Next Generation Optical Access Networks (Invited Paper) Kyeong Soo Kim

David Gutierrez, Wei-Tao Shaw, Fu-Tai An, Yu-Li Hsueh, Matthew Rogge, Gordon Wong and Leonid G. Kazovsky

Advanced System Technology STMicroelectronics San Jose CA 95131, USA [email protected]

Photonics and Networking Research Laboratory Stanford University, Stanford CA 94305, USA {degm, wtshaw, kazovsky}@stanford.edu Abstract— The last mile continues to be a major bottleneck in the Internet. Its low bandwidth and flexibility prevents the deployment of new services and the development of new applications. In this paper we present a summary of current efforts in access networks research, focusing in particular on fiber optic solutions. We present the Stanford University aCCESS (SUCCESS) initiative within the Photonics & Networking Research Laboratory (PNRL). As part of this initiative, two novel network architectures have been developed, SUCCESS-HPON and SUCCESS-DWA, which propose a smooth migration path from current TDM-PONs to future higher bandwidth, cost-efficient, scalable WDM-PONs. In addition, we present SUCCESS-LCO, a spectral-shaping line coding technique that enables a cost-effective shorter-term capacity upgrade of existing TDM-PONs. We discuss as well what we believe are the main open research areas in optical access networks. Keywords - Access networks, PON, TDM-PON, WDM-PON.

I.

INTRODUCTION

Optical access network, including Fiber To The Home / Curb / Node (FTTX), has long been considered a definite solution to the problem of upgrading current congested access networks to ones capable of delivering future broadband integrated services. Optical fibers can provide bandwidths that other transmission media, such as wireless or copper, cannot. Because of their huge bandwidth capacity, optical fibers have already been used in backbone networks, Wide Area Networks (WANs), Metropolitan Area Networks (MANs) and even Local Area Networks (LANs) successfully. The use of fiber in access networks would be the last step needed to the future all-opticalnetwork revolution. Traditional optical network solutions based on point-topoint architectures are expensive for access configurations: besides fiber deployment costs, they need maintenance for the outside plant active systems. These systems consist of many Electrical-to-Optical (E/O) and Optical-to-Electrical (O/E) components prone to failure, which prevent their large-scale deployment. To address these issues and expedite the introduction of FTTX, Time Division Multiplexing Passive Optical Networks (TDM-PONs) have been developed. These TDM-PONs include Ethernet PON (EPON), Broadband PON (BPON) and Gigabit PON (GPON). These solutions are based on similar passive tree topologies, but have different data transport technologies. The TDM-PON architectures mentioned above can break through the economic barrier of traditional fiber optic point-toThis work was sponsored by the Stanford Network Research Center, STMicroelectronics, KDDI and Motorola.

point solutions [1]. They are internationally standardized by the IEEE and ITU, products following the standards have been developed, and are currently being deployed in the field by network service providers in several places around the world, mostly in Japan, Korea and the United States. In a few years, once TDM-PONs are deployed, upgrading these optical access networks will be a challenge when user demand outgrows the existing access network capacities. TDM-PONs have only one wavelength for downstream data and one for upstream data to share among all users, thus limiting the average bandwidth per user to a few tens of Mbit/s [2]. The very high available bandwidth of fiber is thus mostly wasted. Wavelength Division Multiplexing (WDM) technology has been considered an ideal solution to extend the capacity of TDM-PONs without drastically changing the fiber infrastructure. Many architectures incorporating WDM into access have been proposed by both academia and industry (e.g. [3] and [4]). However, how to migrate from TDM-PON to WDM-PON still requires further investigation. This is one of the main issues addressed in this paper. The rest of this paper is organized as follows. Section II describes some current TDM-PON and WDM-PON technologies. While TDM-PON is mostly standardized, WDM-PON is still mostly in its research phase. Sections III, IV and V describe three projects under the SUCCESS initiative for upgrading TDM-PONs: SUCCESS-HPON, SUCCESSDWA and SUCCESS-LCO. Section VI lists some of what we believe are the main research areas in access networks. Section VII concludes this paper. II.

TDM-PON AND WDM-PON TECHNOLOGIES

A. EPON Ethernet-based PONs (EPON) [5] currently offers today the highest average bandwidth for the end customer of any of the TDM-PON alternatives. EPON equipments can support up to 1.25 Gbps symmetrical line rates and 32 Optical Network Units (ONUs) without Forward Error Correction (FEC) or 64 ONUs with FEC per PON. Its maximal physical reach is 20 km. EPON uses the 1510nm wavelength to transmit downstream data and the 1310nm wavelength for upstream data. The 1530 - 1565nm window is reserved for transparent optical overlays. EPON uses an Advanced Encryption Standard (AES) based

encryption scheme in both upstream and downstream directions. EPON uses variable-length Ethernet frames to carry all digital data via the Internet Protocol (IP). Thus, it eliminates the need for more complex and usually more expensive ATM and SONET network elements that may be needed to interface with BPON equipment. In the view of Ethernet convergence advocates, this also dramatically simplifies the overall network architecture. Using the widely adapted, manageable, and flexible Ethernet technology also allow EPON service provider to bundle data, video, and voice service on the same infrastructure and provide a broad range of service offerings. The IEEE and the Ethernet in the First Mile (EFM) study group supported the initial development of the EPON standard. Its simplicity, low-cost, and ability to provide multiservice access solutions hugely contributes to its popularity, especially in Asia. B. BPON and GPON North American operators initiated the BPON Standards through the Full Service Access Network (FSAN) group in 1995. The results of their efforts were later standardized by the ITU in the G.983 recommendations [6]. The underlying transmission technology for BPON is ATM cell-based. Because of this, they were at first named ATM PON (APON). The name was later changed to Broadband PON (BPON) to emphasize that they were not limited to ATM traffic. The BPON standards specify the protocols from physical layer to OAM. The maximum speeds are 622 Mbit/s downstream and 155 Mbit/s upstream. In an effort to make transport more flexible in Layer 2, to increase nominal rate and enhance security, the ITU generated the Gigabit-capable PONs (GPON) standards [7]. GPON is capable of supporting the higher bandwidth requirements of business and residential services, by providing a nominal line rate of 2.4 Gbps downstream and 1.2 Gbps upstream; however, it can support up to 64 users, making its average bandwidth per user a bit lower than EPON. In addition to being able to transport ATM or Ethernet traffic at layer 2, the GPON Encapsulation Method (GEM) maps all traffic across the network using a variant of the SONET/SDH framing structure, which allows more efficient packaging and higher QoS for delay-sensitive traffic such as voice and video communications. In GPON, the physical reach of the PON is designed to fall within 20 km and each system typically support a 1:64 split ratio, but the split ratio is anticipated to be able to support up to 1:128 users in the near future with better optical modules. The standard also specifies encryption to prevent unauthorized snooping in the downstream traffic. The GPON standards are still in development, having the physical and transmission layers already well defined. Since the OAM and other higherlayers have not been defined yet in the standards, operators are reluctant to deploy them for now. Besides the 1490nm downstream and 1310nm upstream traffic, the standard reserves a window of the 1550nm spectrum for “enhancement” services. It might be used for video transmission or for WDM upgrade upon increasing bandwidth

