Varieties and characteristics of discrete spectral encoding (DSE)

Conventional wavelength division multiplexing (WDM) practices use discrete wavelengths to define virtual channels for communication. The number of such WDM virtual channels may be limited by physical effects such as wavelength crosstalk, four-wave-mixing, and the combinations of filter free spectral range and bandwidth over which the associated optical amplifiers have uniform gain, as well as by the number of WDM wavelengths available. Therefore, large (i.e., >10 nodes) networks or interconnection systems which depend on WDM require wavelength re-use, multihops, and/or incorporation of other physical multi-access techniques, such as time division multiplexing (TDM). Alternatively, as we discuss, the few WDM wavelengths can be considered to be a set of “unit vectors”, orthogonal within the WDM crosstalk, in a WDM hyperspace. This alternative approach leads to a variety of new ways to define virtual channels, based on different ways of producing coding formats from these WDM wavelengths to define addresses for the transmitters (senders) or receivers (destinations). Since the number of WDM wavelengths are finite (such as in a WDM transmitter array) and well defined, the term discrete spectral encoding, or “DSE”, is used to describe this new kind of coding and to differentiate it from spectral slicing WDM. Depending on the manner in which the wavelengths of a WDM array are used to define the code, a WDM array of n wavelengths can generate a space with many more than n addresses. This paper describes some varieties of DSE and the characteristics of their multi-access interference