Home and Radio-Over-Fiber Systems”, IEEE Photonics Technology Letters, vol. 19, pp. 610-612, Apr. 2007.

[6] P. W. Shumate, “Fiber-to-the-Home: 1977–2007”, IEEE/OSA J. of Lightwave Technology, vol. 26, pp.1093-1103, May, 2008.

[7] IEEE 802.3ah task force home page [Online]. Available: http://www.ieee802.org/3/efm.

[8] ITU-T G.984.x series of recommendations [Online]. Available: http://www.itu.int/rec/T- REC-G/e.

[9] G. Kramer, and G. Pesavento, “Ethernet Passive Optical Network (EPON): Building a Next-generation Optical Access Network”, IEEE Communications Magazine, vol. 40, pp.

66 - 73, Feb. 2002

[10] B. Lung, “PON Architecture Future-proofs FTTH”, Ligthtwave, vol. 16, pp.104-107, Sep.

1999.

[11] J. Kani, M. Teshima, K. Akimoto, N. Takachio, H. Suzuki, K. Iwatsuki, and M. Ishii, “A WDM-based Optical Access Network for Wide-area Gigabit Access Services,” IEEE Communications Magazine, vol. 41, Feb. 2003, pp. S43–S48.

[12] L. Thylen, G. Karlsson, and O. Nilsson, “Switching Technologies for Future Guided Wave Optical Networks: Potentials and Limitations of Photonics and Electronics”, IEEE Communications Magazine, vol. 34, pp. 106-113, Feb.1996.

[13] L. Thylen, “Some Aspects of Photonics and Electronics in Communications and Interconnects”, Transparent Optical Networks, 1st International Conference on, ICTON '99, 1999.

[14] C. Raffaelli, L. Wosinska, N. Andriolli, F. Callegati, P. Castoldi, W. Kabacinski, G.

Maier, A. Pattavina, and L. Valcarenghi, “Photonics in Switching in NoE e-Photon/One+”, Transparent Optical Networks, 9th International Conference on, ICTON '07, 2007.

[15] A. Huang, L. Xie, Z. Li, and A. Xu, “Time-Space Lable Switching Protocol (TSL-SP) – A New Paradigm of Network Resource Assignment”, Photonic Network Communications, vol. 6, pp. 169-178, Sep. 2003.

[16] L. Wosinska, J. Haralson, L. Thylén, J. Öberg, and B. Hessmo, “Benefit of Implementing Novel Optical Buffers in an Asynchronous Photonic Packet Switch”, European Conference on Optical Communication, ECOC’04, 2004.

(21)

References 7 [17] L. Wosinska, and J. Chen, “Contention Resolution in An Asynchronous All-optical

Packet Switch”, international conference on photonics in switching, 2006.

[18] Chapter 2: Inside Simulation Software, S. Robinson, “Simulation - The practice of model development and use”, Wiley, 2004.

(22)

(23)

9

PART I

Fiber Access Networks

(24)

10 Chapter 2. Fiber Access Network Architectures

Chapter 2 Fiber Access Network Architectures

Fiber to the home (FTTH) is currently experiencing double-digit growth (or higher) [1-3] in the United States, Europe, and several Asian countries, because residential customers require high bit rate connections for broadband services. This demand for bandwidth has exceeded recent predictions, driven mostly by a number of factors, including the huge success of Internet video streaming services such as YouTube, the unanticipated success of HDTV (high-definition television), and the growing popularity of online social media sites where people meet, collaborate and more importantly exchange photographs, video, and audio content with each other. The number of users demanding high bandwidth continues to increase at a rapid pace. Consequently many service providers are planning networks capable of offering 50 Mb/s, 100 Mb/s, or higher bandwidth, per customer. In contrast to many existing broadband technologies, such as DSL (digital subscriber loop) and wireless access, fiber access can easily fulfill such bandwidth requirements, on a per customer basis, while still being capable of offering the higher capability in the future. This “future proof”

characteristic of FTTH has been widely recognized since the concept was first promoted over 30 years ago.

