Excerpt
TABLE OF CONTENTS
ACKNOWLEDGEMENTS
ABSTRACT
TABLE OF CONTENTS
LIST OF SYMBOLS
LIST OF ABBREVIATIONS
LIST OF FIGURES
LIST OF TABLES
LIST OF PUBLICATIONS
CHAPTER 1 INTRODUCTION
1.1 History
1.2 Objective of Thesis
1.3 Organization of Thesis
CHAPTER 2 BACKGROUND AND LITERATURE REVIEW
2.1 Introduction
2.1.1 Traffic Growth
2.1.2 Evolution of the “First Mile”
2.1.3 Next-Generation Access Network
2.2 Overview of PON Technologies
2.2.1 Optical Splitters/Combiners
2.2.2 PON Topologies
2.2.3 WDM vs. TDM PONs
2.2.4 Burst-Mode Transceivers
2.3 Ethernet PON (EPON) Access Network
2.3.1 Why Ethernet?
2.3.2 Principle of Operation Ethernet PON
2.4 BANDWIDTH ALLOCATION
2.4.1 DBA METHODS
2.4.1.1 DBA Algorithms without QoS Support
2.4.1.2 DBA Algorithms with QoS Support
2.4.1.2.1 DBA Algorithms with QoS Support locally
2.4.1.2.2 DBA algorithms with QoS support globally (universally)
CHAPTER 3 BASIC MODEL AND ANALYSIS
3.1 Introduction
3.2 Time Delay
3.3 Throughputs
3.4 Basic Model
3.5 Mathematical model
3.5.1 Arishtat- Limited Service IPACT- FSD-SLA
3.5.2 CDBA-IPACT
3.5.2.1 CDBA Principle
3.5.3 EBDBA-YDBA-SDBA-ADBA
3.5.4 CPBA- IPACT with two stages- IPACT with CBR credit
3.5.5 e-DBA n-DBA
3.5.6 TLBA-Two Stage Queue
3.5.7 IFLDBA- BP
3.5.8 UDBA-MSARF-CPBA SLA
CHAPTER 4 RESULTS AND DISCUSSIONS
4.1 Introduction
4.2 Simulation
4.3 Throughput Performance
4.4 Time delay performance
4.5 Selection of optimum DBA algorithms
4.6 Optimization Process
CHAPTER 5 CONCLUSION AND FUTURE WORK
5.1 Introduction
5.2 Conclusion
5.3 Future Work
ACKNOWLEDGEMENTS
I am very grateful to ALLAH Subhanahu WaTa'ala who gave me an excellent family to live with and provided me the environment where I could finish my M.Sc. and without whose will it would have been impossible to complete my degree.
Second only to ALLAH, thanks to my parents Mr. A.Maher Awad and Mrs. Nadia for what they have contributed towards my education. Their unconditional love, continuous support, great advice and prayers over the years are something that I cannot thank them for enough. Additionally, knowing that my success in completing my M.Sc. would be a source of happiness for them, I was further motivated to work hard.
I must also thank my supervisors Prof. Dr. Moustafa Hussein and Dr. Nazmi Azzam for their continuous guidance and support.
My deepest appreciation goes to my advisor Prof. Dr. Moustafa Hussein for his active contribution to refining my research work and filling in the gaps. At times, when things looked difficult, he was the one who gave me hope. Whether it was studying existing research or writing articles, he was there to listen to my concerns, review the material, provide feedback and show direction.
Similarly, I thank Dr. Nazmi Azzam for his support in improving my understanding of the all-optical network background and his suggestions during the research.
I would also like to thank my wife Mrs. Nadeen for her great support to me, and thank Eng. Ahmed Yossry from AAST Alexandria, Egypt, and Eng. Mohamed Tarek Hawel for fruitful discussions.
Finally, I gratefully acknowledge the staff of the Electronics and Communications Engineering Department for providing me with support and convenience during my studies at the Arab Academy for Science, Technology and Maritime Transport.
ABSTRACT
Dynamic bandwidth allocation in Ethernet passive optical networks (EPON) presents a key issue for providing efficient and fair utilization of the EPON upstream bandwidth while supporting the quality of service QoS requirements of different traffic classes. Rare literatures have addressed a qualitative and quantitative comparison of large numbers of DBA algorithms based on their performance indicators.
This thesis provides a detailed comparison and a classification study for a large number of DBA algorithms with respect to time delay and throughput as performance indicators. The study shows that IPACT WITH CBR, UDBA, IPACT with two stages and CPBA are the optimum DBA algorithms regarding both time delay and throughput at highly loaded scenarios.
These algorithms are enrolled in a parametric optimization process targeting performance enhancement at highly loaded scenarios this increasing upstream line rates, changing distance between the OLT (Optical Line Terminal) and ONU (Optical Network Unit), increasing size of an Ethernet packet and changing maximum cycle time to 1 ms and altering guard time value).
This process reduces time delay around 3.5% for IPACT WITH CBR, 1.725% for UDBA, 1.167% for IPACT with two stages and (1.167% for CPBA. Also, the optimization increases the throughput by 1.3% for IPACT WITH CBR, 1.795% in UDBA, 2.5% for IPACT with two stages and 1.684% for CPBA.
