NFOEC News Analysis   More NFOEC News Analysis

Phaethon Embarks on Tunability

Startup Phaethon Communications emerged from stealth mode today, claiming a unique solution to a troubling problem that's emerged in optical networks (see Phaeton 'Unveils Its Vision').

That problem is how to ensure that wavelengths stay strong across fibers that have different physical characteristics. It's a problem that arises when carriers are designing or installing DWDM (dense wavelength-division multiplexing) gear and when wavelengths are switched. When this happens, wavelengths are often driven across fibers of different types and lengths.

At the root of the problem is a technical issue called dispersion -- the degradation of light signals when they're sent over fiber. There are different types of dispersion. "Chromatic dispersion" happens because a light pulse isn't just a single wavelength; it's a bunch of wavelengths representing slightly different colors. The longer wavelengths travel at slightly different speeds than the shorter ones, so over distance the pulse spreads until it bumps into the adjacent one, at which point the signal is lost.

This problem gets worse with speed. At 40 Gbit/s, for instance, the problem is 16 times worse than it is on 10-Gbit/s links.

To compensate for the inevitable effects of chromatic dispersion, Phaethon has designed a tiny device for use by system vendors. It's a fiber Bragg grating (FBG) that has been etched in a special fashion, then microscopically stretched. The end result is a component that can change the compensation values for different bunches of wavelengths in a single fiber. Hence, the term "tunable" compensator.

Today's chromatic dispersion compensators use a different tack. Most are small modules containing dispersion compensating fiber, which are inserted at the receiver on a DWDM system and in the network itself, right before the amplifiers on the fiber -- typically every 80 to 100 kilometers.

These modules, exemplified by the products of Corning Inc. (NYSE: GLW), are preset with so-called "compensation values" that match the type and length of fiber on which they reside. Accommodating different types or lengths of fiber, or switching to different speeds, requires another module.

Phaethon says its solution is useful because one component will be able to automatically adjust to different compensation values.

"Say you had a 100-channel DWDM system with channels organized in groups of ten," says Bruce Barton, VP of marketing at Phaethon. "You can set different compensation values for all of the channel groups, using one device. There's no need to drive out and reinstall the fiber and modules."

Analysts say that if Phaethon can live up to its promise, it's got something good. "I haven't heard anyone offer remote tunability yet," says Tom Hausken, director of optical communication components at Strategies Unlimited, referring to Phaethon's capability to build a component that can make dynamic adjustments.

There are drawbacks. For one thing, Phaethon won't have automatic tunability in its products right off the bat. Instead, for a few months, customers will be offered the ability to tune the compensation values by hand.

Another problem: Phaethon will offer tunable compensation for only small groups of wavelengths at first. It plans to unveil 4- and 8-channel modules at the upcoming NFOEC show in Baltimore.

Phaethon's devices also don't solve all the problems associated with chromatic dispersion. LaserComm Inc., a startup selling dispersion compensating modules based on higher-order fiber, says its devices will take care of "dispersion slope," which refers to the ability to set different degrees of dispersion compensation for different channels. Phaethon's devices can't tackle this problem right now, although the vendor says that capability is on the drawing board.

Phaethon also doesn't yet handle (but plans to handle) polarization mode dispersion (PMD), a problem caused by the fact that light can travel along slightly different paths inside a fiber, depending on how it bounces around as it travels. Vendors such as Yafo Networks have started to tackle PMD (see Can Yafo Lift Speed Limits?).

Some naysayers hold that fiber Bragg gratings aren't the way to handle chromatic dispersion problems, since they tackle a low number of frequencies at once. Not so, says Phaethon, which insists its FBGs are key to the tunability feature and offer low insertion loss and low cost ("less than $10,000 per 4- or 8-channel module," according to Barton).

Phaethon also points to its other assets, which include more than 15 patents from the University of Southern California, thanks to founder Alan Wilner, a professor at USC. The company also has right of first refusal on any new photonic technologies from USC, and it can opt to license them exclusively for a ten-year period.

Still, the complexity of the problems Phaethon hopes to solve ensures it's going to be a while before full success is achieved. Meantime, its competitors, including the likes of LaserComm and Yafo, also are at work on comprehensive solutions for chromatic problems. And there's always a chance they may get there first, given the headstart they've got.

It's also not unthinkable that some as-yet unimagined combinations may result from all these developments. "We're looking at adaptive optical networks, having a way to sense the quality and respond with a standalone system that works with multiple DWDM devices," says Henry Yaffe, founder, chairman, and CTO of Yafo Networks. "A company such as Phaethon could be partners with us. Or, depending on their viewpoint, they could be competitors. Or a bit of both."

