FTTH: The Overview of Existing Technologies

Dawid Nowaka and John Murphyb

aSchool Of Electronic Engineering, Dublin City University, Dublin 9, IrelandbDepartment of Computer Science, University College Dublin, Dublin 4,Ireland

ABSTRACT

The growing popularity of the Internet is the key driver behind the development of new access methods whichwould enable a customer to experience a true broadband. Amongst various technologies, the access methodsbased on the optical fiber are getting more and more attention as they offer the ultimate solution in deliveringdifferent services to the customers’ premises.

Three different architectures have been proposed that facilitate the roll out of Fiber-to-the-Home (FTTH)infrastructure. Point-to-point Ethernet networks are the most straightforward and already matured solution.Different flavors of Passive Optical Networks (PONs) with Time Division Multiplexing Access (TDMA) aregetting more widespread as necessary equipment is becoming available on the market. The third main contenderare PONs with Wavelength Division Multiplexing Access (WDMA). Although still in their infancy, the laboratorytests show that they have many advantages over present solutions.

In this paper we show a brief comparison of these three access methods. In our analysis the architectureof each solution is presented. The applicability of each system is looked at from different viewpoint and theiradvantages and disadvantages are highlighted.

Keywords: Access Networks, Optical Networks, Fiber-To-The-Home

1. INTRODUCTION

Introducing the fiber in the local loop was envisaged nearly 20 years ago.1–3 As the quality of the optical fiberwas improving, efficient transmitters and receivers appeared, it seemed possible to build an access network thatwould be based on the optical technology. Due to the lack of active units in the light path the architecture ofthe system was simple, cost effective and offered bandwidth that was not, and still is not, possible to achieve byother access methods.

However, the initial progress in the development of optical networks was slowed to a halt by economical andtechnological factors. The Internet was not as widespread then as it is now and customers were not ready topay for the broadband access. It was envisaged that a narrow band Integrated Services Digital Network accesswould be sufficient for most of the users until year 20104 and only handful of business customers would needbroadband access.

Rapid progress in development of different types of the Digital Subscriber Loop (DSL) technology and itswidespread adoption in the nineties was another factor impeding the deployment of FTTH. As the averagetransfer rate of 6 Mbps could be achieved over DSL, the cost of replacing existing copper infrastructure withoptical cables was not justified from the economical point of view.

In recent years there has been a new interest in access methods based on the optical fiber. A couple ofmajor factors are behind the motivation to revive this powerful idea. From the social point of view the Internethas become phenomenally popular and the number of customers requiring broadband access and willing to payfor it is increasing steadily despite the slumps in the global economy.5 New services such as High DefinitionTelevision (HDTV) have been developed that require more bandwidth that can be provided by DSL or CableModems. Moreover, a further growth in the number of people requiring broadband access is forecasted. Muchof the current development of broadband networks based on the optical technology has been achieved as a resultof an active and stimulating role of governmental agencies. They are responsible for implementing policies that

E-mail: {nowakd,murphyj}@ eeng.dcu.ie

encourage the development of the broadband access networks as a way to increase the computer literacy amongstcitizens. This is especially the case in Asian countries, where the penetration of broadband access is the greatest.The Japanese Ministry of Public Management estimates that by March 2006 there will be 7.73 million householdsconnected to the Internet by means of the optical fiber in Japan. Moreover, this number will be higher than thenumber of customers connected via DSL or CATV technologies.5–8

In Korea, 74% of the population already has a broadband connection to the Internet and main telecommu-nications companies are investing heavily in the optical infrastructure in order to maintain their competitiveedge.5, 8–11

In recent years China has been catching up with the leaders. The recent forecasts indicate that in the nearestfuture it will take the first place in terms of number of households with a broadband connection.12

Also in Europe broadband access is getting more and more attention. The “eEurope 2005”13 program waslaunched by the European Council in 2002. Its agenda is primarily concerned with promoting broadband Internetin the member countries. A plan of action endorsed by the Council included:

• most of the public services would be accessed online (e-government,e-learning, e-health);

• an environment for e-business would be created;

• the broadband access would be widespread and available at competitiveprices.