LIST of SYMBOLS
Abbildung in dieser Leseprobe nicht enthalten
LIST of ABBREVIATIONS
Abbildung in dieser Leseprobe nicht enthalten
LIST OF FIGURES
2.1 Fiber to the home (FTTH) deployment scenarios
2.2 8x8 couplers created from multiple 2x2 coupler
2.3 PON topologies
2.4 PON using a single fiber
2.5 Illustration of near-far problem in ATDM PON
2.6 Downstream traffic in EPON
2.7 Upstream traffic in EPON
2.8 DBA taxonomy
3.1 Schematic diagram for the structure of an EPON transmission operation
3.2 CDBA timing diagram as seen by the OLT
3.3 SLA-based DBA scheduling frames in EPONs
3.4 The principle of bandwidth allocation for TLBA
3.5 UDBA scheduling
4.1 Delay as a function of offered load for (Limited Service IPACT, FSD-SLA and Arishtat)
(a) From reference
(b) Obtained simulation result
4.2 Throughput as a function of offered load for (Limited Service IPACT, FSD-SLA and Arishtat)
(a) From reference
(b) Obtained simulation result
4.3 Delay as a function of offered load for (IPACT, CDBA 31 dB and CDBA 26 dB) (a) From reference
(b) Obtained simulation result
4.4 Throughput as a function of offered load for (IPACT, CDBA 31 dB and CDBA 26 dB)
(a) From reference
(b) Obtained simulation result
4.5 Delay as a function of offered load for (EBDBA, YDBA, SDBA and ADBA)
(a) From reference
(b) Obtained simulation result
4.6 Throughput as a function of offered load.
(a) From reference
(b) Obtained simulation result
4.7 Delay as a function of offered load (IPACT with two-stage queue, IPACT with CBR credit and CPBA)
(a) From reference
(b) Obtained simulation result
4.8 Throughput as a function of offered load for (IPACT with two- stage queue, IPACT with CBR credit and CPBA)
(a) From reference
(b) Obtained simulation result
4.9 Throughput as a function of offered load for (e-DBA and n-DBA)
(a) From reference
(b) Obtained simulation result
4.10 Delay as a function of offered load for (e-DBA and n-DBA)
(a) From reference
(b) Obtained simulation result
4.11 Throughput as a function of offered load for (TLBA and TWO STAGE QUEUE)
(a) From reference
(b) Obtained simulation result
4.12 Delay as a function of offered load for (TLBA and TWO STAGE)
(a) From reference.
(b) Obtained simulation result.
4.13 Delay as a function of offered load for (TLBA)
(a) From reference.
(b) Obtained simulation result.
4.14 Throughput as a function of offered load for (TCP-DBA-APC)
(a) From reference.
(b) Obtained simulation result.
4.15 Delay as a function of offered load for (TCP-DBA-APC)
(a) From reference
(b) Obtained simulation result.
4.16 Throughput as a function of offered load for (IFLDBA)
(a) From reference.
(b) Obtained simulation result.
4.17 Delay as a function of offered load for (IFLDBA)
(a) From reference.
(b) Obtained simulation result.
4.18 Delay as a function of offered load for (UDBA)
(a) From reference.
(b) Obtained simulation result.
4.19 Throughput as a function of offered load for (UDBA)
(a) Comparison between different DBA algorithms (throughput)
4.20 (b) Comparison between different DBA algorithms (throughput)
4.21 Comparison between different DBA algorithms (throughput).
4.22 (a) Comparison between different DBA algorithms (Delay).
(b) Comparison between different DBA algorithms (Delay) zoom in range 0.56-0.64 Gbps.
(c) Comparison between different DBA algorithms (Delay) zoom in range 0.65-0.95 Gbps.
4.23 (a) Comparison between different DBA algorithms (Delay)
(b) Comparison between different DBA algorithms (Delay) zoom in range 0.34-0.44 Gbps.
(c) Comparison between different DBA algorithms (Delay) zoom in c) range 0.5-0.6 Gbps. d)
(d) Comparison between different DBA algorithms (Delay) zoom in range 0.75-0.95 Gbps
4.24 Throughputs of different DBA algorithms with original values of simulations parameters
4.25 Throughput of different DBA algorithms with optimized values of simulations parameters
4.26 Time delay of different DBA algorithms with original values of simulations parameters
4.27 Time delay of different DBA algorithms with optimized values of simulations parameters.
List of Tables
2.1 Summary of DBA schemes
3.1 The simulation parameters
4.1 The simulation parameters
4.2 Obtained time delay at different offered loads for (Limited Service IPACT, FSD-SLA and Arishtat)
4.3 Obtained throughput at different offered load load for (Limited Service IPACT, FSD-SLA and Arishtat)
4.4 Obtained time delay at different offered loads for (IPACT, CDBA 31 dB and CDBA 26 dB)
4.5 Obtained throughput at different offered load for (IPACT, CDBA 31 dB and CDBA 26 dB)
4.6 Obtained time delay at different offered loads for (EBDBA, YDBA, SDBA and ADBA)
4.7 Obtained throughput at different offered load for (EBDBA, YDBA,SDBA and ADBA)
4.8 Obtained time delay at different offered loads for (IPACT with two-stage queue, IPACT with CBR credit and CPBA)
4.9 Obtained throughput at different offered load for (IPACT with two-stage queue, IPACT with CBR credit and CPBA)
4.10 Obtained throughput at different offered load for (e-DBA and n-DBA)
4.11 Obtained time delay at different offered loads for (e-DBA and n-DBA)
4.12 Obtained throughput at different offered load for (TLBA and TWO STAGE QUEUE)
4.13 Obtained time delay at different offered loads for (TLBA and TWO STAGE)
4.14 Obtained throughput at different offered load (TLBA)
4.15 Obtained throughput at different offered load (TCP-DBA-APC)
4.16 Obtained time delay at different offered loads for (TCP-DBA-APC)
4.17 Obtained throughput at different offered load for (IFLDBA)
4.18 Obtained time delay at different offered loads for (IFLDBA)
4.19 Obtained time delay at different offered loads for (UDBA)
4.20 Obtained throughput at different offered load for (UDBA)
4.21 DBA algorithms according to their throughput performance.
4.22 DBA algorithms in ascending order according to time delay at offered load = 0.9 Gbps
4.23 DBA algorithms in descending order according to throughput performance
4.24 Original parameters and its optimized values for different algorithms.
LIST OF PUBLICATIONS
1 NAZMI A. MOHAMED, MOHAMED A. MAHER, MOUSTAFA H. ALY "EPON performance optimization: an extensive comparative study for DBA algorithms ," OPTOELECTRONICS AND ADVANCED MATERIALS - RAPID COMMUNICATIONS Journal, Vol. 10, No. 7-8, p. 503 - 508, July-August 2016.
CHAPTER 1 INTRODUCTION
1.1 History
Two major standard groups, the Institute of Electrical and Electronics Engineers (IEEE) and the Telecommunication Standardization Sector of the International Telecommunication Union (ITU-T), develop standards along with a number of other industry organizations.
The Society of Cable Telecommunications Engineers (SCTE) also specified radio frequency over glass for carrying signals over a passive optical network.
