Performance Analysis of Next Generation-PON (NG-PON) Architectures

A. M. Ragheb1 and Habib Fathallahl2

I
King Saud University, KSA
2
Prince Sultan Advanced Technologies Research Institute (PSA TRI), KSA
Abstract-Next generation passive optical networks (NG
PONs) is the natural development of PONs toward achieving
higher data rates, larger counts of wavelength channels, and
longer fber ranges. NG-PON can be implemented as high speed
time division multiplexing (TDM), wavelength division
multiplexing (WDM), Hybrid TDMIWDM, optical code division
multiplexing (OCDM) PON. Many enhancements and
adaptations are occurring in order to offer higher bandwidths
with higher number of subscribers. This includes coverage area
that is increasing to reach of lOOkm and more, e.g. the LR-PON.
In addition, the wireless access network is gradually going to be
integrated with PON systems, e.g. FiWi network. In this paper,
we examine the different promising architectures for NG-PON.
A comparison between key network specifcations such as data
rates and power budget has been reported. Moreover, we study
different state-of-the-art technologies that will potentially be
used for NG-PONs. We discuss the different characteristics of
NG-PON showing the important contributions and challenges of
various NG-PON demonstrations.
Inde terms-FTTH, Next generation passive optical network
(NG-PON), NG-PON architectures, LR-PON, FiWi.
I. INTRODUCTION
Fiber-to-the-Home (FTTH) continued to make great
progress in recent years. During 2011, the number of users
has increased by 18 millions to reach approximately 80
millions [1]. Asia-Pacifc, North America, and Europe are the
leading regions in deploying FTTH. By the end of 2013, it is
expected that the number of FTTH household will reach 30
millions in Japan and approximately 20 millions in US with
FTTH total of 130 millions worldwide [1]. FTTH can be
realized by gigabit-class passive optical networks (PONs).
The two major state-of-the-art networks are the gigabit PON
(GPON), widely deployed in U.S. and Europe, and the
Etheret PON (EPON), adopted in Japan and Korea. Both
networks differ in the data link layer protocol used (i.e.
GPON uses the GPON encapsulation method (GEM) whereas
EPON uses Etheret) [2]. ITU-T and IEEE have generated
the G.984 and IEEE 802.3ah standards, respectively. In
general, PON networks provide a point to multipoint
transmission between the optical line terminal (OL T) and the
several optical network units (ONUs). With the continuous
increasing of bandwidth required by different types of users,
gigabit-class PONs have been evolved into NG-PONs.
NG-PONs promise to satisf higher performance
characteristics such as higher data rates, lower maintenance
cost, and larger number of users. These features can be
achieved by deploying new topologies, larger counts of
wavelength channels, longer fber ranges, and higher splitting
ratios. In general, NG-PON will be deployed in two stages
[3]. The frst stage represents the evolutionary scenario (e.g.
NG-PONl) which supports the coexistence of NG-PON with
legacy PON (gigabit PON) on the same optical distribution
network (OON). In this scenario, the migration is done
incrementally in order to reduce the capital expenditures
(CAPEX) and achieve a minimum service interruption to the
subscribers that are not migrated. Second stage denotes the
deploying of the new technology in a green feld or the total
replacement of the existing gigabit PON in order to take
advantages of NG-PON features (e.g. NG-PON2).
Different architectures have been suggested in the
literature so as to build the NG-PON different topologies.
Firstly, TOM networks use the new standards developed by
IEEE and ITU-T new to deploy a high speed TOM networks.
Secondly, the high bandwidth offered by WOM network
where a dedicated wavelength is assigned for every ONU
attracts vendors to deploy WOM for the NG-PON. A third
category integrates TOM and WOM network in one hybrid
TOM/WOM system. Moreover, the maximum span of the
traditional PON network ( 20 k) has been increased to
reach 100 km by deploying long reach-PON (LR-PON)
architectures. Finally, FiWi networks combine the mobility of
wireless networks and the high capacity of PONs.