- Mary Jander, Senior Editor, Light Reading
http://www.lightreading.com

For more information on NFOEC, please visit the Light Reading NFOEC Site.

Newest Comments First       Display in Chronological Order
Page 1 of 3 Next >
exphoton
User Ranking
Friday June 29, 2001 9:29:04 AM
no ratings
nbw

Thanks - I feel much better now. I will gladly attend NFOEC, go to Pheathon's booth and hopefully grab a few of those intransigent photons.

Very nice.

EP.
Petabit
User Ranking
Friday June 29, 2001 7:37:46 AM
no ratings
To: Milano
Co-Founder
Petabit & Milano International Consultants

Dear Milano,

I wish to issue a minority dissenting report, to prevent our consulting company getting sued by out clients.

1. Whilst the amount of chromatic dispersion variability in a system varies with the system length, it also depnds on the bit rate. A 10G ULH system may have 4 times the CD variation of a conventional reach 10G system, but a 40G system is 16 times more sensitive to CD. So 10G ULH systems may not need tunable CD compensators, but 40G will.

2. Post 1998 fibres will run 40G with no PMD compensation. This has been demonstrated in several of the field trials that have been published. (See LR's 40G report for more details). Since this only accounts for some of the fibres in the ground, a PMD compensator might be a good plan for increasing the number of fibre that the 40G will support. But it is not essential.

3. "tunable Bragg is the main contender to provide PMD compensation". Really? I beg to differ.

4. "only one vendor, Yafo Networks, is considered credible in the current market." Again, really? The main suppliers of PMD compensators are the system vendors themselves. They appear to be the main people publishing papers, and therefore probably hold a lot of IP.

Please edit your report before you send it to Prof. Wilner.

P.
minivan
User Ranking
Thursday June 28, 2001 2:11:15 PM
no ratings
"1- Limited market: the market for tunable chromatic dispersion compensation is very limited, since it is not required or used in many ULH and 40 G architectures."

Disagree ! Tunable chromatic dispersion compensation is required for 40G and would even be usefull for ULH. Make a good spec. device and there's a big market.

"2-Competition by alternative technologies: Dispersion Compensation Fiber (DCF) is a well-established and proven solution. One of the main limitations of DCF, the lack of slope compensation, is also being addessed by vendors such as Lucent and Corning."

This is not the main limitation of DCFs at 40G.

"1-Important market: all 40 G architectures require PMD compensation."

Not true. Many deployed fibres will support 40G without PMD compensation, over reasonable LH distances. Future fibres will be much better (cf new fibre PMD spec improvements as 10G started to be deployed).

"Best Regards,

Milano
Co-President, Petabit & Milano Int'l Consulting"

I hope Mr. Wilner didn't pay for that advice !

40G PMD compensators may eventaully be a big market (but there's lot's of size, cost, and performance impovements to do) - but on the other hand look back at what happened when 10G was first proposed. Everyone was predicting you would need one per 10G lambda but you'd never be able to sell a PMD compensator for 10G these days. Ask Yafo.

Minivan


realguy
User Ranking
Thursday June 28, 2001 1:27:33 PM
no ratings
"Have you heard of General Photonics? It seems to have product that works, if sales are anything to go by. I found their story pretty convincing
What approach to fixing PMD were you working on?
rigby@lightreading.com"

It is a good question. First. There are several orders of PMD. First order PMD is relatively easy to fix with an active polarization differential delay to bring one broken runaway pulse into other. The problem is exasperated when the pulse has broken to several pieces, say 4 or 8 pulses. Here no PMD compensator can do justice. Due to nonlinear interference among pulses (yes nonlinearity is also should be counted) the distortion in the signal becomes uncorrectable. The 40G systems are especially prone to this because they suffer more nonlinearity than the 10G counterpart.

The General Photonics has a good polarization controller. What does that mean? It means one can adjust or maintain the polarization of the light into a device. This would help you to reduce the PDL of a system since there are filters and polarization sensitive elements in the link. It may also help if fiber link has first order linear PMD problem but you will be out of luck if higher order PMD and nonlinearity are present and interplay among each other.


Pauline Rigby
User Ranking
Thursday June 28, 2001 11:19:32 AM
no ratings
Realguy said:

>I don't care who did what regarding the PMD
>issue in the fiber. I have worked on this >problem over several years and believe me the
>theories describing PMD in the fiber is totally
>off the wall. This problem has not been fully
>understood and to say there will be a product
>soon that can addresses the problem is simply
>not realistic. The best solution to PMD is, just
>lower the bit rate while maintaing your spectral
>efficiency

Have you heard of General Photonics? It seems to have product that works, if sales are anything to go by. I found their story pretty convincing (see http://www.lightreading.com/document.asp?site=lightreading&doc_id=6315 ).

