Abstract
In this tutorial, we discuss the evolution of the technology deployed for optical interconnects and the trade-offs in the design of low complexity, low power DSP and implementation for direct detect and coherent, pluggable optical modules for data center applications. The design trade-offs include the choice of modulation format, baud rate, optical link design, forward error correction, signal shaping and dispersion compensation.
Highlights
T HE use of digital signal processing (DSP) in optical links has about a 20-year history [1]–[4]
The first use of DSP was in MLSE implementation for chromatic dispersion compensation in intensity modulated direct detect (IMDD) 10 Gbit/s systems [2], [3]
This may vary between different vendors, we find this is to be generally true for a variety of DSP based modules
Summary
T HE use of digital signal processing (DSP) in optical links has about a 20-year history [1]–[4]. The first use of DSP was in MLSE (maximum likelihood sequence estimation) implementation for chromatic dispersion compensation in intensity modulated direct detect (IMDD) 10 Gbit/s systems [2], [3] This was not very power efficient and provided only limited mitigation of chromatic dispersion for long-haul applications; MLSE was not deployed widely in 10 Gbit/s IMDD systems. The development of high-speed coherent receivers enabled the linear detection of both signal amplitude and phase. The use of DSP in coherent applications to compensate for a variety of linear optical impairments, as well as enabling QAM (Quadrature Amplitude Modulation), was a very powerful innovation which led to the widespread deployment of DSP hardware in optical links [1], [3].
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