A most straightforward way to deploy fiber access network is using a point-to-point (P2P) architecture. In P2P, a dedicated fiber is deployed to connect the central office (CO) to each end user (see Fig. 2.1). Although this is a simple architecture, in most cases it is not cost- efficient due to the fact that it requires significant outside plant fiber deployment as well as a dedicated transceiver at CO for each end user. Considering N end users at an average distance L km from the CO, P2P architecture requires 2N transceivers and a total amount of fiber equal to N x L km (it is assumed that a single fiber is used for bidirectional transmission).

Fig. 2.1 P2P fiber access network architecture

Active optical network (AON) is another common architecture for fiber access. In AON, an electrical switch is deployed as a remote node (RN) close to the end users and only one single fiber is needed for the connection between the CO and the active switch (see Fig. 2.2). Due to the fact that active equipment is used at the RN, this architecture can provide longer reach compared to P2P. In addition, the total amount of deployed fiber is also reduced since only one a single feeder fiber is used. However, N + 1 transceivers are needed for electrical switch at the RN, so the total number of transceivers in an AON needs to be increased to 2N + 2.

Furthermore, AON architecture requires electrical power at the RN. Supply and maintenance of electrical power is considered as one of the key operational costs in access networks.

(25)

Chapter 2. Fiber Access Network Architectures 11

Fig. 2.2 AON architecture

Therefore, it is beneficial to replace the active switch at RN with an inexpensive passive optical component in order to save the operational cost on electrical power consumption in the local loop. Fig. 2.3 shows the typical deployment scenario of a passive optical network (PON).

A PON is a point-to-multipoint optical network with no active devices in the outside plant.

The elements used in the optical distribution network part of the PON are passive optical components, such as optical fiber, splices, and splitters/combiners in TDM PON (see Fig. 2.3).

Obviously, compared with AON and P2P architectures, the total number of transceivers used in a PON can be reduced to N + 1. Furthermore, due to only a single shared feeder fiber connecting the CO to the end users, PON can not only have higher flexibility for resource allocation, but can also easily accommodate some cost-efficient protection schemes in order to increase reliability compared to a P2P scheme. Due to these reasons PON is considered as the most attractive solution for fiber access networks [4-6].

Fig. 2.3 PON architecture

Remainder of the Part I focus on PON technologies. Chapter 3 provides a brief review of PON topologies and different multiplexing techniques over PON (including TDM PON, WDM PON and hybrid WDM/TDM PON). Some specific features related to the resource (i.e., time slot and wavelength) allocation and reliability performance of PONs are discussed in Chapter 4 and 5.

(26)

(27)

13

Chapter 3 Passive Optical Networks

Passive optical network (PON) is one of the most promising solutions for broadband access.

A PON is a point-to-multipoint optical network with no active devices in outside plant. The optical line terminal (OLT) resides in the CO, connecting the optical access network to the metro backbone while the optical network units (ONUs) are located close to the end users and provide customer service interfaces to the end users, converting optical signals to electrical ones.

This chapter provides a brief review of PON topologies and technologies.

3.1 PON topologies

Figure 3.1 shows three basic PON topologies i.e. tree, bus, and ring.

OLT

ONU ONU ONU

ONU

(a)

OLT

ONU ONU

(b)

OLT

ONU ONU

(c)

Fig 3.1 PON topologies: (a) tree with single splitting point, (b) bus, and (c) ring.

(28)

14 Chapter 3. Passive Optical Networks Tree is the most commonly used topology in fiber access networks. Especially, tree with single splitting point (see Fig. 3.1 (a)) is prevalent for PON. It uses a single fiber from the OLT to a remote node (RN), which is an intermediate splitting point. From this splitting point, there is a separate fiber allocated to each ONU connected to the network. The main advantage of this topology is that the splitting is only performed at a single point; thus it is simple to adopt all ONUs to have a similar power budget which means they all transmit or receive approximately the same optical signal power and quality.

Bus topology (see Fig. 3.1 (b)), where each ONU is connected to a tap coupler that can extract a part of power sent by the OLT, can be considered as a special case of tree topology. Its two main advantages are: 1) use of minimal amount of optical fiber; 2) flexible deployments since new ONUs can be connected to the network by adding more taps. However, the problem is that the signals of the ONUs, which have to pass several tap couplers, are degraded and weak.

