TABLE OF CONTENTS In Chapter 5, we have just seen how the perfect transparency of ordinary solitons to one another in lossless fiber begins to break down in real fiber, and thus to cause certain problems and limitations for dense WDM. In Chapter 2, however, we have already seen how the phase mismatch created by the large values of local dispersion tends to make four-wave mixing during soliton-soliton collisions negligibly small in dispersion-managed systems. Thus, in contrast to ordinary solitons, the only nonlinear penalty for dense WDM with dispersion management stems from cross-phase modulation during the collisions, and the timing jitter that results from it. As will soon be evident, however, even in this regard there are great differences, one of the most fundamental of these being the fact that with dispersion-managed solitons, the collision length is essentially independent of channel spacing, again with generally positive consequences. Finally, we shall see how a recently invented trick of dispersion management tends to render the jitter from collisions so small that the performance of a dense WDM system becomes nearly indistinguishable from that of a single, isolated channel.
As we have just seen in the previous two chapters, in WDM with ordinary solitons, pulses in a shorter wavelength channel steadily overtake and pass through the pulses of a longer wavelength channel. With dispersion management, however, the situation is more complicated: In response to the large and rapidly alternating D values, solitons from...