Starting in 1995, work on fiber to the home architectures was done by the Full Service Access Network (FSAN) working group, formed by major telecommunications service providers and system vendors. The International Telecommunications Union (ITU) did further work, and standardized on two generations of PON. The older ITU-T G.983 standard was based on Asynchronous Transfer Mode (ATM), and has therefore been referred to as APON (ATM PON). Further improvements to the original APON standard - as well as the gradual falling out of favor of ATM as a protocol - led to the full, final version of ITU-T G.983 being referred to more often as broadband PON, or BPON. A typical APON/BPON provides 622 megabits per second (Mbit/s) (OC-12) of downstream bandwidth and 155 Mbit/s (OC-3) of upstream traffic, although the standard accommodates higher rates.
The ITU-T G.984 Gigabit-capable Passive Optical Networks (GPON) standard represented an increase, compared to BPON, in both the total bandwidth and bandwidth efficiency through the use of larger, variable-length packets. Again, the standards permit several choices of bit rate, but the industry has converged on 2.488 gigabits per second (Gbit/s) of downstream bandwidth, and 1.244 Gbit/s of upstream bandwidth. GPON Encapsulation Method (GEM) allows very efficient packaging of user traffic with frame segmentation.
By mid-2008, Verizon had installed over 800,000 lines.
British Telecom, BSNL, Saudi Telecom Company, Etisalat, and AT&T were in advanced trials in Britain, India, Saudi Arabia, the UAE, and the USA, respectively. GPON networks have now been deployed in numerous networks across the globe, and the trends indicate higher growth in GPON than other PON technologies.
G.987 defined 10G-PON with 10 Gbit/s downstream and 2.5 Gbit/s upstream - framing is "G- PON like" and designed to coexist with GPON devices on the same network Developed in 2009 by Cable Manufacturing Business to meet SIPRNet requirements of the US Air Force, secure passive optical network (SPON) integrates gigabit passive optical network (GPON) technology and protective distribution system (PDS).
Changes to the NSTISSI 7003 requirements for PDS and the mandate by the US federal government for GREEN technologies allowed for the US federal government consideration of the two technologies as an alternative to active Ethernet and encryption devices.
The chief information officer of the United States Department of the Army issued a directive to adopt the technology by fiscal year 2013. It is marketed to the US military by companies such as Telos Corporation.
In 2004, the Ethernet PON (EPON or GEPON) standard 802.3ah-2004 was ratified as part of the Ethernet in the first mile project of the IEEE 802.3. EPON uses standard 802.3 Ethernet frames with symmetric 1 gigabit per second upstream and downstream rates. EPON is applicable for data-centric networks, as well as full-service voice, data and video networks. 10 Gbit/s EPON or 10G-EPON was ratified as an amendment IEEE 802.3av to IEEE 802.3. 10G-EPON supports 10/1 Gbit/s. The downstream wavelength plan support simultaneous operation of 10 Gbit/s on one wavelength and 1 Gbit/s on a separate wavelength for operation of IEEE 802.3av and IEEE 802.3ah on the same PON concurrently. The upstream channel can support simultaneous operation of IEEE 802.3av and 1 Gbit/s 802.3ah simultaneously on a single shared (1310 nm) channel.
There are currently over 40 million installed EPON ports making it the most widely deployed PON technology globally. EPON is also the foundation for cable operators' business services as part of the DOCSIS Provisioning of EPON (DPoE) specifications.
The OLT is responsible for allocating upstream bandwidth to the ONUs. Because the optical distribution network (ODN) is shared, ONU upstream transmissions could collide if they were transmitted at random times. ONUs can lie at varying distances from the OLT, meaning that the transmission delay from each ONU is unique. The OLT measures delay and sets a register in each ONU via PLOAM (physical layer operations and maintenance) messages to equalize its delay with respect to the other entire ONUs on the PON.
Once the delay of all ONUs has been set, the OLT transmits so-called grants to the individual ONUs. A grant is permission to use a defined interval of time for upstream transmission. The grant map is dynamically re-calculated every few milliseconds. The map allocates bandwidth to all ONUs, such that each ONU receives timely bandwidth for its service needs.
Some services - POTS, for example - require essentially constant upstream bandwidth, and the OLT may provide a fixed bandwidth allocation to each such service that has been provisioned. DS1 and some classes of data service may also require constant upstream bit rate. But much data traffic, such as browsing web sites, is bursty and highly variable. Through dynamic bandwidth allocation (DBA), a PON can be oversubscribed for upstream traffic, according to the traffic engineering concepts of statistical multiplexing. (Downstream traffic can also be oversubscribed, in the same way that any LAN can be oversubscribed. The only special feature in the PON architecture for downstream oversubscription is the fact that the ONU must be able to accept completely arbitrary downstream time slots, both in time and in size.)
In GPON there are two forms of DBA, status-reporting (SR) and non-status reporting (NSR). In NSR DBA, the OLT continuously allocates a small amount of extra bandwidth to each ONU. If the ONU has no traffic to send, it transmits idle frames during its excess allocation. If the OLT observes that a given ONU is not sending idle frames, it increases the bandwidth allocation to that ONU. Once the ONU's burst has been transferred, the OLT observes a large number of idle frames from the given ONU, and reduces its allocation accordingly. NSR DBA has the advantage that it imposes no requirements on the ONU, and the disadvantage that there is no way for the OLT to know how best to assign bandwidth across several ONUs that need more.
In SR DBA, the OLT polls ONUs for their backlogs. A given ONU may have several so- called transmission containers (T-CONTs), each with its own priority or traffic class. The ONU reports each T-CONT separately to the OLT. The report message contains a logarithmic measure of the backlog in the T-CONT queue. By knowledge of the service level agreement for each T- CONT across the entire PON, as well as the size of each T-CONT's backlog, the OLT can optimize allocation of the spare bandwidth on the PON.
EPON systems use a DBA mechanism equivalent to GPON's SR DBA solution. The OLT polls ONUs for their queue status and grants bandwidth using the MPCP GATE message, while ONUs report their status using the MPCP REPORT message
1.2 Objective of Thesis
This work provides a detailed comparison and classification study for a large number of DBA algorithms with respect to time delay and throughput as performance indicators. Seeking to find one DBA algorithm that achieves remarkable results in time delay and throughput performance.