Thus, the number of ONUs that can be connected to the bus PON is limited. Furthermore, it’s not easy to apply the cost-efficient protection scheme to the bus topology.

In ring topology (see Fig. 3.1 (c)), there are two possible ways to reach the OLT from each ONU. Therefore, in case of a fiber cut it is still possible to establish and maintain the connection. However, ring topology has the same drawback as the bus in terms of the power budget. When the optical signal passes through several couplers, it becomes degraded and attenuated. Thus, the total number of ONUs that can be connected to the ring PON is also limited.

3.2 Resource sharing in PONs

Time-division multiplexing (TDM) is a technique for time slot based sharing of a communication channel in network. It allows several users to share the transmission resource by dividing the signal into different time slots. The users transmit signals in rapid succession and each one uses its own time slot. This type of resource sharing technology is widely used in radio, electrical and optical networks. In radio and electrical networks, the same frequency channel is shared, and in an optical network the same wavelength channel is divided for different users. Furthermore, spectrum division multiplexing is another common resource sharing technology. In fiber optic communications, wavelength spectrum can be divided into different channels. In the so called wavelength-division multiplexing (WDM) different users can transmit signals in parallel in time domain, and each one uses its own wavelength channel.

Based on these two resource sharing technologies in a fiber based network, there are three main types of PONs, namely, TDM PON, WDM PON and hybrid WDM/TDM PON.

TDM PON, such as Ethernet-PON (EPON) [1], Gigabit-PON (GPON) [2] offers low per- subscriber cost by sharing a single wavelength channel among multiple subscribers. Due to the trend towards higher bandwidth demand, increasing number of subscribers and advances in the WDM device technology, the WDM PON and hybrid WDM/TDM PON [3-11] are being considered as the next generation solutions for the broadband access. Therefore, TDM PON, WDM PON and hybrid WDM/TDM PON are widely studied by the research community.

(29)

3.2. Resource sharing in PONs 15

(a) (b) Fig. 3.2 Tree TDM PON or WPON (a) and WRPON (b)

3.2.1 TDM PON

Remote node in a TDM PON (Fig. 3.2 (a)) is a splitter/combiner. Downstream traffic is broadcasted from the OLT to all the connected ONUs while in the upstream an arbitration mechanism is required, so that only a single ONU is allowed to transmit data at a given instance of time in the shared upstream channel. Two major standards for TDM PON have emerged, i.e. EPON [7] and GPON [8]. The following two sub sections describe key features of these two competing TDM PON standards along with their next generation counterparts.

3.2.1.1 Ethernet PON vs. Gigabit PON

EPON [7] and GPON [8] are two major standards for TDM PON. The main differences between EPON and GPON are shown in Table 3.1 in [9].

In EPON, both downstream and upstream line rates are 1.25 Gbps, but due to the 8B/10B line encoding the bit rate for data transmission is 1 Gbps. On the other hand, in GPON several upstream and downstream rates up to 2.48832 Gbps are specified, since GPON standard is defined in the ITU-T G.984.x series of recommendations [8] which refer to the bit rates of the conventional TDM systems.

Guard time between two neighboring time slots is used for differentiating the transmission from various ONUs. In EPON, it is composed of laser on-off time, automatic gain control (AGC) and clock-and-data recovery (CDR). IEEE 802.3ah standard [7] has specified values (classes) for AGC and CDR (see Table 3.1). In GPON, guard time consists of laser on-off time, preamble and delimiter. In Table 3.1, it can be seen that GPON has obviously shorter guard time than EPON. However, it requires stricter physical layer constraints than EPON.

Multi-point control protocol (MPCP) is implemented at the medium access control (MAC) layer in EPON to perform the bandwidth allocation, auto-discovery process and ranging. Two control messages REPORT and GATE used for bandwidth allocation are defined in [7].

Normally, a GATE message carries the granted bandwidth information from the OLT to an ONU in the downstream direction, while a REPORT message is used by an ONU to report the bandwidth request to the OLT in the upstream direction. This message exchange allows the time slots to be assigned according to the traffic demand of each individual ONU depending upon the available bandwidth. The size of REPORT and GATE message is 64 bytes which is equal to the shortest Ethernet frame. Furthermore, the EPON standard does not support frame fragmentation. Both OLT and ONUs can directly send and receive Ethernet frames with variable length. Table 3.1 shows the standard values for the frame size and overhead for bandwidth allocation in EPON.