Substantial resources were allocated under Framework Programme 5 and 6 to spur the research and developmentof broadband access and many projects received considerable level of funding:

GigaPON Access Network (GIANT)∗ The primary objective of GIANT is to study and simulate networksbased on the GPON protocol. Partners from five universities are participating in this venture. The totalcost of the project was estimated to be 7.07 million euro with 3.85 million euro funded by the EuropeanCouncil.

Next Generation Optical Networks for Broadband European Leadership† The goal of NOBEL is tocreate intelligent and flexible optical networks that would enable broadband services for all. The totalproject budget is 24.5 million euro with 13.7 million euro of contribution from the European Council .

Multi Service Access Everywhere (MUSE)‡ The scope of MUSE project is to test different types of abroadband access for residential subscribers. With 27 organizations involved, the cost of the project is 34million euro with 18.6 million euro provided by the European Council.

Fiber based access networks are also implemented on national and regional scales in Europe without thesupport of the European Council. Sweden is a leading example as broadband access based on FTTH technologyis available to the largest number of people per 100 inhabitants§. The other Scandinavian countries are closelyfollowing the leader. An interesting example of the local deployment is a city owned FTTH network in Rotterdamand a similar one planned in Amsterdam.5, 8

The USA market is not as strongly regulated as in Europe and there is no clear government policy about thedevelopment of broadband access networks. Despite this fact, optical access networks based on point-to-point orpoint-to-multipoint architecture PONs were deployed by the local communities in around 40 different locationsin years 2001-2002¶. It is envisaged that PON access will achieve the dominant position in the near future dueto the population being more distributed than in European or Asian countries.14

∗http://www.alcatel.be/giant†http://www.ist-nobel.org‡http://www.ist-muse.org§http://www.oecd.org/sti/telecom¶http://www.ponforum.org/market/trials.asp

Another reason for a growing interest in a broadband access was the global crisis after September, 11th. Theburst of the “economy bubble” hit hard big telecommunications companies and they were forced to revise theybusiness strategies. Paltridge15 shows that the main cause of the crisis were inflated expectations about constantdemands for new equipment. Paltridge points out that long haul core networks are no longer the biggest sourceof the revenue. Instead the attention of industry is shifting towards mobile and broadband access networks, asthey recorded the fastest rate of growth and are seen as a new primary source of revenue.

The last factor is technological in nature. In the last twenty years huge progress has been made in theelectronics as well as in the optical signal processing domain. High quality optical cables are generally availabletoday, efficient light sources and receivers can be bought “off the shelf” at a reasonable price and there areelectronic circuits that match the speed of the optical devices. The availability of components makes building aFTTH system not as costly as it used to be. Successful business models were presented16, 17 which have shownthat PON systems could give better revenue than access networks based on DSL or Cable access.

2. POINT-TO-POINT ETHERNET

The widespread adoption of the Ethernet protocol in local networks and the growing demands for fast accessnetworks are the key factors that drove the vendors grouped into Ethernet in the First Mile Alliance (EFMA) towork on new architectures enabling FTTH. Through providing extensions to the existing protocols the propertiesof the new transmission medium were encompassed and two primary goals were achieved:

• Customers and network operators benefited from having a single, well known protocol in their Local Area,Access and Backbone Networks. Such an approach significantly simplified the roll out of the infrastructureand the manageability of the network.

• The reuse of legacy technologies gave the vendors the edge over the competition and new products wereavailable on the market in a shorter time.

As a result of work of the EFMA task group a number of specifications has been released in recent years anddifferent interfaces to the physical layer were defined.