demand. Currently in North America, network operators are using that window for radio frequency (RF) video services (CATV). There is a big push to transmit Internet Protocol Television (IPTV) signals; its deployment via the enhancement spectrum or regular downstream data is imminent. C. WDM-PONs WDM technologies can be used to dramatically increase the throughput of PONs. However, the cost of tunable and wavelength-sensitive WDM components is quite high for access networks. In access networks, the cost of the components is shared among a few tens of residential users, and not among thousands of business users, as is the case with long-haul networks. Thus, developing network architectures, sub-systems and devices that reduce the cost of WDM-PONs is crucial for their successful deployment. Fig. 1 illustrates an example of a WDM-PON. On the right, the Optical Line Terminal (OLT), which resides at the Central Office (CO), has an array of transmitters and receivers. Each transmitter-receiver pair is set at the wavelength band of the port of the multiplexing device, in this case an Arrayed Waveguide Grating (AWG), to which the pair is connected. In some cases, especially when the expected load of the network is low and bursty, the array of transmitters may be replaced by a set of fast tunable transmitters that are dynamically shifted from one wavelength to the next one. In this case, the number of tunable devices can be much less than the number of ONUs. ONU1

OLT

TX1

TX1

RX1

RX1 SMF

TX N

TX N AWG

RXN

AWG

RXN

ONUN

Figure 1. WDM-PON example.

After the OLT, a Single Mode Fiber (SMF) acts as a feeder fiber to a Remote Node (RN, not labeled) where another multiplexing device sits. Each port of the AWG at the RN is connected to a different ONU. Each ONU has a passive splitter that is connected on one end to a tunable transmitter and on the other to a receiver. In this particular architecture, the receiver needs not to be tunable, since the AWG will only allow the assigned wavelength to arrive to the ONU. The receivers may optionally be tunable, if there is a need to have splitters / couplers between the RN and the ONU (as is the case in many deployed TDM-PONs today). However, the transmitter needs to be tunable in this architecture if all ONUs are going to be the same, since each ONU must be able to communicate through whatever port of the AWG it is connected to. One of the key network elements that needs to be produced in large quantities at a low price is the ONUs and Optical Network Terminals (ONTs), which resides close or at the customer premises. Some alternatives for the design of WDMPON ONUs are tunable laser diodes, broadband light sources

with spectral slicing, Fabry-Pérot laser diodes with injectionlocking and Centralized light sources. Table I summarizes some of the key features of TDMPONs and WDM-PON. TABLE I.

PON TECHNOLOGIES TDM-PON

WDMPON

EPON

BPON

GPON

Standard

IEEE 802.3ah

ITU G.983

ITU G.984

Framing

Ethernet

ATM

GEM/ATM

1 Gbit/s

622 Mbit/s

2.488 Gbit/s

16

32

64

None None Specified 1-10 Gbit/s per channel 100’s

60 Mbit/s

20 Mbit/s

40 Mbit/s

1-10 Gbit/s

Maximum Bandwidth Users/PON Average Bandwidth per User Video Estimated Cost Upgradeability

RF / IP

RF

RF / IP

RF / IP

Lowest

Low

Medium

High

Difficult

Difficult

Difficult

Easy

III.

SUCCESS-HPON

The Stanford University aCCESS (SUCCESS) initiative within the Photonics & Networking Research Laboratory encompasses multiple projects in access networks. In this paper, we discuss three of these projects. SUCCESS-HPON and SUCCESS-DWA are two novel network architectures that propose a smooth migration path from TDM-PON to future higher bandwidth, cost-efficient, scalable WDM-PONs. SUCCESS-LCO is a spectral-shaping line coding technique that enables a smooth and cost-effective capacity shorter-term upgrade of existing TDM-PONs. λ 3, λ 4, …

λ3

Central Office

λ 1, λ 2

λ’3, λ 4, …

λ’1, λ 2

C

λ 41

λ4

λ1 λ’1

λ’3

λ 42

C

W

λ 43

λ2 W

TDM-PON ONU

λ 21

λ 22

C

TDM-PON RN

W

WDM-PON RN

WDM-PON ONU

λ 23

Figure 2. SUCCESS-HPON Architecture.

The SUCCESS Hybrid WDM/TDM PON or SUCCESSHPON [8] is a novel optical access network architecture. The two main goals in the design of this architecture are: (1) To provide a smooth migration from current TDM-PONs (i.e., EPON, BPON and GPON) to higher bandwidth WDM-PONs and (2) To make WDM-PON cost-efficient for access. The SUCCESS-HPON architecture proposes a migration path from