(30)

16 Chapter 3. Passive Optical Networks TABLE 3.1

ACOMPARISON OF CURRENT EPON[7] AND GPON[8] STANDARDS

EPON GPON

Line rate

Downstream 1.25 Gbps Downstream 1.24416/

2.48832 Gbps

Upstream 1.25 Gbps Upstream

155.520Mbps/

622.08Mbps/

1.24416Gbps/

2.48832 Gbps Bit rate before

8B/10B line coding

1 Gbps Bit rate before scrambling line coding

155.520Mbps/

622.08Mbps/

1.24416Gbps/

2.48832 Gbps

Guard time

Laser on-off time 512 ns Laser on-off time ≈25.7 ns Automatic Gain

Control (AGC)

96 ns, 192 ns, 288 ns and 400 ns

Preamble & Delimiter 70.7 ns Clock-and-Data

Recovery (CDR)

96 ns, 192 ns, 288 ns and 400 ns

Frame size Ethernet frame 64 -1518 bytes

General Encapsulation

Method (GEM)

GEM

header 5 bytes Frame

fragment ≤1518 bytes Overhead for

bandwidth allocation

GATE/REPORT

64 bytes (Smallest size of Ethernet

frame)

Status report message 2 bytes

In the case of GPON protocol is based on the standard 125 µs periodicity used in the telecommunication industry. This periodicity provides certain efficiency advantages compared with EPON. Messages, such as control, buffer report and grant messages, can be efficiently integrated into the header of each 125 µs frame. In order to pack Ethernet frames into the 125 µs frame, Ethernet frame fragmentation has been introduced. Within GPON each Ethernet frame or frame fragment is up to 1518 bytes and encapsulated in a general encapsulation method (GEM) frame including a 5 byte GEM header. Status report message as the overhead for the bandwidth allocation is only 2 bytes. In addition, upstream QoS awareness has been integrated in the GPON standard with the introduction of the concept of transport containers (T-CONTs), where a type of T-CONT represents a class of service.

Hence GPON can provide a simple and efficient means for setting up a system for multiple service classes.

3.2.1.2 Next generation TDM PON

Both the current GPON and EPON standards are on the verge of evolving to their respective next generation standards providing 10 Gbps bit rate before line coding in downstream direction, along with the higher upstream bandwidth support than it is possible in the current generation TDM PON.

Current EPON based solutions have obtained great market penetration and have been widely

(31)

3.2. Resource sharing in PONs 17 deployed particularly in the Asian market. In order to cater for the ever increasing bandwidth

demand, the 10G EPON Task Force was formed in 2006 known as IEEE 802.3av [10] with an initiative to standardize requirements for the next generation EPON. IEEE 802.3av draft focuses on a new physical layer standard while keeping the changes of the logical layer at a minimum, i.e. maintaining the entire MPCP and operations, administration and maintenance (OAM) specifications as the IEEE 802.3ah standard. 10G EPON will use 64B/66B line coding with a line rate of 10.3125 Gbps instead of 8B/10B line coding with a line rate of 1.25 Gbps used in 1G EPON.

The most likely next generation GPON standard will have a 9.95328 Gbps downstream line rate and a 2.48832 Gbps upstream line rate. This upstream line rate has already been defined in the ITU-T recommendations. For larger upstream line rates, it may also possibly approach 9.95328 Gbps. Similar to the next generation EPON, the next generation GPON is also expected to keep the changes as few as possible in the logical layer.

3.2.2 WDM PON

WDM has been considered as an ideal solution to extend the capacity of optical networks without drastically changing the fiber infrastructure. Many architectures that incorporate WDM into access networks considered as the next generation solutions for the broadband access have been proposed from both academia and industry [11-14]. The fast development pace of WDM technologies opens up for wide deployment of WDM PON for the ultra-high bandwidth access solution.

