OT sort of- Chromatic Dispersion & AVNX
This was written by Scott Shaw of the Fred Hager Tech Report:
CHROMATIC DISPERSION:
Chromatic dispersion is a problem that will face any broadband network as the network moves to 10 and 40 gbps. It involves the spreading and smearing of light as it travels along the fiber-optic network. The higher the bit rate of information contained on the light the exponentially worse the problem becomes. The higher the lambda channel count in the network, and the subsequent decreasing space between channels the worse it gets. The longer between regeneration stops the worse it gets.
LaserComm and Wit Soundview provided an excellent example of this phenomenon in table form. A channel of light with a bit rate of 2.5 gbps can make it from Wall Street to Chicago without any regeneration due to chromatic dispersion. The same channel of light with a bit rate of 10 gbps can make it only a much smaller distance, from Wall Street to Stamford, Connecticut. At 40 gbps that same channel of light will only make it from Wall Street to Brooklyn. The problem is compounded as the number of channels per fiber increases and each channel of light by necessity becomes closer and closer to each other. At 40 gbps the networks tolerance for chromatic dispersion is 256 times less than the networks tolerance at 2.5 gbps. For the mathematically inclined the amount of dispersion an optical network can tolerate is inversely proportional to the square of the transmitted bit rate.
Even in today’s relatively slow and chromatic dispersion tolerant network, dealing with this problem is still one of the more expensive and limiting aspects of the network. Ciena is on the record as stating that the lack of a solution is the primary reason 10 gbps product has not taken off as quickly as hoped. The current solutions for dealing with this problem are very precise in both distance tolerance between regeneration stops and number of channels being deployed and are very expensive. Each new channel of light requires a new component at each and every stop along the network. At 10 and 40 gbps speeds and with increasing channel counts it is simply not economically viable to continue the network build-out in this fashion. Regenerative components would be needed maybe as close as every 80 kilometers are closer, possibly 20 kilometers at 40 gbps. The broadband network would just die a death due to the enormous expense required by current technical solutions to the problem.
It should also be noted that there are also several types of dispersion problems besides chromatic dispersion; polarization mode dispersion foremost among them and becoming more important as we reach 40 gbps and higher. However, for the foreseeable future, chromatic dispersion will be the critical problem that will need to be resolved to reach 10 and 40 gbps. Amongst the problems that need to be dealt with in resolving chromatic dispersion issues are the need to precisely manage both dispersion and particularly the problem of dispersion slope (each wavelength experiences dispersion at a different rate and needs to be individually corrected) and the problem of handling high optical power input (or the ability to handle many channels). The resolutions of these problems are not just a technical hurdle, but also a major economic hurdle that is critical, no essential, for any future broadband network.
One solution to the problem is through the fiber itself. First modal dispersion was resolved in this manner by Corning with multimode fiber; this led to dispersion shifted fiber that dealt with the subsequent chromatic dispersion problems not addressed by mulitmode fiber. However dispersion shifted fiber was only effective with a single channel of light on each strand of fiber. This is a very limiting solution. This led to the development of non-zero dispersion fiber that dealt with the four-wave mixing effect that was created when multiple channels were used in dispersion shifted fiber. Dispersion shifted fiber, however, like previous generations of fiber, has left its own unresolved issues, that of dispersion slope. Best case scenario, even if some miracle fiber is developed that is cost-effective and can resolve all these issues, is that all the legacy fiber, sitting there, unlit an in the ground, will still need correction for chromatic dispersion regardless of whether or not Corning’s latest technology finally resolves this issues or not. And the faster networks get, the less likely that any fiber technology will be able to do so. The problem with resolving the issue via a fiber only solution is also made dicier because as much of a problem that chromatic dispersion is, some of it is actually desirable to combat the four-wave mixing problem that we discussed above. Bottom line: The fiber-optic threads themselves will not resolve these issues.
Another commonly used solution involves the use of dispersion compensating fiber (DCFs). This is an external device, but continuous to the fiber. However this solution is only really effective at very low channel counts, and can be extremely complex and expensive. It does not resolve the problem. Nevertheless, beyond the fiber solutions addressed above DCFs are really the only solution being utilized in the network today. eDCFs (enhanced DCFs) are also coming out to market, but they are really just incremental improvements over DCFs.