These algorithms are enrolled in a parametric optimization process targeting performance enhancement at highly loaded scenarios (i.e. increasing upstream line rates, changing distance between the OLT and ONU from, increasing size of an Ethernet packet and change maximum cycle time to 1 ms and guard time).
1.3 Organization of Thesis
The organization of this thesis is as follows: Chapter 2 presents the background of the topics relevant to the subject of the thesis such as passive optical network and dynamic band width allocation; this chapter also includes the literature review with related works. In Ch. 3, the basic model and mathematical model equations are provided for 23 DBA algorithms. Chapter 4 starts with the throughput and time delay for the DBA algorithms under evaluation, followed by a selection of optimum four DBA algorithms and finally the obtained results of a parametric optimization process for algorithms is carried out to enhance the performance of these algorithms regarding time delay and throughput. A Summary, conclusions and suggestions for some recommended points for future work in this field are presented in Ch. 5. The work is terminated by a list of references and an Arabic summary.
CHAPTER 2 BACKGROUND AND LITERATURE REVIEW
2.1 Introduction
In recent years, the telecommunications backbone has experienced substantial growth; however, little has changed in the access network. The tremendous growth of Internet traffic has accentuated the aggravating lag of access network capacity.
The “last mile” still remains the bottleneck between high-capacity Local Area Networks (LANs) and the backbone network. The most widely deployed “broadband” solutions today are Digital Subscriber Line (DSL) and cable modem (CM) networks. Although they are improved compared to 56 kbps dial-up lines, they are unable to provide enough bandwidth for emerging services such as Video-On-Demand (VoD), interactive jamming or two-way video conferencing. A new technology is required; that is inexpensive, simple, scalable, and capable of delivering bundled voice, data and video services to an end-user over a single network.
Ethernet Passive Optical Networks (EPONs), which represent the convergence of low-cost Ethernet equipment and low-cost fiber infrastructure, appear to be the best candidate for the next-generation access network.
2.1.1 Traffic Growth
Data traffic is increasing at an unprecedented rate. Sustainable data traffic growth rate of over 100% per year has been observed since 1990. There were periods when a combination of economic and technological factors resulted in even larger growth rates, e.g., 1000% increase per year in 1995 and 1996 1.
This trend is likely to continue in the future. Simply, more and more users are getting online, and those who are already online are spending more time online and are using more bandwidth-intensive applications. Market research shows that, after upgrading to a broadband connection, users spend about 35% more time online than before 2. Voice traffic is also growing, but at a much slower rate of 8% annually.
According to most analysts, data traffic has already surpassed the voice traffic. More and more subscribers telecommunicate, and require the same network performance as they see on corporate LANs.
More services and new applications will be available as bandwidth per user increases. Neither DSL nor CM can keep up with such demand. Both technologies are built on top of existing communication infrastructure not optimized for data traffic. In CM networks, only a few RF channels are dedicated for data, while the majority of bandwidth is tied up servicing legacy analog video. DSL copper networks do not allow sufficient data rates at required distances due to signal distortion and crosstalk. Most network operators have come to the realization that a new, data-centric solution is necessary. Such a technology would be optimized for Internet Protocol (IP) data traffic. The remaining services, such as voice or video, will converge into a digital format and a true full-service network will emerge.
2.1.2 Evolution of the “First Mile”
The ‘first mile' once called the last mile, the networking community has renamed this network segment to the first mile, to symbolize its priority and importance. The first mile connects the service provider central offices to business and residential subscribers. Also, referred to as the subscriber access network, or the local loop, it is the network infrastructure at the neighborhood level.
Residential subscribers demand first-mile access solutions that are broadband, offer Internet media-rich services, and are comparable in price with existing networks. Incumbent telephone companies responded to Internet access demand by deploying DSL technology.
DSL uses the same twisted pair as telephone lines and requires a DSL modem at the customer premises and Digital Subscriber Line Access Multiplexer (DSLAM) in the central office (CO). The data rate provided by DSL is typically offered in a range from 128 kbps to 1.5 Mbps. While this is significantly faster than an analog modem, it is well recluse of being considered “broadband” in that it cannot support emerging voice, data, and video applications.
In addition, the physical area that one central office can cover with DSL is limited to distances less than 18000 ft (5.5 km), which covers approximately 60% of potential subscribers. Even though, to increase DSL coverage, remote DSLAMs (R- DSLAMs) may be deployed closer to subscribers, network operators, in general, do not provide DSL services to subscribers located more than a 12000 ft from CO due to increased costs 3.
Cable television companies responded to Internet service demand by integrating data services over their coaxial cable networks, which were originally designed for analog video broadcast. Typically, these hybrid fiber coaxial (HFC) networks have fiber running between a video head-end and a hub to a curb side optical node, with the final drop to the subscriber being coaxial cable, repeaters, and tap couplers.
The drawback of this architecture is that each shared optical node has less than 36 Mbps effective data throughput, which is typically divided between 2000 homes, resulting in frustrating slow speed during peak hours.
To alleviate bandwidth bottlenecks, optical fibers, and thus optical nodes, are penetrating deeper into the first mile. The next wave of local access deployment promises to bring fiber to the building (FTTB) and fiber to the home (FTTH). Unlike previous architectures, where fiber is used as a feeder to shorten the lengths of copper and coaxial networks, these new deployments use optical fiber throughout the access network.
New, optical fiber network architectures are emerging that are capable of supporting gigabit per second speeds, at costs comparable to those of DSL and HFC networks.
2.1.3 Next-Generation Access Network
The optical fiber is capable of delivering bandwidth-intensive, integrated, voice, data and video services at distances beyond 20 km in the subscriber access network. A logical way to deploy optical fibers in the local access network is to use a point-to-point (PtP) topology, with dedicated fiber runs from the CO to each end-user subscriber (Fig. 2.1a). While this is a simple architecture, in most cases it is cost prohibitive due to the fact that it requires significant outside plant fiber deployment as well as connector termination space in the Local Exchange. Considering N subscribers at an average distance L km from the central office, a PtP design requires 2 N transceivers and N*L total fiber length (assuming that a single fiber is used for bi-directional transmission).
Abbildung in dieser Leseprobe nicht enthalten
Figure 2.1 Fiber to the home (FTTH) deployment scenarios 4.