In this paper we provide a study for the state-of-the-art
architectures for NG-PONs. We compare the specifcations
of different next generation TOM-PON. We make a
comparison of different technologies used in WOM, TOM
over WOM, LR-PON, and FiWi networks and their key
characteristics. Finally, the main challenges toward
commercially adoption of fture access networks have been
discussed. The organization of the paper is as follows. In
section II we review the different specifcations and
architectures of high speed TOM, WOM, and TOM over
WOM networks. Section III and IV review the state-of-the
art technologies used in LR-PON and FiWi networks. In
Section V a discussion has been reported. Section VI
concludes the paper.
II. HIGH SPEED TOM, WOM, AND TOM OVER WOM ARCHITECTURES
In recent years, user bandwidth has been increased more
than 10 fold; this huge increment is actually due to different
user's requirements such as video on demand (VoO), high
defnition television (HOTV), Interet, and business
339
applications. In addition, the developed wireless networks
such as IEEE 802.11n wireless LANs need EPON as a
backhaul in order to achieve lOOMbps or higher per access
point [4]. Different trends have been proposed as in follows:
A. High speed TDM-PON
In September 2009, IEEE 802.3av Task Force completed
the development of the 10G-EPON specifcations. 10G
EPON can support symmetric 10Gbps downstream and
upstream, and asymmetric 10Gbps downstream and 1 Gbps
upstream data rates. The main differences between lOG and
IG-EPON are summarized in TABLE 1[5].
TALE!
MAIN DIFFERENCES BETWEEN IG-EPON AND 10G-EPON
Data rate
Line coding
lit ratio
Wavelengths
FEC
Power budget classes
IC-EPON IOC-EPON
1/1symmctrc Jl/Jlsymmctrc
Jl/1 asymmctrc
8H/1 lHovcrhcad(Z0) 1H/Hovcrhcad(J)
1:1 1:I/1.JZ
1180-illnm (O5) 1575-i80 nm(O5)
1Z0-1J0nm(!5) 1Z0-1Z80nm(I5)
R5 Z55,ZJv R5 Z55,ZZJ
Zci asscs Jci asscs
lR(10/Z0) lR(i0/Z0/J0)
lRX(10/Z0/J0)
For the same objective, ITU-T completes its standard, ITU
T G.987, in 2010 which addessed the general requirements
of 10 Gigabit-passive optical networks (XG-PON) [6]. ITU-T
G.987 supports symmetric lOGbps, and asymmetric, lOGbps
downstream and 2.5Gbps upstream, date rates. The main
differences between GPON and XG-PON are summarized in
TABLE II. A block diagram of high speed TDM PON is
shown in Fig. 1.
TALE II
MAIN DIFFERENCES BETWEEN GPON AND XG-PON
Data rate
Reach distance
!plit ratio
Wavelengths
FEC
Power budget classes
CPON
Z.5/i.Z
10km,Z0km
(physicaIdistanccs)
I.JZ, 1.1
1180-1500nm(O5)
1Z0-1J0nm(L5)
XC-PON
il/Z.5X-lNi
Il/IlX-lNZ
Z0km(0kmin
thcIogicaIIaycr)
i.1,(i.Z5in
IogicaIIaycr
157-1580nm(O5)
1Z0-1Z80nm(L5)
R5(Z18,Z1) O5
R5(18,ZJZ U5
CIass H+ (1J-Z8dH NominaI (Zv-JidH)
opticaIbudgct) Ixtcndcd(JJ-J5dH)
Fig. I Next gcncrationTOM-PON
978-1-4577-1169-5/11/$26.00 2011 IEEE
B. WDMPON
WDM-PONs have been investigated as a potential
technology for NG-PON. Different architectures have been
proposed to implement WDM-PONs[7], some of them are
based on the legacy TDM-PON (i.e. broadcast and select) but
with a band pass flter at each ONU with small insertion loss,
as shown in Fig. 2 (a). A band splitter (BS) is used to
differentiate between up and downstream wavelengths. This
architecture is suffered fom passive splitter loss, lack of
security, as well as the difculty of employing identical
ONUs (colorless NU). Another type of WDM-PONs is
based on the use of an arayed waveguide grating (A WG)
instead of power splitter in order to specif a dedicated pair
of wavelengths for each ONU, as shown in Fig. 2 (b). Tis
confguration provides lower A WG' s insertion loss and
simpler NU's receivers. Nonetheless, diferent NU's
transmitters are still needed. The mentioned confgurations
are based on dense WDM. The cost of WDM network can
highly reduce by considering the coarse WDM (CWDM)
network where the wavelength spacing is sufciently high to
relax the transmitters to be accurately controlled. TABLE III
summarized the basic features of different WDM networks
[7]. Note that the ONU link cost is defned with respect to
GPON cost (GPON cost = 100 percent for 20 dB power
budget).