TDM to WDM-PON using a centralized light sources (CLS) approach and tunable WDM component sharing for costefficiency. A. Architecture Overview The overall architecture of SUCESS-HPON, including TDM-PONs and WDM-PONs as its subnetworks, is shown in Fig. 2. A single-fiber collector ring with stars attached to it formulates the basic topology. The collector ring strings up Remote Nodes (RNs), which are the centers of the stars. The ONUs attached to the RN on west side of the ring talk and listen to the transceiver on the west side of OLT, and likewise for the ONU attached to the RNs on the east side of the ring. There is a point-to-point WDM connection between the OLT and each RN. No wavelength is reused on the collector ring. Optional semi-passive RNs may be used to sense fiber cuts and flip the orientation, if extra reliability is desired. Fig. 3(a) shows the logical block diagram for the WDMPON portion of the SUCCESS-HPON OLT. Tunable components, such as fast tunable lasers and tunable filters are employed. Since the average load of access networks is generally low, using tunable components minimizes transceiver count and thus minimizes total system cost. This arrangement is also good in terms of scalability: as more users join the network, or their traffic increases, more tunable lasers and receivers are added at the OLT. Upstream optical signals are separated from the downstream signals by circulators. The scheduler controls the operation of both tunable transmitters and tunable receivers. Note that the tunable transmitters at the OLT generate both downstream frames and CW optical bursts to be modulated by ONU for upstream data. With this configuration, half duplex communication is possible at the physical layer between each ONU and the OLT. Compared to a similar architecture [4] with a two-fiber ring, two sets of light sources, and two sets of MUX/DEMUX to perform full-duplex communications, the SUCCESS-HPON architecture dramatically lowers costs. As a tradeoff, it needs carefully designed MAC protocol and scheduling algorithms to provide efficient bidirectional communication. Fig. 3(b) illustrates the RN design for the TDM-PON RNs (marked "C" in Fig. 2) and Fig. 3(c) illustrates the design of the WDM-PON RNs (marked "W" in Fig. 2). A TDM-PON RN has a pair of CWDM band splitters to add and drop wavelengths for upstream and downstream transmissions, respectively. A WDM-PON RN has one CWDM band splitter, adding and dropping a group of DWDM wavelengths within a CWDM grid, and a DWDM MUX/DEMUX device such as an Arrayed Waveguide Grating (AWG). Each WDM ONU has its own dedicated wavelength for both upstream and downstream transmissions on a DWDM grid to communicate with the OLT. Since the insertion loss of an AWG is roughly 6 dB regardless of the number of ports, one with more than eight ports can be used to enjoy better power budget compared to a passive splitter. Each RN generally links sixteen to sixty four WDMPON ONUs. Fig 3(d) illustrates the logical block diagram for the WDMPON ONU. The ONU has no local optical source and uses

instead an optical modulator to modulate optical CW burst received from OLT for its upstream transmission. A Reflective Semiconductor Optical Amplifier (RSOA) can be used as an amplifier/modulator for this purpose with the assumption that its integration with electronics would decrease its production costs when mass-produced.

strings the RNs served by this CO. Note that distribution fibers are untouched during this migration. From the ONUs point of view, the functionality of the optical access network is exactly the same; only a short downtime for upgrade is needed. Therefore, existing TDM-PON ONUs can virtually work the same as before without a major upgrade.

Note that the ONU does not need a tunable receiver. The WDM-PON RN removes extraneous wavelengths and allows only a specific wavelength to reach each WDM-PON ONU. The receiver at the ONU just needs to have enough optical spectral bandwidth to receive any DWDM channel used in the network. Note as well that the MAC block in the ONU not only controls the switching between upstream and downstream transmissions but also coordinates with the scheduler at the OLT through polling and reporting mechanisms. For further implementation details of the SUCCESS-HPON, especially at the physical layer, readers are referred to [8].

Fig. 4(c) describes the second phase of migration. As more users demand high-bandwidth for future broadband applications, WDM-PON RNs are inserted in the network. In this case, as explained above, there is a dedicated DWDM channel between each ONU and the OLT. If protection and restoration functionality is implemented in the existing RNs using semi-passive switches, inserting a new RN in the network will not disturb the network operation in general. Fig. 4(d) shows the possible extension of the network. Since there is a dedicated wavelength at the output of the AWG, it is possible to use the collector ring as a backhaul for the PON with tree topology. The two feeder fibers of the PON can connect to different RNs to form a protection path. To upgrade the capacity even further, the RSOA modulator can be replaced by a stabilized laser source to perform full-duplex operation.

Tunable Transmitter 1

M:2 Passive Splitter

...

...

Downstream Traffic Queues

...

Tunable Transmitter M

Upstream Grant Queues

Ring - East Circulator

Scheduler

Ring - West

CO

CO

...

Tunable Receiver 1

Upstream Traffic

Tunable Receiver N

Passive splitter

N:2 Passive Splitter

C

W W

C

(a) CWDM RN Upstream Add 1310 nm

C

Co-existing CWDM/DWDM Optical Access Networks

W W DWDM, AWG

DWDM RN (a)

Downstream Drop e.g. 1498.96 nm

(c)

λ CO

C

NxN AWG

2×N

CO

C

Old ONUs and distributor fibers are preserved.

C

W W

C

C

Cost efficient, scalable, protected WDM/TDM PON

C

W 2×N

C CWDM, splitter

...

TDMTDM- PON ONUs

WDMWDM-PON ONUs

(b)

To DWDM Remote Node

(c) Upstream Traffic Queue RSOA

PD

MAC

BM Receiver

(b)

(d)

Figure 4. SUCCESS-HPON migration from TDM to WDM/TDM.

Downstream Traffic

(d)

Figure 3. SUCCESS-HPON components.

B. SUCCESS-HPON Migration from TDM-PON to Hybrid WDM/TDM-PON Fig. 4 illustrates the network migration scenario of optical access networks based on the SUCESS-HPON architecture. Fig. 4(a) shows the existing tree TDM-PONs connected to the same CO. Each TDM-PON has its own cabling and OLT inside CO. Fig. 4(b) shows the first migration step of the existing network infrastructure. The passive couplers of the PONs are replaced with the TDM-PON RNs described before. The feeder fibers of PON are replaced with a single fiber ring that

C. MAC Protocol and Scheduling Algorithms In the SUCCESS-HPON architecture, all tunable transmitters and receivers are located at the OLT. The sharing of these tunable components to service all the ONUs, and their use for both upstream and downstream data transmission pose a great challenge in designing an adequate scheduling algorithm. Such an algorithm has to keep track of the status of all shared resources through time (i.e., tunable transmitters and tunable receivers), ONU wavelength assignments and round trip times and arrange them properly in both the time and wavelength domains for both downstream and upstream data transmissions. Extensive work has been done on developing efficient scheduling algorithms for SUCCESS-HPON, as described in [8, 9, 10 and 11]. Here we briefly describe the best scheduling algorithm thus far, Sequential Scheduling with Schedule-Time Framing (S3F). This algorithm has relatively low computational complexity and can provide high throughput and low delay.

period, and (2) a higher computational complexity and memory is needed, which may affect the implementability of the algorithm.