• 100BASE–LX

• 100BASE–BX

• 1000BASE–LX

• 1000BASE–BX

In the proposed extensions the two end stations communicate with each other over a point-to-point con-nection in full duplex mode over a single (LX) or dual (BX) strand of single or multi mode optical fiber. Thesupported maximum transmission speed is 100Mbit/s for slower links (100BASE–X) or 1000Mbit/s for fasterlinks (1000BASE–X). In comparison with the different DSL architectures the proposed specifications enableconnecting stations which are up 10km apart. The typical architecture of network based on point-to-pointconnections is shown in Figure 1.

3. PASSIVE OPTICAL NETWORKS

The most important aspect of PON architecture is its simplicity. The Optical Line Terminal (OLT) is the mainelement of the network and it is usually placed in the Local Exchange. Optical Network Units (ONUs) serve asan interface to the network and are deployed on a customer’s side. ONUs are connected to the OLT by meansof optical fiber and no active elements are present in the link. A single ONU can serve as point of access for one(Fiber to the Home) or multiple (Fiber to the Block or Curb) customers and be deployed either at customer’spremises (Fiber to the Home or Block) or on the street in a cabinet (Fiber to the Curb). Although PONs canexist in three basic configurations (tree, bus and ring), the tree topology is favored due to smaller variation inthe signal power from different end stations.

Figure 1. Configuration of Point-to-Point FTTH Network.

The main element that distinguishes PONs from other wired networks is the unidirectionality of a coupler.This results in an architecture that is analogous in nature to a satellite or mobile communication. In thedownstream direction the signal sent by the OLT arrives at the splitter’s input and later the same signal reachesevery ONU. The signal is attenuated but otherwise its structure and properties are the same.

In the other direction, from ONUs to the OLT, the situation is different. The signals from different ONUsarrive at inputs of the splitter. Although the signals can not reach different ONUs, as they traverse through thesplitter they get mixed with each other and the superposition of all signals is received at the OLT. Hence, in theupstream direction the TDMA method is used to avoid the interference of signals from different ONUs.

A schematic architecture of PON access is depicted in Figure 2. While waiting for its opportunity, an ONUbuffers all incoming data. The contents of the queues are transmitted in a single burst using the whole availablebandwidth of the channel upon the start of an allocated transmission window.

The existing standards are the results of efforts of two different groups of network providers and equipmentvendors. The standards represent different views and attitudes towards the problem and the possible future ofthe telecommunication market.

3.1. Broadband PON Standard

Historically, the Broadband Passive Optical Network (BPON)18 standard was introduced first. It was acceptedby International Telecommunications Union (ITU) in 1999. The standard was endorsed by a number of networkproviders and equipment vendors which cooperated together in the Full Service Network Access (FSAN) ‖ group.The FSAN group proposed the Asynchronous Transfer Mode (ATM) protocol should be used to carry user data,hence sometimes access networks based on this standard are referred to as APONs.19, 20

The architecture of BPON is flexible and adapts well to different scenarios. The underlying ATM protocolprovides support for different types of service by means of adaptation layers. The small size of ATM cells andthe use of virtual channels and links allow allocating available bandwidth to the end users with a fine granularity.Moreover, the deployment of ATM in a backbone of metropolitan networks and easy mapping into SDH/SONETcontainers allows the use of only one protocol from one end user to another.

‖www.fsanweb.org

Figure 2. Configuration of PON FTTH Network.

Yet, the advantages of ATM proved to be the main obstacle in deployment of BPON and despite many fieldtrails21, 22 BPON did not gain much popularity. The complexity of the ATM protocol was hard to implementand in many cases superfluous. Much simpler, data only oriented Ethernet protocols found a widespread use inlocal area networks and started to replace ATM in many metropolitan area and backbone networks.

3.2. Gigabit PON Standard

The progress in the technology, the need for larger bandwidths and the unquestionable complexity of ATMforced the FSAN group to revise their approach. In the outcome a new standard called Gigabit Passive OpticalNetwork (GPON)23 was released and adopted by ITU in 2003.