1A0ii!!!
IIA1LRI5 IHA5!CVOlNI1VR!5
5pI ittcr-bascd AV-bascd CVOl-lN
VOllN VOllN
O5/!5bitratc 1,Z.5,1,10' 1,Z.5,1,10' 1,Z.5,1,10'
VavcIcngthgrid 100, 50'Hz 100,50'Hz Z0nm
lax. spIitting 1.JZ 1.10 1.8
ratio
lax.budgct 8dH 10dH 1JdH
ONL Iink cost Z1Z(1'pcr1) Z5(1'pcr1) 1ZZ(1'pcr1)
jrcIativc|
So as to decrease the cost, simplif the maintenance, and
unif the design of ONUs, different technologies have been
considered to design a colorless ONU (i.e. wavelength
independent ONU). Techniques for implementing a colorless
ONUs include injection locking and wavelength
seeding/remote modulation, spectrum slicing, and wavelength
tuning[8]. In wavelength-seeding scheme a refective
semiconductor optical amplifer (RSOA) is used in every
NU's transmitter. The received wavelength is amplifed,
modulated by the upstream data, and refected back to the
OLT as shown in Fig. 2(c). The up and down streams can be
transmitted on different wavelengths or they can have the
same wavelength with different extinction ratios (ER). In the
spectrum slicing technique, each NU's transmitter employs
a broadband light source such as super luminescent diode
(SLD). The lightwave is modulated by the upstream data and
sent to the OL T using MUX/DeMUX device [9]. The
MUX/DeMUX is used to slice the optical spectrum of each
340
CM\
tC L
FCwer/
wvelengt!
oivioer
(c)
ML
tCL
Fig. ZNext generation WDM-PON; (a) splitter-based WDM-PON, (b) AWG-based WDM-PON,
(c) RSOA-based ONU, (d)spectrum slicing- based ONU
ONU at different wavelengths, as shown in Fig. 2(d).The
simplest way to incorporate a colorless-ONU is by using a
tunable wavelength transmitter at each ONU. Different
designs have been proposed to realize a low cost tunable laser
source [10].
C Hybrid TDMWDM PON
As shown in Fig. J, hybrid TDM/WDM PON has the
benefts of the high capacity offered by TDM-PON and the
high number of wavelengths offered by WDM-PON.
Generally, hybrid TDM/WDM PONs can be categorized into
static and dynamic hybrid schemes. In static scheme, the
wavelengths (US and DS) assigned for each ONU can't be
changed during operation. Whereas, in the dynamic case the
wavelengths are changed dynamically according to operation
and communication [8]. Dynamic scheme offers some
advantages to PON including load balancing, resilience, and
power saving fnction, but in the same time the ONU s must
employ burst mode receivers due to the fequent change in
wavelengths. In [11], Lee et al. proposed the frst commercial
colorless gigabit TDMWDM hybrid PON to be used in
Korea Telecom. Hybrid WDM/TDM can operate on DWDM
or CWDM grid wavelengths with the main differences listed
in TABLE IV [7].