The two most promising solutions for the future appear to be High Order Mode fiber (HOM) that is being pioneered by Lasercomm and Virtual Imaged Phased Arrays (VIPAs) that is jointly pioneered by Avanex and Fujitsu. Both Lasercomm and Avanex are sampling their products to manufacturers.
HOM involves a continuous fiber solution like DCF but that enables the handling of up to 16 times the power of DCFs enabling the handling of much higher channel counts. What I have uncovered about this technology is mostly from Lasercomm marketing material. If they are correct HOM may play a major role in resolving the critical chromatic dispersion problem including the correction of dispersion slope across all channels. I will continue to investigate this issue and update this report regarding the strengths and weaknesses of this technique. Lasercomm originally expected product to be shipping by sometime last quarter. No product shipment announcements have been made. The promises of the technology are very impressive and Lasercomm holds what appears to be a proprietary lock on the technology. But all we really have so far is extensive marketing claims in regard to efficacy of the product without any cost/benefit analysis of the product or whether or not it really works as well as Lasercomm marketing material indicates.
Avanex’s VIPA approach promises similar results as those promised by Lasercomm, including the resolution of the difficult dispersion slope problem. The technology is a collaboration between Fujitsu and Avanex and involves tunable micro-optic technology. Basically shorter wavelengths travel down the fiber faster than the longer wavelengths. These wavelengths, in the simplest terms, are reflected by mirrors in the VIPA device in such a way as to allow the longer wavelengths to catch up. Avanex’s approach provides a high degree of dispersion accuracy, enables the correction of variable dispersion slope, and can compensate for many different levels of dispersion. The technology is tunable and can be used to reconstruct degraded signals off of any quality of fiber-optic string, at any bit rate, channel spacing and wavelength. Of the approaches I’ve examined only the Lasercomm HOM solution appears competitive at this time to VIPA; in fact, if you listen to the marketing material from Lasercomm, the VIPA approach appears to be the second best way to compensate for chromatic dispersion trailing HOM. Stay tuned and we will keep you updated as real world developments occur.
There are also other solutions being developed such as solutions using chirped fiber bragg gratings. Phaethon is leading the charge here. JDSU and others are researching similar techniques. There are drawbacks to this technique like dispersion ripple, a problem that only gets worse as transmission speeds increase, as well as an inability to scale to higher channel counts. Phaethon is attempting to develop a system based on this technology that does not suffer from these limitations. We shall have to see how it develops. Certainly no one can discount what may come out of the R&D labs for companies developing products using these techniques, but at the present chirped fiber gratings do not appear to be in the running as a premiere solution to the problem of chromatic dispersion.
In conclusion, the correction of chromatic dispersion is one of the key enabling technologies, nay, maybe THE KEY ENABLING TECHNOLOGY required to move to 10 and 40 gbps systems. Without an economical solution to this problem network technologies will stall where they are in terms of bit rate and channel count. It is far too early to say that any one technology will be the winner in this space. HOM technology from Lasercomm, and VIPA technology from Avanex are the two leading choices. Both companies own proprietary control of their respective technologies. Lasercomm has yet to be report any product wins, and is a private company that can pretty much say what they please about the technology. And Lasercomm has basically asserted that HOM is the best solution for all problems in regard to chromatic dispersion - don’t look any further. Avanex’s VIPA technology is fully disclosed and protected by patents between Avanex and Fujitsu. The strengths and weaknesses of the technology are known and it is a very elegant solution to the problem; the PowerExpress Metro includes the VIPA technology built-in (PowerShaper is the standalone version of the product). Other technologies exist or are under development by competitors, but none seem likely to equal either VIPA or HOM in either effectiveness or price/performance with the codicil that we don’t quite know how HOM will stand-up to this criteria. We will have to wait for more information in this regard. This is an absolutely critical market to the future of the broadband build-out and will offer a billion dollar market opportunity to those companies that can best solve this problem. Avanex is well positioned to be a leading player in this market and can also leverage its VIPA technology to better create whole solutions by integrating the VIPA technology into its other product offerings. Updates regarding this market will be forthcoming as developments make themselves known.
---------------------------------------------------------
Customers for Avanex photonic processors include WorldCom, Inc., Nortel Networks Corporation, Lucent Technologies, Fujitsu Ltd., Hitachi Ltd., Sycamore Networks Inc., Sorrento Networks, Cogent Communications and Cisco Systems, Inc. |