The GPON’s functionality is heavily based on its predecessor, although it is no longer reliant on ATM asan underlying protocol. Instead a much simpler Generic Framing Procedure (GFP) is used to provide supportfor both voice and data oriented services. A big advantage of GPON over other schemes is that interfaces toall the main services are provided and in GFP enabled networks packets belonging to different protocols can betransmitted in their native formats. The functionality is provided which allows seamless interoperability withother GPONs or BPONs. As in modern networks the security of transmitted data is a key issue. A sophisticatedmechanism based on Advanced Encryption Standard and a complex exchange of unique keys is built into theGPON architecture.

Also in comparison with the BPON standard, higher transmission rates are specified making GPON capableof supporting transfer rates of up to 2.48 Gbps in the downstream as well as the upstream direction.

3.3. Ethernet PON Standard

The Ethernet Passive Optical Network (EPON) standard has been endorsed by the EFMA. The final version ofthe new protocol and necessary amendments to the existing ones were accepted by Standard Body and released asIEEE 802.3ah in September 2004.24 The main goal was to achieve a full compatibility with other Ethernet basednetworks. Hence, the functionality of Ethernet’s Media Access Control layer is maintained and the extensionsare provided to encompass the features of PONs. The achieved solution is simple and straightforward, and thelegacy equipment and technologies can be reused similar as in 100BASE–X and 1000BASE–X networks.

Figure 3. Configuration of WDM-PON FTTH Network.

4. WAVELENGTH DIVISION MULTIPLEXING ACCESS PASSIVE OPTICALNETWORKS

Wavelength Division Multiplexing Passive Optical Network (WDM PON) are the next generation in developmentof access networks. Ultimately, they can offer the largest bandwidth at the lowest cost. In principle, thearchitecture of WDM PON is similar to the architecture of the PON. The main difference is that ONUs operateon different wavelengths and hence higher transmission rates can be achieved. The illustrative example of thisarchitecture is shown in Figure 3.

The main problem with WDM PONs is that usually the wavelength is assigned to an ONU in a fixed manner.This makes upgrades in the network topology difficult as they require manual reconfiguration of the equipmentin the customers’ premises, which significantly increases the cost of maintenance. The solution to this is thedevelopment of so called “colorless” ONUs. In such a scheme the ONU detects what wavelength is used inthe downstream direction and sends its data on this wavelength in the upstream direction. WDM PON with“colorless” ONUs is presented by Park et al.25 The authors proposed to use amplified spontaneous emissionFabry-Perot laser diode. They showed that such a diode can be locked to the signal from a spectrum-slicedbroadband light source. A different approach is presented by Kani et al. The authors showed a system where128 different wavelengths were used to carry the signal to the customers’ premises.26 In the upstream direction,an ONU modulates the carrier wavelength provided by the OLT with its data. The advantage of such an approachis that ONUs do not have to be equipped with expensive light sources. This not only lowers the overall costof the equipment but also makes ONUs transparent to the signal and different wavelengths can be used at anytime.

The disadvantage of WDM PONs is the high cost of equipment. Much research was focused on enhancingWDM PONs ability to serve larger numbers of customers in attempt to increase revenue from invested resourcesand its cost efficiency. As a result, some hybrid structures have been proposed where both WDMA and TDMAmodes are used to increase the number of potential users. The typical architecture of such a network is presentedin Figure 4. Shin et al. presented a system where 16 different wavelengths27 were used in the WDMA mode. Thenumber of possible end-points was further increased by employing a TDMA mode. This allowed 8 ONUs to utilizethe same wavelength. In total, in such a Wavelength Division Multiplexing/Time Division Multiplexing PassiveOptical Network (WDM-TDM PON), 75Mbit/s access rates were provided to 128 customers. Less complex andexpensive equipment was needed than in WDM PONs with a similar number of connected customers.

Figure 4. Configuration of WDM-PON FTTH Network.