Total throughput [Gbps]

We consider a SUCCESS-HPON WDM-PON system with W ONUs (therefore W wavelengths), M tunable transmitters, and N tunable receivers. Because the tunable transmitters are used for both upstream and downstream traffic but the tunable receivers are only used for upstream traffic, it will usually be the case that W ≥ M ≥ N. A guard band (tens of ns) between consecutive frames takes into account the effects of unstable local ONU clock frequencies and tuning times of tunable transmitters and receivers at the OLT. Like in B/G/EPON systems, the SUCESS-HPON OLT polls to check the amount of upstream traffic stored inside ONUs and sends grants (with appended optical CW bursts) to allow the ONUs to transmit upstream traffic. Since there is neither a separate control channel nor a control message embedding scheme using escape sequences as in [12], the SUCCESS-HPON WDM-PON MAC protocol employs in-band signaling and uses frame formats with Report and Grant fields defined for polling and granting, respectively. A Frame Type field in the downstream frame header indicates whether the frame is to be used for upstream data or not. If it is going to be used for upstream data, the frame has appended a CW long enough for the granted data to be modulated onto it.

Total network load [Gbps] Figure 5. SUCCESS-HPON throughput simulation results.

3

The S F algorithm operates as follows. At the end of each SUCCESS-HPON frame transmission or when a payload frame (e.g., Ethernet) arrives at an empty virtual output queue (VOQ): 1.

Select the earliest available transmitter and receiver.

2.

Taking into account the transmitter and channel availability times, schedule the next transmission time for data in this VOQ; if the VOQ is for upstream grants, take into account the receiver availability and ONU round trip times as well.

3.

Encapsulate queued data (e.g., Ethernet frames) into a SUCCESS-HPON frame; the maximum size of the latter is determined by the VOQs downstream transmission counter.

4.

Update the status variables for transmitter, receiver and channel availabilities and the downstream traffic counter.

The downstream transmission counter in Step 3 ensures fairness among downstream and upstream traffic. The counter is increased based on the corresponding VOQ length when upstream traffic requests arrive and decreased as downstream traffic leaves. It is ensured, though, that downstream traffic can be sent even in the absence of upstream traffic. This counter, and the fact that there are separate VOQs for each ONU in each direction (downstream and upstream) ensures fairness between downstream and upstream traffic. Fig. 5 illustrates the simulation results for the S3F algorithm. For a more detailed description of these and previous algorithms (Sequential and BEDF), please see [10]. Future work regarding the MAC protocol and scheduling algorithm will deal with the important issue of weighted fairness among ONUs and the tradeoffs of batch mode scheduling. In this mode, every batch period all the frames that can be scheduled are considered simultaneously, which allows for some optimization. Two tradeoffs are of importance in this case: (1) better throughput might be achieved, but at the expense of possibly higher average delays due to the batch

D. Experimental Results

(a)

(d) 800 ps

(b)

(e)

(c) Figure 6. SUCCESS-HPON Experimental results.

The SUCCESS-HPON testbed was developed to: (1) demonstrate the feasibility of bi-directional transmission of upstream and downstream traffic on the same wavelength for access networks, (2) demonstrate the possibility to modulate upstream data onto CWs provided by the OLTs, (3) demonstrate the functionality of the MAC protocol and scheduling algorithms and (4) explore possible SUCCESSHPON implementation issues. In this paper we describe the first version of the SUCCESSHPON testbed and the experimental results obtained, mostly

related to objectives (1) and (2) above. We are currently working on a second version of the testbed, in which several subsystems will be improved and where objectives (3) and (4) above will be addressed. For a detailed explanation of the SUCCESS-HPON testbed, readers are referred to [13]. In this paper we limit our discussion to the experimental physical layer results shown in Fig. 6 for downstream and upstream modulation and transmission. Fig. 6(a) is the downstream data and continuous wave sent on λ1 to ONU1; Fig. 6(b) is the downstream data and continuous wave sent on λ2 to ONU2; Fig. 6(c) is the upstream data on λ1 and λ2 that the OLT receives. Fig. 6(d) is the downstream eye diagram, while Fig. 6(e) corresponds to the upstream eye diagram at a 1.25 Gbit/s rate. In conclusion, the SUCCESS-HPON architecture provides a smooth migration path from current TDM-PONs to future WDM-PONs. Simulation results for the scheduling algorithms, and experimental results at the physical link layer validate our approach. IV.

SUCCESS-DWA

SUCCESS-DWA PON is another novel optical access network architecture designed for the next-generation access networks, which employs Dynamic Wavelength Allocation (DWA) [14, 15, 16, 17 and 18]. As is shown later, the scalability of the SUCCESS-DWA PON allows the network to easily bridge the large gap between TDM and WDM-PONs. In addition, the architecture provides excellent cost efficiency and network performance by sharing bandwidth across multiple physical PONs. Existing arbitrary field-deployed PON infrastructures remain intact when brought together into a SUCCESS-DWA PON. Central Office

Figure 7. SUCCESS-DWA downstream architecture.

A. Architecture of SUCCESS-DWA PON for downstream transmission Fig. 7 shows the architecture of SUCCESS-DWA PON. Fast tunable lasers (TLs), cyclic AWG, and thin-film WDM filters constitute the key components of the system. The TLs and AWG reside in the CO, while the WDM filters are within the ONU. Note that the field-deployed infrastructure is compatible with TDM-PONs. In the basic architecture, the cyclic AWG multiplexes the TLs and routes the TL outputs to

different physical PONs depending on the wavelength. Each ONU within a single PON contains a unique fixed-wavelength filter and a burst-mode receiver. The key lies in the fact that the passband of the ONU filter encompasses the free spectral range of the AWG. For example, 200 GHz ONU filters would work with a 4 X 4 cyclic 50 GHz AWG. The relative filter shapes are illustrated in Fig. 8. In this architecture, any tunable laser can individually address any ONU across separate physical PONs at any given time [17]. ONU WDM Channel 1 1

2

3

4

ONU WDM Channel 2 5

6

7

8

ONU WDM Channel 16 .....

61

62

63

64 λ

OLT AWG Channels

Figure 8. Wavelength bands for AWG channels and thing-film WDM filter.