Primarily, there are two types of WDM PON architectures, namely, WDM broadcast and select PON (WPON), and wavelength routed PON (WRPON). The outside plant of WPON can be identical to the standard TDM PON shown in Fig. 3.2 (a), with power splitter/combiner at the RN and the WDM equipment located at the OLT and ONUs. Fixed or tunable WDM filters at the ONUs select their own wavelength that is allocated either statically or dynamically. At the same time, WDM interfaces at the OLT transmit on different wavelengths. For the WRPON (the typical deployment case is shown in Fig. 3.2 (b)), the wavelength MUX/DEMUX (e.g., array waveguide grating AWG) at RN is replacing the power splitter used in WPON. Compared with WPON, the main advantages of WRPON are the improved power budget and the full duplex transmission. However, each ONU in WRPON may need a different laser source, increasing enormously the complexity of the stock control and decreasing the flexibility of wavelength allocation (e.g., dynamic wavelength allocation cannot be easily supported).

(32)

18 Chapter 3. Passive Optical Networks

Fig. 3.3 Possible approach for migration from TDM PON to hybrid WDM/TDM PON in Paper IV

3.2.3 Hybrid WDM/TDM PON

The main motivation for the hybrid WDM/TDM PON is: 1) a smooth migration from the current PON (i.e. TDM PON) to the future PON by incorporating the WDM technology; 2) supporting ultra-large number of ONUs.

Upgrading the access networks based on TDM PON with tree topology can be a challenge when user demand eventually outgrows the existing network capacity. Installing new fibers in the field is a most straightforward way to expand the PON coverage, but it is a very expensive approach. WDM has been considered as an ideal solution to extend the capacity of optical networks without drastically changing the fiber infrastructure. Hybrid WDM/TDM PON integrating WDM and TDM technologies could be a possible option to ensure the flexibility and a smooth transition from current PON to the future PON. Fig. 3.3 shows a possible approach for the smooth and graceful migration from TDM PON to the hybrid WDM/TDM PON in Paper IV of this thesis. The hybrid PON can be seen as either the intermediate stage or final phase in the migration towards the next generation PON.

In TDM PON and WPON, the total number of ONUs is limited by the splitting ratio of the power splitter. The power budget for the connection between OLT and ONUs can be a problem if splitting ratio is too high. In WRPON, the number of available wavelengths determinates the number of users supported. According to the state of the art, the number of ONUs in both TDM PON and WDM PON is strictly limited to a few of tens. On the other hand, hybrid WDM/TDM PON can support a large number of users. There are mainly two deployment alternatives for hybrid WDM/TDM PON, namely, embedded TDM PONs in a WDM PON [15-17] as shown in Fig. 3.3 and combined TDM PONs and WDM PONs in a ring [18-19], e.g. SUCCESS PON proposed in [18] shown in Fig. 3.4. In the first one, the upper bound on the number of ONUs can be a product of the upper bound on the number of ONUs in TDM PON and WDM PON while in the second one, it can be equal to the sum of the number of ONUs in each TDM PON and WDM PON hosted in one single hybrid PON.

(33)

3.2. Resource sharing in PONs 19

OLT

RN

RN RN

RN ONU

ONU

ONU ONU ONU ONU ONU

ONU ONU

ONU

ONU ONU ONU

ONU RN

ONU

RN ONU WDM PON

TDM PON

Fig. 3.4 SUCCESS PON proposed in [18]

(34)

Design, analysis and simultion for optical access and wide-area networks.

Design, Analysis and Simulation of Optical Access and Wide-area Networks

JIAJIA CHEN

©

Abstract

Acknowledgements

Contents

List of Papers

Acronyms

Chapter 1

Introduction

1.1 High capacity networks

1.1.1 Fiber access networks

1.1.2 Switched optical networks

1.1.2.1 Optical circuit switching

1.1.2.2 Optical burst switching

1.1.2.3 Optical packet switching

1.2 Discrete event driven simulation

1.3 Organization of the thesis

References

PART I

Fiber Access Networks

Chapter 2

Fiber Access Network Architectures

Chapter 3

Passive Optical Networks

3.1 PON topologies

3.2 Resource sharing in PONs

3.2.1 TDM PON

3.2.1.1 Ethernet PON vs. Gigabit PON

3.2.1.2 Next generation TDM PON

3.2.2 WDM PON

3.2.3 Hybrid WDM/TDM PON