5. TECHNOLOGY COMPARISON

In this section we will present a comparison of key aspects of different FTTH architectures. The brief overviewis shown in Table 1.

5.1. Available Bandwidth

With advancements in the development of HDTV, it can be envisaged that future access networks will have todeliver between 70 and 100Mbit/s per customer. This will enable the customer to watch 3 to 4 HDTV channels,browse the Internet and carry a number of video calls. Currently, all types of FTTH architectures are capable ofproviding such a bandwidth. In Ethernet-based point-to-point networks 100Mbit/s rates are explicitly supportedin 100BASE–X networks. The rates offered by PONs are somewhat lower and in the typical configuration around70Mbit/s can be delivered. The results recorded for WDM PON show that much larger transfer rates than thenecessary minimum can be provided.25–27

5.2. Network Reach

The maximum network reach is limited by the power budget of a particular system. In 100BASE–X and1000BASE–X networks, it is specified that the network should have a reach of up to 10km. In standardsspecifying PON’s functionality two distances 10 and 20km are considered. In experimental WDM PONs, asuccessful transmission was possible over similar distances.

5.3. Cost

The total cost of the infrastructure is dependent on several factors such as the price of the materials, availabilityof underground or aerial ducts, the terrain topography and local labor rates.

Table 1. Comparison of different FTTH architectures in their typical configurations.

Architecture Link Rate Split Ratio CustomerBandwidth

Reach Availability

100BASE–X 100Mbit/s 1:1 100Mbit/s 10km on the market1000BASE–X 1000Mbit/s 1:1 1000Mbit/s 10km on the market

BPON↑ 622Mbit/s↓ 150Mbit/s 1:16

40Mbit/s9.3Mbit/s 10-20km on the market

GPON 1.2Gbit/s 1:16 75Mbit/s 10-20km on the marketEPON 1.2Gbit/s 1:16 75Mbit/s 10-20km in development

WDM-PON 1.25Gbit/s/λ 1:128 1.25Gbit/s 10km laboratoryWDM/TDM-PON 1.25Gbit/s/λ 1:128 75Mbit/s 10km laboratory

In terms of the initial expenditure on necessary infrastructure, PONs and WDM PON have the advantageover the other access methods as their architecture permits to spread the total cost over a larger number ofcustomers. This is especially important when the cost of the fiber infrastructure is compared. In 100BASE–X and 1000BASE–X networks, the end stations must be connected with a separate fiber. Also for each fiberseparate terminating equipment must be provided on both sides. These two factors significantly increase theinitial cost. In PONs and WDM PONs, even if the initial cost was higher, it could be shared amongst a largernumber of customers as one OLT can serve multiple ONUs.

The maturity of technology has also an impact on the total cost and the choice of one technology over anotherone. The equipment for well known solutions tends to be less expensive, more reliable and readily available. Fromthis point of view, 100BASE–X and 1000BASE–X networks have an advantage over others as these solution arewell matured and easily available on the market. The growing interest in PONs resulted in quick developmentof the equipment and the gap between these two architectures is closing fast. Despite the fact that WDM PONsare still in the laboratory phase, systems utilizing the legacy protocols are already available on the market.

6. OUTLOOK

Recent market research shows that, in the near future, the development of FTTH access will be slower thanexpected. The report presented by Yankee Group28 shows that introduction of mass broadband access based onoptical fiber could be inhibited by a number of factors.

• Services such as cable television, voice and data access are widely available and they utilize the existinginfrastructure e.g. copper and radio links. As these solutions are already on the market it is going to bedifficult to persuade customers’ to embark on using new technology without offering substantial benefits.

• The regulatory policy remains unclear. The decision has not been made whether impose unbundling onfuture optical networks owned by an operator with large market share or not. This uncertainty discouragespossible investors.

• Lack of applications that would drive the demand for larger bandwidth with the exception of HDTV. In anumber of countries (England, Germany) small interest in other advanced technologies (video on demand)has been recorded.