For each downstream frame, the TLs tune to the appropriate wavelengths and transmit the data to the corresponding end users. The transmission durations for end users are globally managed to achieve optimal network performance. The traffic scheduling allows free moving of bandwidth among physical PONs, which is very useful to accommodate bursty Internet traffic. All TLs share the load, shifting bandwidth across the separate physical PONs as necessary — we call this technique Dynamic Wavelength Allocation or DWA. To illustrate the flexibility of this architecture, compare an initial deployment of four TDM-PONs to a SUCCESS- DWA PON that spans four physical PONs. In the initial deployment stage, the first several subscribers would likely scatter across multiple PONs, and in the worst case they may be spread across all four PONs. In the four TDM-PON cases, then, all four OLTs (lasers) must be activated, despite the fact that some OLTs may only be serving few subscribers. With the SUCCESS-DWA PON, on the other hand, it is sufficient to deploy only one TL and AWG in the CO, and the subscribers across multiple PONs are all serviced by the single TL. As demand grows, additional TLs can be added to the AWG. When the subscription rate is high enough, the two scenarios seem to converge — both have four transmitters serving all subscribers. However, the SUCCESS-DWA PON enjoys the benefit of statistical multiplexing over a larger customer base, so its performance will exceed that of the four TDM-PONs. For the performance of statistical multiplexing gain, readers are referred to [15] and [17]. When demand dictates, the SUCCESS-DWA PON can be scaled far beyond the conventional TDM-PONs. The grayedout AWG in Fig. 7 illustrates the concept. By connecting some of the physical PONs to additional AWG + TLs sets, the architecture shifts from four TLs serving four PONs to four TLs serving each a single PON. If 8x8 AWGs are utilized, sixteen end users can be served by eight TLs, which results in a very high-performance network close to a full WDM-PON. The excellent scalability provides a smooth and graceful upgrade from a TDM-PON towards a WDM-PON. The stepby-step system upgrade easily tracks user demands, and the initial overhead can be even lower than that of conventional TDM-PONs. In addition, the TLs provide protection for each

other, maintaining service to all physical PONs in the event of a failure. B. Architecture of SUCCESS-DWA PON for upstream transmission In [18] it was reported that the measured upstream traffic rates were highly related to the downstream traffic rates in access networks. Users enjoying high-speed downloads also expect a commensurate data rate in the upstream. This suggests that a high-performance upstream scheme is desirable. Depending on the performance requirements, there is a wide range of possible scenarios for upstream transmission. In [17] we investigated 5 different upstream schemes for SUCCESSDWA PON and evaluated them in terms of cost, performance, and scalability. These schemes are summarized in Fig. 9. ONU

(A)

LD λ ONU

(B)

Some users are equipped with TLs and the others are equipped with a single FP

TL λ ONU

(C)

Some users are equipped with dual FPs at different λs, and the rest are equipped with a single FP

λm λn

λ

wavelengths and require different channel spacings. Hence separate AWGs are necessary for up/downstream. Similar to the downstream, only one PD + receiver module is activated in the initial deployment. To cover all ONUs residing on different physical PONs, the first several subscribers from different PONs are assigned specific upstream wavelengths, i.e., the first several subscribers from PON1 are assigned to λ1, the first several subscribers from PON2 are assigned λ4, and so on. When the number of users increases, a second PD + receiver module can be installed in the OLT. For new subscribers to be served by the second PD, they are assigned λ2 on PON1, λ1 on PON2, and so on. Similarly, the upstream architecture can further scale up as the downstream. When demand dictates, additional AWG + PD/RX sets can be added (illustrated by the grayed-out AWG in Fig. 4) to shift from four PDs serving four PONs to four PDs serving each PON. For the bandwidthdemanding users, tunable devices can replace the fixedwavelength transmitters, as in Schemes B and C. However, unlike the downstream architecture that desires fast TLs (~10 ns tuning times) to achieve DWA on the order of µs, the upstream DWA schemes perform slower wavelength reallocations (in the order of ms) due to the more involved communications between the OLT and ONUs for transmission scheduling.

OLT

(D)

DeMux λ

Tunable DeMux in OLT (Consecutive channels)

OLT

(E)

DeMux

λ

Tunable DeMux in OLT (Arbitrary channels)

Figure 9. Five upstream schemes for SUCCESS-DWA.

For scheme A, users within a group transmit on the same upstream wavelength and time-share the same PD in the OLT. It is static in wavelength allocation and provides a baseline for comparison with the other schemes. Schemes B and C are distributed DWA schemes, in which the tunability is spread across the ONUs. The distributed schemes allow more flexibility since the deployment of tunable devices can be judged by user demands. On the other hand, Schemes D and E are centralized DWA schemes, in which the tunability resides in the OLT. The description of the tunable DeMux can be found in [17]. Qualitatively, full tunability results in the best performance. Scheme E exhibits full-tunability, but at the cost of an expensive, centralized tunable DeMux. Scheme C can provide equal performance if all users are equipped with TLs. While the centralized schemes subject all ONUs to the high cost of the tunable device, the distributed schemes require only those ONUs which demand high performance to be upgraded. Therefore, the added design flexibility of distributed schemes B and C makes them more preferable. A complete upstream SUCCESS-DWA PON architecture is shown in Fig. 10. Four physical PONs are connected to the OLT, and an AWG functions as the DeMux for routing the incoming wavelengths to the corresponding PDs. It is worth noting that the AWG does not require cyclic features, and the upstream and downstream AWGs pass completely different

Figure 10. SUCCESS-DWA upstream architecture for users across multiple PONs. The grayed-out PDs and AWG demonstrate the system upgrade path.