TABLE IN
H`HR!OOL^5LA^O CAR5L1Ol/VOl PON
CVOl/1Ol OVOl/1Ol
O5/!5bit rate
VavcIcngthg t .d

Z

0

Max. s Iittngatio 1+
lax. budgct 11dH
N! Ink cost 1JZ(Z.5per A)
[rcIatvc|
Up to1b s
100(50)Hz
1.10
11dH
ZJ+(Z.5per A)
Fig. JNext generation WDMlTDM PON
Ill. LR-PON
In order to overcome the limitations of TDM and WDM
passive optical networks, such as low splitting ratio and
range, LR-PON has been developed as a more cost effective
solution for optical access network [12]. LR-PON can
simplif the network by combining the metro and access
segments into one large network. This architecture can reduce
the total number of optical-electrical-optical (OEO)
conversions, at the expense of mid-span optical amplifers
[2].
Different LR-PON architectures have been developed,
the photonics local access network (PLANET) project was
initiated in mid of 1990 to develop a super passive optical
network (SuperPON). SuperPON represents an upgrade
version of G.983 broadband PON (BPON). It can support up
to 2048 split size over a range of 100km. A long reach PON
was developed for British Telecom (BT) in order to satisf
their predicted bandwidth and consolidate the number of
central ofces through the country. In this architecture, the
split ratio was targeted to be 1024 with 100km range and
10Gbps bit rate in both down- and upstream, as shown in Fig.
341
4 (a). A hybrid DWDM/TDM long-reach PON was
developed by Tail and Townsend [13]. This demonstration
can support 17 TDM PONs each with 256 customers, total of
4352 ONUs, with total reach of 100kI, shown in Fig. 4 (b).
The reach of the GPON system has been increased fom
20km to 135km using an OEO conversion and amplifer at
the local exchange point, as shown in Fig. 4 (c). The OEO
converts the burst mode GPON network to a DWDM
network. Instead of OEO converter, a wavelength converter
has been used in the wavelength converter PON (WC-PON)
proposed by Shea and Mitchell [14]. In this demonstration, a
cross-gain modulation (XGM) wavelength converter (WC)
was used to convert the ONU wavelength to a standard
DWDM wavelength. The reach of this network is 120 k
with a total of 20 PONs each with 64 users, total of 1280
users, as shown in Fig. 4 (d). A comparison between different
LR-PON architectures is summarized in TABLE V. The
propagation delay time between the OL T and different ONUs
b challenging for LR-PON. By increasing the distance to
100k the round trip time (R TT) will increase to fom
O.2msec in traditional PON to Imsec. A multithread
algorithm has been proposed to increase the number of
polling threads in order to decrease the propagation delay
time [15].
IV. FiWi NETWORKS
Optical access networks can provide huge bandwidth but it
can't go everywhere, on the other hand wireless access
networks can go everywhere but its offered bandwidth is
limited [16]. By combining the capacity of optical fber
network and the mobility of wireless network fber wireless
(FiWi) network can provide customers with a quad-play
services (data, video, voice and mobility) [17]. FiWi network
can be developed as radio over fver (RoF) or radio and fber
(R&F).RoF has been developed as an approach to combine
optical fber and wireless networks.
TALE N
DIFFERENT FEATURES OF LR-PON ARCHITECTURES
Reach distance
Splitting ratio
Bit rate
Bit rate/user
BER
wavelengths
SuperPON
100 km
2048 (i.e. 1:16 x 1:64
x 2)
2.5 Gbps OS, 311
Mb s US
1.22 Mbps (OS)
0.15 Mbps( US)
10.9
1550 nm (DS)
1310 nm (US)
90km

Long reach PON
II0km
1024 ( i.e. 1:4 x I: 16 x
I:I
10 Gbps symmetric
9.76 Mbps(OS & US)
10.10
1550 nm for both OS,
US
1 sp|N er 10 km

1:16

sp|mer
Hybrid OWOMffOM GPON extension
PON
100 km 135 km
4352 (i.e. 17 PON x 64
256 userl
10 Gbps symmetric 2.5 Gbps OS
1.25 Gb s US
39 Mbps(OS & US) 39 Mbps (OS)
19.5 Mbps (US)
Better than I 0.9 Better than 10.10
OWOM grid OW DMgrid
125 km
:,
BackhLr----------------------
!