What approach to fixing PMD were you working on?

rigby@lightreading.com
Milano
User Ranking
Thursday June 28, 2001 9:14:47 AM
no ratings
To: Alan Wilner
Co-Founder
Phaethon Communications

From: Petabit & Milano International Consultants

Dear Alan,

Please find attached the result of the analysis of Phaeton’s business plan that was conducted by our firm. In summary, we believe that by adjusting its direction, Phaethon has all the elements to be a very successful venture.

First, we conclude that the current direction of bringing tunable chromatic dispersion compensators to the market should not be pursued further for the following reasons:

1- Limited market: the market for tunable chromatic dispersion compensation is very limited, since it is not required or used in many ULH and 40 G architectures.

2-Competition by alternative technologies: Dispersion Compensation Fiber (DCF) is a well-established and proven solution. One of the main limitations of DCF, the lack of slope compensation, is also being addessed by vendors such as Lucent and Corning.

3-Competition by other vendors: in the Bragg-based tunable compensation, Phaethon would also face competition from established vendors such as JDS Uniphase.

4-Resources allocation: bringing a product to the market would monopolize scarce technical resources as well as capital. The effort needed to qualify a product should not be underestimated.

Second, we conclude that Phaethon should focus its business plan on being a PMD-compensation pure play for the following reasons:

1-Important market: all 40 G architectures require PMD compensation.

2-Competition by alternative technologies: tunable Bragg is the main contender to provide PMD compensation.

3-Competition by other vendors: only one vendor, Yafo Networks, is considered credible in the current market.

4-Resources allocation: focussing resources on R&D only will limit the capital requirements and increase the morale of R&D people.

The following actions should be taken to re-inforce the direction:

1-Deepen your Board by binging an experienced manager with extensive contacts in leading 40G manufacturers.

2-Establish a partnership with a 40G vendor for the development of the product specification and the interface to the system.

3-Increase marketing efforts to project an image of constant progress in the PMD compensation, directed primarily at the financial community.

Best Regards,

Milano
Co-President, Petabit & Milano Int'l Consulting
minivan
User Ranking
Wednesday June 27, 2001 5:47:04 PM
no ratings
Milano,

Dynamic chromatic dispersion compensation is also required at 40G, due to the narrow dispersion tolerance window of most 40G modulation formats and the variable nature of the dispersion of many system elements over temperature/aging etc.

Even at 10G, in ULH systems, one must watch out that a dispersion budget isn't eaten up by extra contributions to the end-to-end total from things like fiters, amplifiers, etc. etc. A dynamic chromatic dispersion compensation unit (probably per channel) would help here too, making it easier to provision a system. The key requirements then are size, cost, and minimal penalty relative to the alternative of fixed DCMs and careful budgeting. I don't believe such devices exist.

Minivan
flanker
User Ranking
Wednesday June 27, 2001 3:54:20 PM
no ratings
...And my point was, there is non need to go to 40G and be tackling PMD actively when you can avoid all together at 10G...

I think you unintentionally touched on the issue that there is no business case for deploying 40G until after 4q2003 for most carriers (something LR wont talk about lest it piss off their advertisers). Why even plan to deploy 40G when 10G will be humming by next year? Maybe Nortel's Roth would have sounded more intelligent had he said "today there's no demand for technology allowing exponential capacity increases" instead of "IP traffic is shrinking".

realguy
User Ranking
Wednesday June 27, 2001 3:06:23 PM
no ratings
"PMD compensation is a must for 40G systems and given its statistical nature, must be dynamic with control from the system. My point was: who needs tunable *chromatic* dispersion compensation - I mean, in real networks?"

And my point was, there is non need to go to 40G and be tackling PMD actively when you can avoid all together at 10G using far simpler approaches such as duo-binary formats that achieve the same spectral efficiency. The 40G system will never be deployed on the existing fibers suffering from high PMD. Only will be reserved for more modern fibers of the future.
Photonboat
User Ranking
Wednesday June 27, 2001 1:48:38 PM
no ratings
>>My point was: who needs tunable *chromatic* dispersion compensation - I mean, in real networks?<<

"...and when wavelengths are switched. When this happens, wavelengths are often driven across fibers of different types and lengths."