To reduce fiber deployment, it is possible to deploy a remote switch (concentrator) close to the neighborhood. This will reduce the fiber consumption to only L km (assuming a negligible distance between the switch and customers), but will actually increase the number of transceivers to 2 N +2, as there is one more link added to the network (Fig 2.1b). In addition, curb-switched network architecture requires electrical power as well as back-up power at the curb switch. Currently, one of the highest costs for Local Exchange Carriers (LECs) is providing and maintaining electrical power in the local loop.
Therefore, it is logical to replace the hardened (environmentally protected) active curbside switch with an inexpensive passive optical splitter.
A Passive Optical Network (PON) is a technology viewed by many as an attractive solution to the first mile problem [4, 5]. A PON minimizes the number of optical transceivers, central office terminations and fiber deployment. A PON is a point- to-multipoint (PtMP) optical network with no active elements in the signals' path from source to destination. The only interior elements used in PON are passive optical components, such as optical fibers, splices and splitters. An access network based on a single-fiber PON only requires N + 1 transceivers and L km of fiber (Fig 2.1c).
2.2 Overview of PON Technologies
2.2.1 Optical Splitters/Combiners
A PON employs a passive (not requiring any power) device to split optical signal (power) from one fiber into several fibers and reciprocally, to combine optical signals from multiple fibers into one. This device is an optical coupler. In its simplest form, an optical coupler consists of two fibers fused together. Signal power received on any input port is split between both output ports. The splitting ratio of a splitter can be controlled by the length of the fused region and therefore is a constant parameter.
N xN couplers are manufactured by staggering multiple 2 x 2 couplers (Fig. 2.2) or by using planar waveguide technology.
Abbildung in dieser Leseprobe nicht enthalten
Figure 2.2 8x8 couplers created from multiple 2x2 couplers 6.
Couplers are characterized by the following parameters: splitting loss - power level at the coupler's output vs. power level at its input, measured in dB. For an ideal 2 x 2 coupler, this value is 3 dB. Figure 2.2 illustrates two topologies for 8 x 8 couplers based on 2 x 2 couplers. In a 4-stage topology (Fig 2.2.a), only 1/16 of the input power is delivered to each output.
Figure 2.2.b shows a more efficient design called multistage interconnection network 6. In this arrangement, each output receives 1/8 of the input power. Insertion loss - power loss resulting from imperfections of the manufacturing process.
Typically, this value ranges from 0.1 dB to 1 dB. Directivity in the amount of input power leaked from one input port to another input port. Couplers are highly directional devices with the directivity parameter reaching 40 - 50 dB. Very often, couplers are manufactured to have only one input or one output. A coupler with only one input is referred to as a splitter. A coupler with only one output is called a combiner. Sometimes, 2 x 2 couplers are made highly asymmetric (with splitting ratios 5/95 or 10/90). This kind of couplers is used to branch off a small portion of signal power, for example, for monitoring purposes. Such devices are called tap couplers.
2.2.2 PON Topologies
Logically, the first mile is a point to multi point (PTMP) network, with a CO servicing multiple subscribers. There are several multipoint topologies suitable for the access network, including tree, tree-and branch, ring, or bus (Fig. 2.3). Using 1:2 optical tap couplers and 1:N optical splitters, the PONs can be flexibly deployed in any of these topologies.
In addition, PONs can be deployed in redundant configurations such as double rings or double trees; or redundancy may be added to only a part of the PON, say the trunk of the tree (Fig. 2.3d) also refer to 7 for more redundant topologies.
Abbildung in dieser Leseprobe nicht enthalten
Figure 2.3 PON topologies 7.
All transmissions in a PON are performed between an Optical Line Terminal (OLT) and Optical Network Units (ONUs) (Fig 2.3). The OLT resides in the CO and connects the optical. The Access network to the metropolitan area network (MAN) or wide area network (WAN) is also known as backbone or long-haul network. The ONU is located either at the end-user location (FTTH and FTTB), or at the curb, resulting in fiber to the curb (FTTC) architecture. The advantages of using PONs in subscriber access networks are numerous 7:
- PONs allows long reach between the CO and customer premises, operating at distances over 20 km.
- PONs minimize fiber deployment in both the CO and the local loop.
- PONs provide higher bandwidth due to deeper fiber penetration, offering gigabit-per second solutions.
- Operating in the downstream as a broadcast network, PONs allow for video broadcasting either as IP video, or analog video.
- PONs eliminate the necessity of installing active multiplexers at the splitting locations, thus relieving network operators from the gruesome task of maintaining active curbside units and providing power to them. Instead of active devices in these locations, PONs use small passive optical splitters, located in splice trays, and deployed as part of the optical fiber cable plant.
- Being optically transparent end-to-end, PONs allow upgrades to higher bit rates or additional wavelengths.
2.2.3 WDM vs. TDM PONs
In the downstream direction (from OLT to ONUs), a PON is a point-to- multipoint network. The OLT typically has the entire downstream bandwidth available to it at all times. In the upstream direction, a PON is a multipoint-to-point network: multiple ONUs transmit all towards one OLT. Directional properties of a passive splitter/combiner are such that an ONU's transmission cannot be detected by other ONUs. However, data streams from different ONUs transmitted simultaneously still may collide. Thus, in the upstream direction (from user to network), the PON should employ some channel separation mechanism to avoid data collisions and fairly share the trunk fiber channel capacity and resources.
One possible way of separating the ONU upstream channels is to use a wavelength division multiplexing (WDM), in which each ONU operates on a different wavelength. While it is a simple solution (from a theoretical perspective), it remains cost-prohibitive for an access network. A WDM solution would require either a tunable receiver, or a receiver array at the OLT to receive multiple channels. An even more serious problem for network operators would be wavelength specific ONU inventory: instead of having just one type of ONU, there would be multiple types of ONUs based on their laser wavelength. Each ONU will have to use a laser with a narrow and controlled spectral width, and thus will become more expensive.
It would also be more problematic for an unqualified user to replace a defective ONU because a unit with wrong wavelength may interfere with some other ONU in the PON. Using tunable lasers in ONUs may solve the inventory problem, but is too expensive at the current state of technology. For these reasons, a WDM PON network is not an attractive solution in today's environment.