Distribution section
OLT
section
Local
exchange
Core
exchange
(b)
(a)
17' PON
Fig. 4 Next generation LR-PON
Core
exchange
(c)
(d)
WC-PON
120 km
1280 (i.e. 20 PON x
64 user)
2.5 Gbps OS
2.5 Gb s US
39 Mbps (O5&!5)
10.9
OW OM grid
10km
Distribution section
PON1
342
In RoF, radio fequencies are carried over optical fber, by
means of hybrid fber radio (HFR), between central ofce
(CO) and multiple low cost remote antenna units (RAUs).
RoF networks suffer fom propagation delay which
deteriorates network performance. This limitation can be
avoided by employing R&F network, where the access to
optical and wireless media is controlled separately fom each
other by using different MAC protocols in the optical and
wireless media [17].
Different FiWi architectures have been proposed in the
literature. Muralidharan et al. [17] proposed a hybrid
integrated fber wireless networking (HiFi-WiN), where the
connection between the central ofce (CO) and the wirless
access points (W APs) is done via a unidirectional fber ring,
as shown in Fig. 5 (a). Each WAP unit provides a wireless
access to the mobile users in the service area. In this
architecture the CO is responsible for assigning channels to
accommodate diferent services. A self-healing architecture
has been proposed by Lin et al. in [18]. As shown in Fig. 5
(b), the confguration is based on a two level bidirectional
path protected ring (BPR). The CO is connected to the remote
nodes via dual fber ring. Each R connects many W AP by
means of concentration nodes (CN). Finally, each WAP
|=)
|I
wirelessly served different users. In this architecture, the
W AP assigns at least 5 MHz chanel bandwidth and covers
up to 16 users. The distributed controllers placed at CO and
each R provides a self-healing property for this architecture.
Bhandari and Park proposed a hybrid star/ring wireless
access network that has the property of providing bandwidth
on demand [19]. As shown in Fig. 5 (c), the central ofce
connects many ring networks via optical switches, which has
fll wavelength capability. In addition, the optical switch
connects each ring network to two neighbor rings in order to
provide network protection. The network is periodically
monitored in order to balance the load or add new lightpath
according the load and bandwidth offered, respectively. In
[20], Shaw et at. proposed a hybrid optical wireless access
network that consists of optical WDM ring and wireless mesh
networks (WMNs) connected together via a different PONs,
as shown in Fig. 5 (d). Wireless gateways are used to bridge
the PONs and the WMNs. The system provides
confgurability according the load trafc. The features of the
different state-of-the-art architectures are summarized in
TABLE VI.
CACM
|)
|g.5Ncxtgcncraton |Wnctwork
ccongurat o n __
lrotccton
LhanncIass nmcnt
lNuscd
VMN
HIi-Vl
ycs
^o
L
No
^o
1mLNI
Ol|||K|N1||A1K|5||lWlARLHl1|L1K|5 [171
HdircctionaIring
lo
ycs
VAl
lo
lo
Hybrid star/ring
ycs
ycs
VAl
lo
lo
Hybridring/VMN
ycs
ycs
N!-vircIcssgatcvay
ycs
ycs
343
V. DISCUSSION
According to the different implementations discussed in
the literature, a comparison between different NG-PON
architecture's features is summarized in TABLE VII. The
new TDM standards, IEEE 802.3av and ITU-T G.987,
enforce technical challenges in terms of power budget and
design of dual rate burst mode OL T's receivers. NU's cost
is another challenging aspect for high speed TDM networks.
10GE-PON and XG-PON provide smooth migration fom
gigabit class access networks to higher data rates PON. The
migration will maintain the main infastructure components
that are already deployed. WDM network provide the main
user requirements in terms of higher bandwidth and security.