My guess is there are no current networks that would use a tunable chromatic dispersion compensator--but I'm not too knowledgeable on architecture issues. I suppose a tunable chromatic dispersion compensator allows for quick provisioning in a future mesh network. As new transmit paths are established, some lasers will be transmitting significantly longer distances than others. As changes are made as to how a wavelength path is used, then the chromatic dispersion compensation will need to be adjusted as the path length for a laser-receiver pair will have been altered.

What such a network would allow is that the core backbone would be indifferent to provisioning and changes at the edge. This "indifferent" core network would need "sexy" amplifiers that have fast compensation for transient spikes (results when channels are added) and dynamic gain flattening as well.

Photonboat

ps I really have enjoyed Milano and Petabit posts, especially in the Marconi thread. Thorough, authoritative, and knowledgeable.
Page 1 of 3 Next >
LIGHT READING MARKET PLACE
The Time is Now for FCoE
Join Cisco and its partners for a live informative webcast on 12/10/09
Conferencing System
Enter Now to Win Two Polycom Video Conferencing Systems. Details Here!
Your Customer Experience Defines You
OnProcess helps market leaders proactively improve their customers' experiences
Used and Refurbished Nortel Routers
Purchase Your Routers From Network Liquidators. Savings of Up to 90% with a Lifetime Warranty!
Used and Refurbished HP ProCurve Switches
Lifetime Warranties, Professional Testing & Shipping on all HP Equipment Purchases!
The blogs and comments are the opinions only of the writers and do not reflect the views of Light Reading. They are no substitute for your own research and should not be relied upon for trading or any other purpose.
White Papers SPONSORED CONTENT
Featured
Podcasts SPONSORED CONTENT
Services Transformation - by Alcatel-Lucent Communications service providers want to be able to bring new services to...
Rural Ops Bridge the Digital Divide - by Tellabs Tellabs helps IOCs build triple play networks
Driving Network Transformation - by Alcatel-Lucent In order to deal with competitive pressures, the change in service models...
Back(haul) to the Future - by Tellabs Tellabs works with Vodafone to meet growing mobile broadband demands.
MRS Logistica - by Tellabs Tellabs helps MRS Logistica transform its existing, largely outdated TDM networks to IP.
Carrier Ethernet Offers an Enterprising Solution - by Tellabs What is VPLS and how does it work? Tellabs takes a closer look.
Swisscom’s Network Makeover - by Tellabs Fresh off the launch of 7.2 Mbps HSDPA, Swisscom sees 3G as an opportunity to launch a unifying ...
Telecom in Namibia - by Tellabs Tellabs helps Telecom Namibia with next-gen challenges
Companies
Alcatel-Lucent (5872), AT&T (1948), BellSouth (848), BT (1287), Cablevision (615), Cisco (5297), Comcast (1910), Cox Communications (858), Deutsche Telekom (807), eBay (Skype) (345), Ericsson (1617), France Telecom (964), Google (489), Huawei (1045), Intel (1127), Juniper (2022), Microsoft (1115), Motorola (1486), Nokia Siemens Networks (2645), Nortel (3956), NTT (173), Siemens (1359), Sprint (1059), Telefonica (439), Time Warner Cable (969), Verizon (2587), Vodafone (510), Yahoo (339)

Broadband
Access equipment (2169), Access technologies (2378), Broadband loop carriers / multiservice access nodes (388), Cable modem termination systems (CMTSs) (1104), Cable TV chips (286), DSL (2425), DSL chips (227), DSLAMs (703), Free-space optics (35), FTTx (3265), Gaming consoles (58), Gaming servers (22), Media adapters (23), Municipal networks (106), PON (1364), PON chips (217), Satellite (497), WiMax (880), Wireless LAN (354)

Cable Digital
Cable Modems (681), Cable/MSO equipment (2802), CableLabs (470), Compression (MPEG-2 and MPEG-4) (279), Docsis (1046), Embedded multimedia terminal adapters (E-MTAs) (213), Head-ends (233), PacketCable (129), QAM (307)

Chips, Components & Subsystems
ASICs & FPGAs (101), ATCA (480), ATM chips (13), Comm chips (2360), Dispersion compensators (149), Lasers (920), Modulators (163), Mux/demuxes (299), Network processors (933), Optical amplifiers (349), Optical channel monitors (92), Optical components (2824), Speciality fiber (94), Switches & OADMs (397), Transceivers (1247), Transmission fiber (419), Variable optical attenuators (139)