Several alternative solutions based on WDM have been proposed, like wavelength routed PON (WRPON). A WRPON uses an arrayed waveguide grating (AWG) instead of wavelength-independent optical splitter/combiner. We refer the reader to 8 for a detailed overview of these approaches.
In one variation, ONUs use external modulators to modulate the signal received from the OLT and send it back upstream. This solution, however, is not cheap because it either requires additional amplifiers at or close to the ONUs to compensate for signal attenuation after the round-trip propagation. It also requires more expensive optical components to limit the reflections, since both downstream and upstream channels use the same wavelength.
Also, to allow independent (non-arbitrated) transmission from each of N ONUs, the OLT must have N receivers - one for each ONU. In another variation, ONUs contain cheap light-emitting diodes LEDs whose wide spectral band was sliced by the AWG on the upstream path. This approach still requires multiple receivers at the OLT.
If, however, a single tunable receiver is used at the OLT, then a data stream from only one ONU can be received at a time, which in effect makes it a time-division multiplexed (TDM) PON.
In a TDM PON, simultaneous transmissions from several ONUs will collide when reaching the combiner. In order to prevent data collisions, each ONU must transmit in its own transmission window (time slot). One of the major advantages of a TDM PON is that all ONUs can operate on the same wavelength and be absolutely identical component-wise. The OLT will also need a single receiver. A transceiver in an ONU must operate at the full line rate, even though the bandwidth available to the ONU is lower. However, this property also allows the TDM PON to efficiently change the bandwidth allocated to each ONU by changing the assigned timeslot size, or even employ statistical multiplexing to fully utilize the bandwidth available in the PON.
In a subscriber access network, most of the traffic flows downstream (from network to users) and upstream (from users to the network), but not peer-to-peer (user to user). Thus, it seems reasonable to separate the downstream and the upstream channels. A simple channel separation can be based on space division multiplexing (SDM) where separate PONs are provided for downstream and for upstream transmissions. To save optical fiber and reduce cost of repair and maintenance, a single fiber may be used for bi-directional transmission. In this case, two wavelengths are used: typically 1310 nm (11) for the upstream transmission and 1550 nm (X2) for the downstream transmission (Fig 2.4). The channel capacity on each wavelength can be flexibly divided between the ONUs.
Abbildung in dieser Leseprobe nicht enthalten
Figure 2.4 PON using a single fiber 8.
Time-sharing appears to be the preferred method today for optical channel sharing in an access network as it allows a single upstream wavelength, such as 1310 nm, and a single transceiver in the OLT, resulting in a cost-effective solution.
2.2.4 Burst-Mode Transceivers
Due to unequal distances between CO and ONUs, optical signal attenuation in the PON is not the same for each ONU. The power level received at the OLT will be different for each timeslot (called the near-far problem). Figure 2.5 depicts power levels of four timeslots received by the OLT from four different ONUs in a TDM PON.
Abbildung in dieser Leseprobe nicht enthalten
Figure 2.5 Illustration of near-far problem in a TDM PON: a snapshot of received power level from four time slots 8.
As shown, one ONU signal strength is lower at the OLT most likely due to its longer distance. If the receiver in OLT is adjusted to properly receive high-power signal from a close ONU, it may mistakenly read ones as zeros when receiving weak signal from a distant ONU.
In the opposite case, if the receiver is trained on a weak signal, it may read zeros as ones when receiving a strong signal. To properly detect the incoming bit stream, the OLT receiver must be able to quickly adjust its zero-one threshold at the beginning of each received timeslot, i.e., it should operate in burst mode. A burst mode receiver is necessary only in the OLT. The ONUs read a continuous bit stream (data or idles) sent by the OLT and do not need to re-adjust quickly. An alternative approach is to allow ONUs to adjust their transmitter powers such that power levels received by OLT from all ONUs become the same. This method is not particularly favored by transceiver designers as it makes the ONU hardware more complicated. It will require special signaling protocol for feedback from the OLT to each ONU, and most importantly, may degrade the performance of all ONUs to that of a most distant unit.
Another issue is that it is not enough just to disallow ONUs to send any data. The problem is that, even in the absence of data, lasers generate spontaneous emission noise. Spontaneous emission noise from several ONUs located close to the OLT can easily obscure the signal from a distant ONU (capture effect). Thus, an ONU must shut down its laser between the time slots.
Because a laser cools down when it is turned off, and warms up when it is turned on, its emitted power may fluctuate at the beginning of a transmission. It is important that the laser be able to stabilize quickly after being turned on.
2.3 Ethernet PON (EPON) Access Network
Ethernet PON (EPON) is a PON-based network that carries data traffic encapsulated in Ethernet frames (defined in the IEEE 802.3 standard). It uses a standard 8b/10b line coding (8 user bits encoded as 10 line bits) and operates at standard Ethernet speed.
2.3.1 Why Ethernet?
Passive optical networking has been considered for the access network for quite some time, even well before the Internet spurred bandwidth demand. The Full Service Access Network (FSAN) recommendation (ITU G.983) defines a PON-based optical access network that uses ATM as its layer 2 protocol. In 1995, when the FSAN initiative was started, ATM had high hopes of becoming the prevalent technology in the LAN, MAN and backbone. However, since that time, Ethernet technology has leapfrogged ATM. Ethernet has become a universally accepted standard, with over 320 million port deployments worldwide, offering staggering economies of scale 9.
High-speed Gigabit Ethernet deployment is widely accelerating and 10 Gigabit Ethernet products are becoming available. Ethernet, which is easy to scale and manage, is winning new grounds in MAN and WAN. Considering the fact that 95% of LANs use Ethernet, it becomes clear that ATM PON may not be the best choice to interconnect two Ethernet networks. One of ATM shortcomings is the fact that a dropped or corrupted ATM cell will invalidate the entire IP datagram.
