Digital Signal Processing
Digital signal processing (DSP) in communication systems refers to the mathematical manipulation of discretized signals, in order to extract the information they contain. The field of DSP was the catalyst that enabled the ‘coherent revolution’ which took place for optical telecom networks in the mid-2000s. In coherent detection, a local oscillator is used in order to detect data that is encoded in both the amplitude and the phase of a carrier signal – a principle that has been used in electrical wireline and wireless communication systems for decades. Researchers in the 90s thought about applying this in optical communications, since it offers significantly higher speeds and longer reach. However, it required digitizing and processing of the received signals in the electrical domain with specialized algorithms; the data-rates of fiber optic systems were orders of magnitude higher than their electrical counterparts, making practical implementation of the concept impossible at the time. The advent of energy-efficient, ultra-high-speed CMOS digital electronics in the following decade changed everything: Real-time DSP could finally be applied at the rates required by optical transmission systems. It goes beyond digital demodulation of signals: DSP enables the compensation of linear and nonlinear fiber transmission impairments (e.g. chromatic dispersion, polarization effects, nonlinear phase noise, etc.), some of which were either impossible to mitigate with direct-detection systems, or required specialized and expensive hardware. The first coherent optical transponder operating at 100 Gb/s appeared on the market in 2010, and today’s core optical networks are dominated by coherent technology. Research into DSP algorithms for optical communications is currently flourishing, with the aim of further increasing bandwidth-efficiency and performance, while at the same time reducing power consumption of the digital receiver. DSP is also starting to impact short-to-medium-reach optical links for datacom applications, where low-complexity direct-detection schemes are the norm: Research into appropriate forward error correction (FEC) and equalization DSP is ongoing.