Ethernet
10-Gbit/s Ethernet switches (1454), Access devices (272), ATM switches (333), Circuit emulation (16), Converged access (103), Ethernet chips (573), Ethernet equipment (2212), Ethernet over copper (231), Ethernet PONs (160), Ethernet services (1909), Ethernet technologies (568), Multipoint (131), Multiservice edge equipment (143), Multiservice provisioning platforms (622), Multiservice switches (389), PBT (Provider Backbone Transport) (256), Point-to-point (139), Pseudowire (Layer 2 tunnels) (132)

IP & Convergence
B-RASs (229), Cell/WLAN (77), Compression equipment (13), Core routers (1294), DNS (56), Edge routers (1686), ENUM (53), Fixed/Mobile Convergence (485), GMPLS (76), IMS (1088), IMS Control Layer (27), IMS Service Layer (27), IP equipment (1224), IP software (381), IP technologies (1482), IPv6 (99), Layer 3 VPNs (194), MPLS (1774), MPLS (687), Multicast (36), P2P (258), Pseudowire (Layer 2 tunnels) (132), QOS (350), SIP (396), Traffic managers (808), Wireline/Wireless (59)

Mobile/Wireless
3G Evolution (175), Broadcast (Mobile TV, etc.) (189), Carrier WiFi (226), CDMA (3G) (367), Core Network (173), EV-DO (126), Femtocells (30), Fixed Wireless (Microwave, etc.) (71), Fourth Generation (4G) Wireless (70), GSM/EDGE (430), HSDPA/HSUPA (321), IMS Core (47), Long-Term Evolution (LTE) (188), Mobile Advertising (24), Mobile Music (31), Mobile TV (130), Mobile Video (65), Mobile WiMax/WiBro (92), Mobile/Wireless (5877), Packet Core (61), Radio Access Network (236), TD-SCDMA (Chinese 3G) (67), Transmission (38), Ultra-Mobile Broadband (UMB) (8), UMTS(3G) (340), Voice Core (21), WiMax (880), Wireless Backhaul (272), Wireless Chips (191), Wireless LAN (354)

Optical Networking
40-Gbit/s transmission (452), Core optical switches (760), CWDM (289), DWDM (1842), Long-haul WDM equipment (654), Metro optical switches, ROADMs (1173), Metro WDM equipment (773), Multiservice provisioning platforms & add/drop muxes (375), Optical equipment (2191), Optical switches & crossconnects (398), Optical technologies (417), Sonet/SDH (1036), Sonet/SDH chips (351), Wavelength services (305)

Security
Anti-virus (29), Denial-of-service attacks (44), Encryption (97), Endpoint security (22), Firewalls (61), Intrusion detection & prevention (45), IPSec VPN (801), Security (1835), SSL VPN (862), URL filtering (12), User authentication (24)

Services Software
Activation (415), Billing systems (761), Content/software downloads (231), Customer relationship management (231), Data Integrity (61), Element management systems (36), Fault management (69), Inventory management (153), Mediation systems (204), Messaging (231), Middleware (72), Mobile location (41), OSS (2584), Performance monitoring (335), Policy control (269), Provisioning (553), Revenue assurance & fraud management (334), Service delivery platforms (SDPs) (328), Service management (220), Service-oriented architectures (310), Services (2480), Web gateways (56), Web services (124), XML (51)

Test & Measurement (Sponsored by Etaliq Inc)
Access equipment Access test & measurement equipment (126), Comm chips Comm chips test & measurement equipment (29), Ethernet equipment Ethernet test & measurement equipment (170), IP equipment IP test & measurement equipment (122), MPLS MPLS test & measurement equipment (14), Optical components Optical components test & measurement equipment (113), Optical equipment Optical test & measurement equipment (886), OSS OSS test & measurement (1059), Sonet/SDH Sonet/SDH test & measurement equipment (1599), Test & measurement (1755), VOIP equipment VOIP test & measurement equipment (145)

Video (Sponsored by Ericsson Televisionary)
Broadcast (Mobile TV, etc.) (189), Broadcast video equipment (including encoding) (730), Content delivery network (CDN) (394), Content protection (270), DVRs (665), Internet Video (840), IPTV (3461), Middleware & business support systems (845), Set-top boxes (1624), Stored video servers (379), TV (3581), Video equipment (2448), Video services (4130), Video software (1349), Videophone (185), VOD (2635)

VOIP
Application servers (186), Centrex (198), Conferencing (78), Contact centers (38), Enhanced voice (34), Enterprise (637), Media gateways (357), Messaging (73), Presence management (43), Residential (835), Session border controllers (398), Signaling gateways (104), Softswitches (1090), VOIP chips (167), VOIP equipment (3423), VOIP services (3768), VOIP software (620), VOIP VPNs (28), Wholesale (220)