However, the remaining cells carrying the portions of the same IP datagram will propagate further, thus consuming network resources unnecessarily. Also, ATM imposes a cell tax on variable-length IP packets. For example, for the tri-modal packetsize distribution reported in 10, the cell tax is approximately 13%, i.e., to send the same amount of user's data an ATM network must transmit 13% more bytes than an Ethernet network (counting 64-bit preamble and 96-bit minimum inter-frame gap (IFG) in Ethernet and 12 bytes of overhead associated with ATM adaptation layer 5 (AAL- 5)). Finally, perhaps most importantly, ATM did not live up to its promise of becoming an inexpensive technology - vendors are in decline and manufacturing volumes are relatively low. ATM switches and network cards are significantly (roughly 8 times) more expensive than Ethernet switches and network cards 9.
On the other hand, Ethernet looks like a logical choice for an IP data-optimized access network. Newly-adopted quality-of-service (QoS) techniques have made Ethernet networks capable of supporting voice, data and video. These techniques include full duplex transmission mode, prioritization (P802.1p), and virtual LAN (VLAN) tagging (P802.1Q). Ethernet is an inexpensive technology, which is ubiquitous and interoperable with a variety of legacy equipment.
2.3.2 Principle of Operation Ethernet PON
The IEEE 802.3 standard defines two basic configurations for an Ethernet network. In one configuration, it can be deployed over a shared medium using the Carrier Sense Multiple Access with Collision Detection (CSMA/CD) protocol. In another configuration, stations may be connected through a switch using full-duplex point-to-point links. Properties of EPON are such that it cannot be considered either a shared medium or a point-to-point network; rather, it is a combination of both.
In the downstream direction, Ethernet frames transmitted by the OLT pass through a 1:N passive splitter and reach each ONU. The value of N is typically between 4 and 64. This behavior is similar to a shared-medium network. Because Ethernet is broadcast by nature, in the downstream direction, it fits perfectly with the Ethernet PON architecture: packets are broadcast by the OLT and extracted by their destination ONU based on the media-access control (MAC) address (Fig. 2.6).
Abbildung in dieser Leseprobe nicht enthalten
Figure 2.6 Downstream traffic in EPON 12.
In the upstream direction, due to the directional properties of a passive optical combiner, data frames from any ONU will only reach the OLT, and no other ONUs. In that sense, in the upstream direction, the behavior of EPON is similar to that of a point- to-point architecture. However, unlike in a true point-to-point network, in EPON data, frames from different ONUs transmitted simultaneously still may collide. Thus, in the upstream direction (from users to network) the ONUs need to employ some arbitration mechanism to avoid data collisions and fairly share the fiber-channel capacity.
A contention-based media access mechanism (something similar to CSMA/CD) is difficult to implement because ONUs cannot detect a collision at the OLT (due to directional properties of optical splitter/combiner). An OLT could detect a collision and inform ONUs by sending a jam signal. However, propagation delays in PON, which can exceed 20 km in length, can greatly reduce the efficiency of such a scheme. Contention-based schemes also have a drawback of providing a non-deterministic service, i.e., node throughput and channel utilization may be described as statistical averages.
There is no guarantee of a node getting access to the media in any small interval of time. It is not a problem for CSMA/CD-based enterprise networks where links are short, typically over-provisioned, and traffic predominantly consists of data. Subscriber access networks, however, in addition to data, must support voice and video services, and thus must provide some guarantees on time delivery of these traffic types. To introduce determinism in the frame delivery, different non-contention schemes have been proposed. Figure 2.7 illustrates an upstream time-shared data flow in an EPON.
Abbildung in dieser Leseprobe nicht enthalten
Figure 2.7 Upstream traffic in EPON 12.
All ONUs are synchronized to a common time reference and each ONU is allocated a time slot. Each time slot is capable of carrying several Ethernet frames. An ONU should buffer frames received from a subscriber until its time slot arrives. When its time slot arrives, the ONU would “burst” all stored frames at full channel speed which must correspond to one of standard Ethernet rates (10/100/1000/10000 Mbps). If there are no frames in the buffer to fill the entire time slot, idles 10-bit characters are transmitted.
The possible time slot allocation schemes could range from a static allocation (fixed time-division multiple access (TDMA)) to a dynamically adapting scheme based on instantaneous queue size in every ONU (statistical multiplexing scheme). There are more allocation schemes possible, including schemes utilizing notions of traffic priority, Quality of Service (QoS), Service-Level Agreements (SLAs), oversubscription ratios, etc. Decentralized approaches to implement a dynamic slot assignment scheme are also possible, in which ONUs decide when to send data and for how long. These schemes are somewhat similar to a token-passing approach, except that in this case it is a passive ring. In such a scheme, every ONU, before sending its data, will send a special message announcing how many bytes it is about to send. The ONU that is scheduled next (say, in round-robin fashion) will monitor the transmission of the previous ONU and will time its transmission such that it arrives to the OLT right after the transmission from the previous ONU. Thus, there will be no collision and no bandwidth will be wasted. This scheme is similar to hub polling 11.
However, this scheme has a major limitation: it requires connectivity (communicability) between ONUs. This imposes some constraints on PON topology; namely, the network should be deployed as a ring or as a broadcasting star. This requirement is not desirable as (a) it may require more fiber to be deployed, or (b) fiber plant with different topology might be already pre-deployed. In general, a preferred algorithm shall support any point-to-multipoint PON topology. In an optical access network, one can count only on connectivity from the OLT to every ONU (downstream traffic) and every ONU to the OLT (upstream traffic). This is true for all PON topologies.
Therefore, the OLT remains the only device that can arbitrate time-division access to the shared channel. The challenge of implementing an OLT-based dynamic arbitration scheme is in the fact that the OLT does not know how many bytes of data each ONU has buffered. The burstiness of data traffic precludes a queue occupancy prediction with any reasonable accuracy. If the OLT is to make an accurate time slot assignment, it should know the state of a given ONU exactly. One solution may be to use a polling scheme based on Grant and Request messages. Requests are sent from an ONU to report changes in an ONU's state, e.g., the amount of buffered data. The OLT processes all Requests and allocates different transmission windows (time slots) to ONUs. Slot assignment information is delivered to ONUs using Grant messages.
The advantage of having centralized intelligence for the slot-allocation algorithm is that the OLT knows the state of the entire network and can switch to another allocation scheme based on that information; the ONUs don't need to monitor the network state or negotiate and acknowledge parameters. This will make ONUs simpler and cheaper and the entire network more robust 13.
