I think you have left one important variable out of the trade-off-DWDM requires tradeoffs among speed, channel count, and distance. Starting with Nortel, the OC-192 expertise in itself is not the key, because it is much harder to fit 10 Gbs signals into the wavelength spacings required by high channel-count systems because it is more sensitive to fiber dispersion and nonlinearity. These are the key challenges to overcome, IMHO.
Moving to Lucent, the traditional DWDM approach to modulation is to use NRZ (non-return to zero) signal encoding for each wavelength. Works fine at OC-48, but at 10 Gb/s the NRZ signal is very susceptible to dispersion, which causes the pulse to spread, which limits the effective distance over which the signal can be transmitted.
So it looks to me as though each of them has related problems to overcome in going from what they have to where they need to be, and the key is going to be who can do the best job of managing the effect of dispersion and non-linearities on speed, channel count and distance. Some of this (the distortion effects of fiber non-linearities) can be addressed by dispersion compensation devices, and Nortel has been working in this area for awhile. It's not easy, since where you put them varies with the different fiber types (there is a very promising company called LaserComm with a new approach to this). The best solution, however, may lie in the use of ....
Solitons!
Much has been written about the ability of soliton technology to extend the transmission range of optical systems, for example Gilder (the master of the blindingly obvious) believes this is what NT was after in acquiring Qtera. As noted above, however, the same technology could be used to increase the transmission speed at current distances (OC-768 is theoretically do-able, and and possibly more) or to achieve some better trade-off among the 3 key variables. I expect that this played a role in the decision to acquire Qtera.
|