The Irish market does not differ much from the European average. In last two years many companies startedoffering broadband access based on DSL and wireless technology. Cable television operators are not behind andtheir customers can access Internet with Cable modems. All these factors inhibits the roll out of access networksbased on optical fiber.

7. SUMMARY

In this paper the comparison of the main FTTH architectures has been presented. The common features aswell as the differences between three main architectures were outlined. We showed that although point-to-pointEthernet optical networks have many advantages their development could be more expensive in the long runthat the potential benefits.

It is envisaged that, in the nearest future, access based on the PONs will be dominant. Primarily becausethis technology is maturing fast, it allows the total cost to be shared amongst a larger number of customers. It ispredicted that as a number of available HDTV channels grows even more bandwidth will be required. This couldprovide an incentive to quicken the development of FTTH networks based on WDMA mode and significantlyreduce their time to market. Despite being in the laboratory phase, this technology might become available onthe market in the next couple of years.

In comparison with other access technologies such as DSL or Cable, FTTH offers much bigger bandwidth.The main obstacle faced by network operators are the prohibitive costs of the rolling out the infrastructure. Themassive roll out of FTTH infrastructure in nearest future is difficult to anticipate in current conditions. Thehigh penetration of existing services and lack of new “killer applications”, that would increase the demand forbandwidth, and slower than expected start of HDTV can be named as major obstacles.

ACKNOWLEDGMENT

The support of both Science Foundation Ireland and the Informatics Research Initiative of Enterprise Ireland isgratefully acknowledged.

REFERENCES1. J.R. Stern, J.W. Ballance, D.W. Faulkner, S. Hornung, and D.B. Payne, “Passive Optical Local Networks

for Telephony Applications and Beyond,” Electronics Letters 23, pp. 1255–1257, Nov. 1987.2. Donald. E.A. Clarke and Tetsuya Kanada, “Broadband: The Last Mile,” IEEE Communications Magazine

31, pp. 94–100, Mar. 1993.3. Yih-Kang Maurice Lin, Dan R. Spears, and Mih Yin, “Fiber-Based Local Access Network Architectures,”

IEEE Communications Magazine 27, pp. 64–73, Oct. 1989.4. Borgar T. Olsen, Alcibiade Zaganiaris, Kjell Stordahl, L. Aa. Ims, D. Myhre, T.

Øverli, M. Tahkokorpi, I. Welling, M. Drieskens, J. Kraushaar, J. Mononen, M. Lahteenoja, S.Markatos, M.De Bortoli, U. Ferrero, M. Ravera, S. Balzaretti, F. Fleuren, N. Gieschen, M. De Oliveira Duarte, and E.de Castro, “Techno-Economic Evaluation of Narrowband and Broadband Access Network Alternatives andEvolution Scenario Assessment,” IEEE Journal on Selected Areas in Communications 15, Aug. 1996.

5. OECD – Organisation For Economic Co-Operation and Development, “OECD Communications Outlook,”June 2003.

6. Y. Maeda, “Broadband Access Deployment in NTT.” http://www.fsanweb.org/Atlanta Workshop.asp,Mar. 2003. FSAN PON Workshop, Atlanta, USA.

7. Y. Maeda, “Overwiew of Optical Broadband in Japan,” in Proceedings of ECOC, 1, Sept. 2004.8. B. Whitman, “Fibre Access Deployment Worldwide: Market Drivers, Politics and Technology Choices,” in

Proceedings of ECOC, pp. 6–9, Sept. 2004.9. J. Hongbeom, “Broadband access, service developments and trends in Korea,” in Proceedings of ECOC, 5,

Sept. 2004.10. C. H. Yim, “Convergence Acess Network in Korea.” http://www.itu.int/ITU-T/worksem/asna/index.

html, June 2004. All Star Network Access Workshop, Geneva, Switzerland.11. J.-W. Yang, “Broadband convergence Network (BcN) - a Korean National Broadband Project,” in Proceed-

ings of OECC/COIN, pp. 50–51, July 2004.12. M. Kunigonis, “International Broadband Deployments. Lessons learned around the globe..” http://www.

apectelwg.org/apecdata/telwg/28tel/bw/telwg28-BB-12.pdf, Oct. 2003. APEC Telecommunicationsand Information Working Group, Taipei, Taiwan.