2.4 BANDWIDTH ALLOCATION
Since IEEE 802.3ah does not specify any method to allocate the time slot from OLT to the ONUs 14, it allows them to be vendor-specific. The directional properties of the splitter in EPON make the conventional Carrier Sense Multiple Access with Collision Avoidance (CSMA/CA) and Carrier Sense Multiple Access with Collision Detection (CSMA/CD) methods difficult to implement in EPON 15. This is because an ONU is unable to inform other ONUs about the collision. Although OLT is able to detect or avoid a collision and inform the ONUs about the collision by sending a jam signal, propagation delays in EPON which can exceed 20 km in length can greatly reduce the efficiency of the system. Therefore, there is no guarantee that an ONU can get an access to the medium without any delay. Since EPON must support real time traffic such as voice and video traffic, it must be able to guarantee the delivery of these traffic types.
Thus, the conventional CSMA/CD is not suitable to be used in EPON. Although CSMA/CA is able to support real time applications in EPON, it is still not efficient to be used in EPON due to its propagation delay.
The optical looping-back technique was then proposed by Desai et al. 16 to achieve high channel efficiency with CSMA/CD. With this method, a portion of upstream signal power transmitted by each ONU is looped back to the other ONUs by using a coupler. Two ports of the coupler are connected together via isolator. If two or more ONUs transmit data simultaneously, collisions will be detected at each ONU and all data transmissions will be stopped immediately. However, to implement this technique, each ONU has to use an additional receiver and a carrier sensing circuit. This is not a preferred method since it increases the network cost and is unable to provide guaranteed bandwidth.
Another possible solution is to use wavelength division multiplexing (WDM) method 17. This method allows each ONU to operate at a different wavelength to avoid collision. In order to receive the data transmitted in multiple channels, a tunable receiver or a receiver array is required at the OLT. It also requires each ONU to use a fixed transmitter operating at a different wavelength, which would result in an inventory problem. Although the inventory problem can be solved by using tunable transmitters, these devices are costly, making the solution cost-ineffective.
At this moment, time division multiple access (TDMA) is considered to be the most effective solution for upstream EPON 18. In TDMA, OLT allocates a time slot or a transmission window for data transmission in ONU. Upon the arrival of its time slot, the ONU will send out its buffered packets at the full transmission rate of the upstream channel. If there are no frames in the buffer to fill the entire time slot, idles are transmitted. TDMA can be either static or dynamic, depending on the arbitration mechanism implemented by the OLT.
Static bandwidth allocation (SBA) is simple to implement 19. With SBA, once bandwidth is assigned to an ONU, it will be unavailable to other ONUs in EPON. Due to the bursty nature of network traffic, it may result in a situation where several time slots overflow even under very light load, causing packet delay for several time slots, while other time slots are not fully utilized even under very heavy traffic. For this reason, the SBA is not preferred.
To increase bandwidth utilization, it is desirable that the OLT dynamically allocates a variable time slot to each ONU based on the instantaneous bandwidth demand of the ONUs. Dynamic bandwidth allocation (DBA) is suitable because the IP traffic is burst traffic. With DBA, when a particular ONU is not using its allocated bandwidth, that bandwidth can be reassigned to another ONU. This feature enables the flexibility of DBA to meet the different type of traffics in EPON. Without DBA, the unused bandwidth would be stranded and unusable by other ONUs on the network.
2.4.1 DBA METHODS
Figure 2.8 shows the taxonomy of DBA algorithms that is used as a framework to discuss the research on DBA. The DBA methods are divided into two categories; with quality of service (QoS) support and without QoS support.
Abbildung in dieser Leseprobe nicht enthalten
Figure 2.8 DBA taxonomy obtained from study.
2.4.1.1 DBA Algorithms without QoS Support
The first DBA algorithm for EPON that can be found in the literature is Interleaved Polling with Adaptive Cycle Time (IPACT) 20. In IPACT, the OLT polls ONUs and grants the bandwidth to each ONU in a round-robin fashion according to the ONU bandwidth demand. Each ONU is served once per round-robin polling cycle. The length of the polling cycle is not fixed where it adapts to the bandwidth requirements of the ONUs. The dynamic cycle length may result in the monopolization of bandwidth for ONUs with high traffic load. In order to prevent this, IPACT introduces maximum transfer window, Wmax. IPACT studies on several bandwidth allocation schemes namely fixed, limited, gated, constant credit, linear credit, and elastic.
Among all the six disciplines, limited scheduling discipline exhibits the best performance in IPACT. The advantages of the IPACT algorithm are that it improves bandwidth utilization by reducing the overhead occurrence from propagation delay and it allows statistical multiplexing. It also deploys an efficient in-band signaling approach that prevents the usage of extra Ethernet frames for control.
Other algorithms that do not support QoS are [21-23]. In 21, a multiple-access control scheme is proposed to provide bandwidth guaranteed (BG) service for high demand ONUs, while providing best effort (BE) service to low-demand ONUs according to the service level agreement (SLA). SLA is a contract between a network service provider and a customer that specifies what services the network service provider will furnish in measurable terms. ONUs are partitioned into two subgroups with some overlap. While frames are collected from the ONUs in one subgroup, the OLT performs DBA for ONUs in the other. Hence, the OLT continuously receives frames from the ONUs without any interruptions in 22. REPORT messages are arranged by the request length at the next transmission cycle as long as at least one ONU requests a long enough Wmax. Alternatively, when no grant length is long enough, then some requests are laid out together in the idle period to utilize the wasted idle time in 23.
The disadvantage of the DBA algorithms that do not support QoS is that they do not support the service differentiation needs of the subscribers. This causes delay in real time traffic and thus, causing the overall performance of the system to degrade.
In order to support QoS, Kramer et al. proposed two different methods; which are one logical link identifier LLID per ONU and one LLID per queue 24. The first method allocates one LLID to the entire ONU. It presents a hierarchical scheduling structure where an OLT assigns bandwidth to the ONUs and the ONUs then further subdivide the bandwidth to the multiple queues inside the ONU. The second method allocates a single LLID to each queue. It is considered as the simplest and most robust solution where it eliminates any needs for the low-level scheduler. The scheduling will only be done in OLT, where it receives a separate REPORT message from each queue and then issue a separate GATE message for each queue.
[...]