The high cost of WDM networks can be reduced by
deploying CWDM networks in combination with DWDM
networks. The design of low cost and high efciency
colorless ONU are the main challenges in order to have a
cost-effective user's ONU. Recently, CIP company
introduced a TO-CAN packaged refective SOA at about 200
$ price. In addition, the deployment of hybrid TDM over
WDM networks will provide users with higher bandwidth
and lower cost access network. The main challenges in this
demonstration will be in the design of efcient dynamic
bandwidth allocation (DBA) schemes as well as dynamic
wavelength allocation (DWA) algorithms for the dynamic
hybrid system. Also, tunable transmitters are still required for
the dynamic hybrid scheme.
LR-PON has the capability to reduce the maintenance of
the network and increase its distance. The deployed
architecture should be accurately selected so as to reduce the
propagation time which is a main challenging point in LR
PON. A tradeoff between the splitting ratio and the power
budget for different LR-PON layouts has to be considered.
FiWi networks should consider new approaches to take
benefts of the huge bandwidth offered by optical access
network. Moreover, new reconfgurability algorithms and
routing protocols are still needed so as to optimally balance
the unexpected trafc conditions. Self-healing FiWi network
is another challenging point to increase the network
survivability. The cost of FiWi network is a major challenge
toward widely deploying FiWi networks commercially.
Mid-term and long-term PONs can support higher bit
rates, 40, 100, and even Tera bit/s, by considering bandwidth
effective modulation techniques. Various modulation
schemes can achieve better performance characteristics in
terms of bit rate, span distance, and cost. Multilevel
modulation techniques such as MPSK and MQAM in
addition to polarization diversity ease the speed requirements
on electronic/optoelectronic components and achieve higher
bit rates. Additionally, OCDMA has unique features such as
asynchronous transmission, low latency access, sof capacity
on demand, and optical layer security. These features make
OCDMA a good candidate for NG-PON networks. The
demand of higher data rates over the dispersive optical
channel and the development of DSP technology over optical
channel will encourage researchers to implement NG-PON
networks based on optical OFDM (OOFDM). OOFDM not
only increases the spectral effciency and provides good
resistance to fber impairments such as chromatic dispersion
(CD) and polarization mode dispersion (PMD), but also it
provide a natural compatibility with DSP-based systems.
VI. CONCLUSION
NG-PON is the best solution for the ever-increasing
number of interet users and hungry bandwidth applications.
Different architectures have been considered to fulfll the
NG-PON requirements including high speed TDM, WDM,
Hybrid TDM/WDM, LR-PON, and FiWi PON networks.
High speed TDM-PON provides the easiest migration
scenario to short term incrementally upgrades for exiting
PONs. Colorless-ONU WDM network is a promising
candidate for NG-PON and already found a rich market in
Korea. In addition, TDM over WDM provides a cost
effective solution for NG-PON. The reach distance of LR
PON topologies can simplif the network by combining the
metro and access segments. The combination of wire and
wireless networks can support future broadband services and
wideband mobiles on the same infastructure. Different
modulation schemes are extremely needed to increase the
performance of NG-PON in order to increase the bit rate
within the minimum bandwidth.
VII. ACKNOWLEDGMENT
This research is supported by The National Plan for
Science and Technology (NPST) program by King Saud
University, Saudi Arabia, Project Number 09-ELE855-02.
The authors acknowledge Prince Sultan Advanced
Technologies Research Institute (PSA TRI) for accessing its
facilities.
J^IL NII
A COMPARISON BETWEEN NG-PON ARCHITECTURES
1Ol VOl
Cost |ov High
Hitratc Lov High
Rcachdistancc |ov |ov
5pIittingratio High |ov
5ubscribcrsI 1orhighcr Lcssthan 1
backhauI!bcr
HybridVOl/1Ol
lcdium
lcdium
Lov
High
lorcthan 1Z8
High
Lov
High
High
~1housand
|iVi
High
|ov
|ov
|ov
LcssthanJZ
J44
III
[2)
[3)
[4)
[5)
"I
[7)
[8)
[9)
[10)
[1 I)
\\ lIlahk hii"
VIIl. REFERENCES
N. Ghazisaidi, et al., "Survivability Analysis of Next-Generation
Passive Optical Networks and Fiber-Wireless Access Networks,"
Reliabilit, IEEE Transactions on, vol. 60, pp. 479-492, 2011.