C. Traffic scheduling and QoS Support in SUCCESS-DWA PON To support QoS on SUCCESS-DWA PON, strict-priority traffic scheduling algorithms were investigated in [15]. In [16] the traffic is categorized into high-priority (HP) and best effort (BE) traffic. Specifically, to improve the scalability and reduce the complexity of the buffering strategy in the OLT, we propose the TL-buffering scheme for the OLT of the SUCCESS-DWA PON. The block diagram of the TL-buffering scheme is illustrated in Fig. 11. The incoming traffic is divided and stored in separate HP and BE buffers before scheduling. Depending on the packet arrival time and priority, and the availability of the TLs, the HP traffic can be serviced at the earliest possible time and the prioritized traffic is highly differentiated. The performance of the TL-buffering scheme is

evaluated in comparison with conventional virtue-outputqueuing (VOQ) scheme, and the results show that the average latency and jitter of HP traffic outperforms that of the VOQ buffering strategy [16]. HP Buffer

TL1 Buffer

TL1

TL2 Buffer

TL2

BE Buffer

PON2 AWG

Scheduler Traffic Stripper

PON1

TL3 Buffer

TL3

PON3

TL4 Buffer

TL4

PON4

Figure 11. Block diagram of the TL-buffered scheme in the OLT. DWA scheduling algorithm supporting QoS Modulation Data FPGA PCB

TL1 MZ1

WDM traffic flow WDM Filter AWG

TL2 MZ2

OLT

equipment to keep operating while only the overloaded nodes/equipment are upgraded to higher capacity. It is also possible that a service provider intends to jointly service a deployed network infrastructure that is already operated by another service provider. The new service launched by the new provider should not affect the existing service and subscribers. In either case, it is preferable that all necessary changes occur only in the CO or in the newly deployed ONUs but not in the existing ONUs to minimize intrusions to subscribers. Consider the network architecture in Fig. 13, in which a new OLT and new ONUs share the same infrastructure with the existing equipments. Compared to the upstream service upgrade which can be done by simply assigning different wavelengths for the new and existing ONUs, it is not trivial to smoothly upgrade the downstream service. In conventional TDM PON, existing ONUs typically have a broadband transmit/receive optical filter (1.3µm /1.5µm), as opposed to a narrowband wavelength selective filter. Therefore any enhancements must take into account that existing ONUs will see all wavelengths that make it onto their leg of the fiber infrastructure.

Streaming Video

Existing ONU Existing OLT

WDM Filter

PD CDR FL

Existing ONU

FPGA PCB

AMP

ONU

New OLT New ONU

Figure 12. SUCCESS-DWA experimental testbed. Central Office

D. Experimental Testbed of SUCCESS-DWA PON In [18] we described the details of design and implementation issues of the key building blocks of SUCCESS-DWA PON, including fast tunable lasers, burstmode receivers, and scheduling algorithms with QoS support. A testbed is constructed and video streaming experiments in the downstream are performed to show the feasibility of the integrated system. As shown in Fig. 12, the experimental testbed consists of two fast TLs coupled by an AWG. Each TL is externally modulated at 1.25 Gb/s by a Mach-Zehnder modulator. Four user wavelength channels are defined, and 16 subscribers share bandwidth in both wavelength and time domains. The WDM filter in the OLT separates the downstream and upstream traffic occupying different wavelength bands. The WDM filter in the ONU not only separates the up/downstream traffic but also provides the necessary filtering corresponding to the ONU’s user wavelength channel. Burst mode level recovery and CDR are implemented in both the OLT and ONUs. In the ONU, a laser driven by a DC-coupled amplifier realizes a burst mode transmitter. Processors are necessary to perform the framing procedure, line coding, and scheduling algorithm. We designed PCBs with field programmable gate arrays (FPGAs) as processors in OLT and ONUs. The average packet latency and jitter of a realistic SUCCESS-DWA PON system are also measured on the testbed [18]. V.

SUCCESS-LCO

In a TDM PON system, when demand dictates, equipment must be upgraded to allow more bandwidth for some or all the end users. A smooth upgrade process allows the existing

Existing infrastructure

Subscribers

Figure 13. Upgrading existing TDM-PON by using an overlay.

To enable a smooth upgrade and service co-existence, three techniques were discussed in [19], including: (1) launch the new services in new wavelengths and retrofit the existing equipment and/or infrastructure. (2) Sub-carrier multiplexing (SCM) technique can be used to move the spectrum of the new service out of the baseband of existing ONU. (3) Using spectral line coding to shape the spectrum of the new service so that its interference to the existing users is minimized [19, 20 and 21]. Compared to the spectral line coding technique, the first two methods is less favorable due to the cost of hardware replacement and optical/electrical components. In principle, using spectral line code enables a graceful upgrade path without replacing existing PON equipment. During the upgrade period, only end users demanding enhanced services are required to replace their equipment. Line codes are flexible and can be reconfigured whenever necessary. Therefore we focused on investigating the spectral line code technique to allow coexistence of the existing lower bit-rate signal and a higher bitrate overlay signal. A. System Architecture of SUCCESS-LCO and the algorithm of spectral line code In an effort to capitalize on the advantages of both SCM and WDM, we developed novel line codes that suppress low frequency components while maintaining DC balance [19]. The goal is to push the frequency content of the new high bit-rate stream out of the baseband of the existing signal with minimal overhead. Fig. 14 illustrates the spectral-shaping line codes assisted with WDM to provide a graceful upgrade for PONs.

The existing OLT sends its traffic in nonreturn-to-zero (NRZ) format at a certain downstream wavelength. The new OLT, on the other hand, transmits its traffic at another wavelength, and the spectral power in the baseband region is purposely avoided. On the ONU side, the high-bit-rate signal from the new OLT is rejected by the LPF in the existing ONUs, while the signal from the existing OLT is filtered out by the WDM filter in the new ONUs. One clear advantage of this scheme is that the existing OLT, ONUs, and fiber plant remain unchanged. As subscribers upgrade their ONUs, they become members of the high-bit-rate PON. After all subscribers upgrade their ONUs, the old OLT is removed, and the new OLT can be reconfigured to maximize the data throughput. Received electrical signal.

Baseband transmission at a certain wavelength.

Pelect

Popt

f ONUOLD P elect

λ

OLTOLD

PD

LPF

f

Received signal.

OLTNEW

WDM

Popt

PD: PD: Photo PhotoDetector Detector LPF: LPF:Low LowPass PassFilter Filter WDM: WDM:WDM WDMfilter filter

Popt Coded signal at λ another wavelength.

PD

LPF

Pelect

ONUNEW Coexistence of signals.

