Continuously tunable 1.16 μs optical delay of 100 Gbit/s DQPSK and 50 Gbit/s DPSK signals using wavelength conversion and chromatic dispersion
Spotlight summary: Correct timing is a priority when you have to handle many tasks in a short time. If you are a router of an optical fiber network, you must intelligently handle thousands of millions of small messages per second, each delivered at tens or even hundreds of gigabits per second (Gbps). Even if you are very fast, a correct timing for each message is crucial to avoid a collapse. The work by S. R. Nuccio and colleagues from Alan Willner’s laboratory in the University of Southern California is a step forward toward achieving a correct timing in these vital network elements.
Optical fiber networks nowadays are composed of high-capacity optical fiber links that connect to huge electronic nodes (routers). These electronic nodes are in charge of all the switching and routing functions, i.e., directing the right signal to the right port at the right time. Such intelligent functions of the network are conventionally done in the electrical domain, thus needing bandwidth-limited optical-electrical conversions in every step. To avoid this bottleneck and benefit from the virtually unlimited bandwidth of optics, all optical signal processing is considered indispensable. Although for many years—even decades—a promise, all-optical signal processing is still catching up with the real picture of modern network demands.
The development of all-optical signal-processing techniques has been seriously impaired because of the lack of timing tools. The ability to temporarily store a signal and recall it at an exact point in time is an essential function of any type of signal-processing system. This is known as buffering. Buffering is extremely simple and cost-effective in electronics (current embedded random-access memory modules accommodate 100 million bits in less than 1 mm2 with storage times of at least tens of microseconds) but is extremely complicated in optics. However, progress to achieve all-optical buffering devices is under way. One example of this progress is the continuously tunable delay line of 1.16 µs S. R. Nuccio et al. now present.
The scheme used by this group makes use of a very simple yet powerful idea: wavelength conversion and chromatic dispersion. The incoming data signal (as fast as 50 or 100 Gbps) is wavelength converted, propagates through a strongly dispersive medium (a dispersion compensating fiber in this case) and is then wavelength converted back to its original wavelength. By changing the conversion wavelength one can effectively “change” the group velocity of the medium seen by the signal, and thus the end-to-end delay suffered by the bit stream. Of course, that the signal propagates through a strongly dispersive medium is an issue, since it creates intersymbol interference. However, use of mid-span optical phase conjugation completely compensates for the chromatic dispersion. This scheme has two important advantages: (1) it can be directly used for higher bit rates (with the only limit of higher-order dispersion) and (2) it is essentially independent of the modulation format used. These are unique features that make this approach extremely attractive for optical communications, since it can operate with virtually any signal and advanced modulation format that we can produce in the near future.
Further progress in this area should go toward reducing both the size of the setup and the power consumption. The most critical part seems to be the achievement of energy and size efficient wavelength-conversion steps. Recent record-breaking results of wavelength conversion in silicon photonics may be the route to bringing this longstanding promise of all-optical signal processing even closer to reality.
-- Miguel González Herráez
Technical Division: Optoelectronics
ToC Category: Fiber Optics and Optical Communications
|OCIS Codes:||(060.5060) Fiber optics and optical communications : Phase modulation|
|(070.4340) Fourier optics and signal processing : Nonlinear optical signal processing|
You must log in to add comments.