1. i. Kani, et al., "Next-generation PON-part I: Technology roadmap
and general requirements," Communications Magazine, IE vol. 47,
pp. 43-49, 2009.
L. Rujian, "Next Generation PON in Emerging Networks," in Optical
Fiber communicationNational Fiber Optic Engineers Conference,
2008. OFC/NFOEC 2008. Conference on, 2008, pp. 1-3.
K. Tanaka et al., "IEEE 802.3av 10G-EPON Standardization and Its
Research and Development Status," Lightave Technolog, Journal
of vol. 28, pp. 651-661, 2010.
\
K. Grobe and 1. P. Elbers, "PON in adolescence: from TDMA to
WDM-PON," Communications Magazine, IE vol. 46, pp. 26-34,
2008.
1. i. Kani, "Enabling Technologies for Future Scalable and Flexible
WDM-PON and WDM/TDM-PON Systems," Selected Topics in
Quantum Electronics, IEJoural of vol. 16, pp. 1290-1297,2010.
1. S. Lee, et al., "Spectrum-sliced fber amplifer light source for
multichannel WDM applications," Photonics Technolog Letters,
IE vol. 5, pp. 1458-1461, 1993.
K. lun-ichi and K. Iwatsuki, "A wavelength-tunable optical
transmitter using semiconductor optical amplifers and an optical
tunable flter for metro/access DWDM applications," Lightave
Technolog, Joural of vol. 23, pp. 1164-1169, 2005.
L. lie Hyun, et al., "First Commercial Deployment of a Colorless
Gigabit WDM/TDM Hybrid PON System Using Remote Protocol
Terminator," Lightwave Technolog, Journal of vol. 28, pp. 344-351,
2010.
[12)
[13)
[14)
[15)
[16)
[17)
[18)
[19)
[20)
D. P. Shea and 1. E. Mitchell, "Long-Reach Optical Access
Technologies," Network, IEEE, vol. 21, pp. 5-11,2007.
G. Talli and P. D. Townsend, "Hybrid DWDM-TDM long-reach PON
for next-generation optical access," Lightwave Technolog, Journal
of vol. 24, pp. 2827 -2834, 2006.
D. P. Shea and 1. E. Mitchell, "Experimental Upstream Demonstration
of a Long Reach Wavelength-Converting PON with DWDM
Backhaul," in Optical Fiber Communication and the National Fiber
Optic Engineers Conerence, 2007. OFC/NFOEC 2007. Conference
on, 2007, pp. 1-3.
H. Song, et al., "Multi-thread polling: a dynamic bandwidth
distribution scheme in long-reach PON," Selected Areas in
Communications, IEEE Journal on, vol. 27, pp. 134-142,2009.
N. Ghazisaidi, et al., "Fiber-wireless (FiWi) access networks: A
survey," Communications Magazine, IEEE, vol. 47, pp. 160-167,
2009.
N. Ghazisaidi and M. Maier, "Fiber-wireless (FiWi) access networks:
Challenges and opportunities," Network, IE vol. 25, pp. 36-42,
2011.
L. Wen-Piao, et al., "A reliable architecture for broad-band fber
wireless access networks," Photonics Technolog Letters, IE vol.
15, pp. 344-346,2003.
S. Bhandari and E. K. Park, "Hybrid Optical Wireless Networks," in
Networking, International Conerence on Systems and International
Conference on Mobile Communications and Learning Technologies,
2006. ICN/ICONS/MCL 2006. Interational Conference on, 2006, pp.
113-113.
S. Wei-Tao, et al., "Hybrid Architecture and Integrated Routing in a
Scalable Optical–Wireless Access Network," Lightwave
Technolog, Joural of vol. 25, pp. 3443-3451, 2007.
345