λ

f Received signal.

hardware. Under the same hardware constraint, one can easily tune the data rate by reconfiguring the line code in firmware or software. In all cases, the 1-GBd/s baseband signal source emulating existing service is a bit-error-rate-tester (BERT) BERT1, which modulates the distributed-feedback (DFB) laser DFB1. The interferer or new service is realized by programming different line codes in BERT2, which modulates DFB2 at 10 GBd/s—the actual data rate depends on the code rate. DFB1 and DFB2 operate at DWDM wavelengths 1550.9 and 1551.7 nm, respectively. In the first test, a pseudorandom bit sequence (PRBS 223 − 1) is modulated onto the interferer in NRZ format. Since this interferer is not coded at all, it has a code rate of 1, but the spectral overlap in the baseband of the existing ONU is unacceptable. This scenario results in the worst performance. The other three codes include SUCCESSLC68, SUCCESSLC58, and Manchester coding, with code rates of 6/8, 5/8, and 1/2, respectively. The measurement results of BER v.s. signal-to-interference ratio (SIR) is shown in [19]. It shows that for a specific tolerable interference level, the optimal line code can easily be determined, which maximizes the data throughput. The service overlay by the linecoding technique provides a very flexible and elegant approach to smoothly upgrade existing PONs. 1 Gbaud/sec BERT1

Dn – raw data

1550.9nm DFB1 MZ Attn

WDM Filter amp

Coupler Optical

(20km SMF)

PD

LPF

Sn – coded signal Processing Window

: xn – inserted bits

PD

DFB2 MZ Attn 1551.7nm BERT2

Figure 14. Coding hierarchy consisting of the processing window, cells, data bits, and inserted bits.

The basic idea of the proposed line-coding technique is periodically inserting extra bits to tailor the spectral shape of the coded signal. The data stream is divided into fixed length processing windows of L bits (e.g., L = 256), which consist of small cells of C bits (e.g., C = 8). Fig. 14 shows the coding hierarchy. A long processing window ensures long-term spectral properties, while small cells simplify bit processing. The processing windows can be partially overlapped, but the difference between overlapping and non-overlapping windows is insignificant for reasonable window sizes. A cell contains data bits from the raw data stream and the inserted bits whose values are to be determined. Data bits and inserted bits are located in predefined positions in each cell. Decoding on the receiver side is straightforward—simply extracting data bits from the predefined positions in each cell. During network power up or upon system reconfiguration, cell-level synchronization and signaling are necessary for the receiver to acquire correct positions of data bits in the bit stream. The details about the algorithm and analysis of the spectral line code can be found in [19]. B. Experimental setup and measurement of SUCCESS- LCO Fig. 15 depicts the experimental setup used to compare four different overlay schemes, each with different efficiency and spectral overlap. We choose to compare codes at the same baud rate (10 GBd/s) because the baud rate is determined by the

10 Gbaud/sec

amp

LPF

Electrical DFB: DFB Laser MZ: Mach-Zehnder Modulator BERT: Bit Error Rate Tester Attn: optical attenuator PD: Photo Diode LPF: Low Pass Filter amp: RF amplifier

Figure 15. Experimental setup of SUCCESS-LCO. The new service is transmitted in 10 GBaud/s, and the existing is transmitting at 1 GBaud/s.

VI.

FURTHER RESEARCH ISSUES

In FTTX PONs, issues similar to those that arise in other networks, such as QoS, fairness, security and reliability, are present. Beyond these issues, however, there are some that are particular to access networks. TDM-PONs. The EPON standards specify only the network’s physical layer; the algorithms and protocols used to assign bandwidth to the users, ensuring fairness and QoS, are left open to the implementer. Considerable work has gone into developing Dynamic Bandwidth Allocation algorithms for this. A separate issue is how to provide open access for multiple ISPs to share the same fiber infrastructure when the local legislation requires it. Large-scale IP Video Networks. As discussed above, some service providers plan to deploy large-scale IP video networks to provide HDTV and VoD services. Even though this has been done in limited-size trials, the concept of a national IP video network is new and challenging. Some have suggested that the ideal situation is a combination of analog and IP video (this also depends on legislation and copyright issues). Some of the technical difficulties are the deployment of scalable multicasting protocols, mapping multicast groups to

conventional video channels, the development of inexpensive integrated set-top boxes and the ability to change video channels quickly. Integrated ONU / Wireless Base Station / Home Gateway / DSLAM. The integration of various components provides an opportunity to develop combined scheduling algorithms and MAC protocols that take into account multiple interfaces and users to optimize performance. Research on architectures, protocols and algorithms for such an integrated unit will be very valuable. Hybrid TDM/WDM-PON Architectures. To provide a smooth transition from TDM to WDM, the proposed network architectures will most likely need to efficiently share expensive tunable components. In the case of the SUCCESS architectures proposed, by sharing the most expensive components, the overall cost of the network decreased; however, this increased the computational complexity of the system. This tradeoff between number of components and computational complexity is a matter of further research. WDM-PON. Research that lowers the cost of devices that are passive and WDM functional (e.g., athermal AWGs) and that facilitates the integration of optics with electronics (e.g., SOAs in an ONU) is very important. As in the case of Hybrid TDM/WDM PONs, how to select which devices to share and how to share them with fast scheduling algorithms is also a matter of further research. VII. CONCLUSION In this paper we have presented the major projects carried out under the SUCCESS initiative at the Stanford University Photonics & Networking Research Laboratory (PNRL). As part of this initiative, two novel network architectures have been developed, SUCCESS-HPON and SUCCESS-DWA, which propose a smooth migration path from current TDMPONs to future higher bandwidth, cost-efficient, and scalable WDM-PONs. In addition, we presented SUCCESS-LCO, a spectral-shaping line coding technique that enables a costeffective shorter term capacity upgrade of existing TDM PONs. There are still many open areas of research in optical access networks, including TDM to WDM migration, video transmission, and integration with wireless and hybrid networks.

[5] [6] [7] [8]

[9]

[10]

[11]

[12]

[13]

[14]

[15]

[16]

[17]

[18]

[19]

REFERENCES [1] [2]

[3]

[4]

Kyeong Soo Kim, "On the evolution of PON-based FTTH solutions," (Invited Paper) Information Sciences, vol. 149/1-2, pp. 21-30, Jan. 2003. David Gutierrez, Kyeong Soo Kim, Salvatore Rotolo, Fu-Tai An, and Leonid G. Kazovsky, "FTTH standards, deployments and research issues," (Invited paper) Proc. of JCIS 2005, Salt Lake City, UT, USA, pp. 1358-1361, Jul. 2005. N.M. Froberg, S.R. Henion, H.G. Rao, B.K. Hazzard, S. Parikh, B.R. Rornkey and M. Kuznetsov, "The NGI ONRAMP test bed: reconfigurable WDM technology for next generation regional access networks,'' IEEE Journal of Lightwave Technology, Vol. 18/12, pp/ 1697-1708, Dec. 2000. J. Kani, M. Teshima, K. Akimoto, N. Takachio, H. Suzuki, and K. Iwatsuki, "A WDM-based optical access network for wide-area gigabit access services,'' IEEE Optical Communications Magazine, vol. 41/2, pp. S43 - S48, Feb. 2003.