13. C. O. T. E. Communities, “eEurope 2005: An information society for all.” http://www.europa.eu.int/information society/eeurope/2005/index en.htm, June 2002.

14. John A. Jay, “An Overview of International Fiber to the Home Deployment,” in Proceedings of FTTHConference, Oct. 2002.

15. S. Paltridge, “Industry in Crisis (?),” in Proceedings of CEPT Conference, 2003.16. Dana Bisaro, “FTTH-”HFC vs. PON”,” in Proceedings of FTTH Conference, Oct. 2002.17. Patrick T. Garvey, “Economics of FTTH,” in Proceedings of FTTH Conference, Oct. 2004.18. ITU-T, “G.983.1 - Broadband Passive Optical Networks (BPON): General characteristics,” June 1999.19. David Faulkner, Rajendrakumar Mistry, Tom Rowbotham, Kenji Okada, Wsewolod Warzanskyj, Albert

Zylbersztejn, and Yves Picault, “The Full Services Access Networks Initiative,” IEEE CommunicationsMagazine 35, pp. 58–68, Apr. 1997.

20. Yoichi Maeda, Kenji Okada, and David Faulkner, “FSAN OAN-WG and future issues for broadband opticalaccess networks,” IEEE Communications Magazine 39, pp. 126–132, Dec. 2001.

21. Ingrid Van de Voorde and Gert Van der Plas, “Full Service Optical Access Networks: ATM Transport onPassive Optical Networks,” IEEE Communications Magazine 35(4), pp. 70–75, 1997.

22. Hiromi Ueda, Kenji Okada, Brian Ford, Glenn Mahony, Stephen Hornung, David Faulkner, Jacques Abiven,Sophie Durel, Ralph Ballart, and John Erickson, “Deployment status and common technical specificationsfor a B-PON system,” IEEE Communications Magazine 39, pp. 134–141, Dec. 2001.

23. ITU-T, “G.984.1 - Gigabit-capablePassive Optical Networks (GPON): General characteristics,” Mar. 2003.24. The Institute of Electrical and Electronics Engineers, “IEEE 802.3ah - IEEE Standard for Local and

Metropolitan Area Networks. Part 3: Carrier Sense Multiple Access with Collision Detection (CSMA/CD)Access Method and Physical Layer Specifications. Amendment: Media Access Control Parameters, PhysicalLayers, and Management Parameters for Subscriber Access Networks,” Sept. 2004.

25. S.-J. Park, C.-H. Lee, K.-T. Jeong, H.-J. Park, J.-G. Ahn, and K.-H. Song, “Fiber-to-the-Home ServicesBased on Wavelength-Division-Multiplexing Passive Optical Network,” IEEE/OSA Journal of LightwaveTechnology 22, pp. 2582–2591, Nov. 2004.

26. J. ichi Kani, M. Teshima, K. Akimoto, N. Takachio, H. Suzuki, K. Iwatsuki, and M. Ishii, “A WDM-based Optical Acess Network for Wide-Area Gigabit Access Services,” IEEE Communications Magazine41, pp. 43–48, Feb. 2003.

27. D. J. Shin, D. K. Jung, H. S. Shin, J. W. Kwon, S. Hwang, Y. Oh, and C. Shim, “Hybrid WDM/TDM-PONWith Wavelength-Selection-Free Transmitters,” IEEE/OSA Journal of Lightwave Technology 23, pp. 187–195, Jan. 2005.

28. J. Doran, “Mass-Market Fiber Remains Distant on the European Horizon,” Dec. 2004. http://www.yankeegroup.com.