[20]

[21]

IEEE Standard 802.3ah. ITU G.983 Recommendations. ITU G.984 Recommendations. Fu-Tai An, Kyeong Soo Kim, David Gutierrez, Scott Yam, Eric Hu, Kapil Shrikhande, and Leonid G. Kazovsky, "SUCCESS: A nextgeneration hybrid WDM/TDM optical access network architecture," IEEE/OSA Journal of Lightwave Technology, vol. 22, no. 11, pp. 25572569, Nov. 2004. Kyeong Soo Kim, David Gutierrez, Fu-Tai An, and Leonid G. Kazovsky, "Batch scheduling algorithm for SUCCESS WDM-PON," Proc. of GLOBECOM 2004, Dallas, TX, USA, Nov. 2004. Kyeong Soo Kim, David Gutierrez, Fu-Tai An, and Leonid G. Kazovsky, "Design and performance analysis of scheduling algorithms for WDM-PON under SUCCESS-HPON architecture," IEEE/OSA Journal of Lightwave Technology, vol. 23, no. 11, pp. 3716-3731, Nov. 2005. Jung Woo Lee, David Gutierrez, Kyeong Soo Kim, and Leonid G. Kazovsky, "Achieving 100% throughput in WDM-PON under the SUCCESS-HPON architecture," Proceedings of GLOBECOM 2006, Nov. 2006. G. Kramer, B. Mukherjee, and G. Pesavento, "IPACT a dynamic protocol for an Ethernet PON (EPON),'' IEEE Communications Magazine, vol. 40/2, pp. 74 - 80, Feb. 2002. Fu-Tai An, David Gutierrez, Kyeong Soo Kim, Jung Woo Lee, and Leonid G. Kazovsky, "SUCCESS-HPON: A next-generation optical access architecture for smooth migration from TDM-PON to WDMPON," IEEE Communications Magazine - Optical Communications Supplement Special Issue on Optical Networking Testbeds (Part 2), vol. 43, no. 11, pp. S40-S47, Nov. 2005. M. S. Rogge, Y.-L. Hsueh and L. G. Kazovsky, "A novel Passive Optical Network with Dynamic Wavelength Allocation," Optical Fiber Communication Conference (OFC 2004), Los Angeles, CA, p.FG1, February 2004. Y.-L. Hsueh, M. S. Rogge, W.-T. Shaw, S. Yamamoto and L. G. Kazovsky, "Quality of Service Support over SUCCESS-DWA: A Highly Evolutional and Cost-Effective Optical Access Network," Optical Fiber Communication Conference OFC 2005, Anaheim, CA, p.OThG4, March 2005. Y.-L. Hsueh, M. S. Rogge, W.-T. Shaw, L. G. Kazovsky and S. Yamamoto, "SUCCESS-DWA: A Highly Scalable and Cost-Effective Optical Access Network," IEEE Communications Magazine, 42, No. 8, pp. S24-S30, August, 2004. Y.-L. Hsueh, M. S. Rogge, S. Yamamoto and L. G. Kazovsky, "A Highly Flexible and Efficient Passive Optical Network Employing Dynamic Wavelength Allocation," Journal of Lightwave Technology, 23, No. 1, pp. 277-286, January, 2005. Y.-L. Hsueh, W.-T. Shaw, L. G. Kazovsky, A. Agata and S. Yamamoto, "SUCCESS PON demonstrator: experimental exploration of nextgeneration optical access networks," IEEE Communications Magazine, 43, No. 8, pp. S26-S33, August, 2005. Y.-L. Hsueh, M. S. Rogge, W.-T. Shaw, J. Kim, S. Yamamoto and L. G. Kazovsky, "Smooth upgrade of existing passive optical networks with spectral-shaping line coding service overlay," Journal of Lightwave Technology, 23, No. 9, pp. 2629-2637, September, 2005. Y.-L. Hsueh, M. S. Rogge, W.-T. Shaw, J. Kim, L. G. Kazovsky and S. Yamamoto, "Spectral Shaping Line Codes for Instant Upgrade of Existing Optical Passive Networks," Optical Fiber Communication Conference OFC 2004, Los Angeles, CA, p.FG2, February 2004. Y.-L. Hsueh, M. S. Rogge, W.-T. Shaw, J. Kim, L. G. Kazovsky and S. Yamamoto, "SUCCESS-LCO: Instant and Cost-Effective Upgrade of Existing Passive Optical Networks by Spectral Shaping Line Codes," IEEE Global Telecommunications Conference Globecom 2004, Dallas, TX, p.1907-1911, November 2004.

Next Generation Optical Access Networks

on similar passive tree topologies, but have different data transport technologies. ... Wavelength Division Multiplexing (WDM) technology has been considered an .... in large quantities at a low price is the ONUs and Optical. Network Terminals ...

501KB Sizes 1 Downloads 70 Views

Recommend Documents

Next Generation Networks
Email (for orders and customer service enquiries): [email protected] ..... services based on the Internet Protocol (IP) are replacing traditional telecom.

pdf-0941\optical-and-wireless-communications-next-generation ...
... the apps below to open or edit this item. pdf-0941\optical-and-wireless-communications-next-ge ... g-applied-signal-processing-series-by-matthew-no.pdf.

Next Generation Crowdsourcing for Collective Intelligence.pdf ...
... the smartphone sensor hardware, and autonomously supply the. data content through wifi/mobile networks to the Crowdsourcing endeavor through time.

THE NEXT 10 - New Generation Plantations
Trees are amazing machines for taking carbon out of the atmosphere ... NGP and like-minded organizations contribute to change the rules of the finance game,.

Carbon Nanotube Cantilevers for Next Generation Sensors
This clean separation of intrinsic noise from external factors ... Finally, we discuss other sources of dissipation and their effect on experiments ... mode analysis of the. Timoshenko beam equations10,22 and show the potential energy. U = 1. 2.