Modulation Schemes for Single-Laser 100 Gb/s Links: Multicarrier

Abstract

We evaluate multicarrier modulation methods for 100 Gb/s single-wavelength data center interconnects. We consider two different orthogonal frequency-division multiplexing (OFDM) techniques: DC-biased OFDM (DC-OFDM) and asymmetrically clipped optical OFDM (ACO-OFDM). We also consider two different techniques for bit loading and power allocation: fixed bit loading with preemphasis and optimized bit loading and power allocation. We first present a semianalytical performance and complexity evaluation of these OFDM methods including the effects of linear filtering, clipping, and quantization. We then include the effects of chromatic dispersion and chirp, as well as intensity and shot noises. Performance is quantified in terms of the required average optical power to achieve a target bit-error probability for a given modulator bandwidth. Complexity is quantified in terms of the resolution and sampling rate required of digital-to-analog (DAC) and analog-to-digital (ADC) converters, as well as the number of signal processing operations required. For each OFDM technique, we adjust the clipping ratio to minimize the optical power requirement. For DC-OFDM, taking into account the DAC frequency response reduces the optical power requirement up to 2 dB. ACO-OFDM is more power efficient and requires lower DAC/ADC resolution than DC-OFDM, but ACO-OFDM requires prohibitively high sampling rates owing to its poor spectral efficiency.

Hermitian symmetry free optical-single-carrier frequency division multiple access for visible light communication

“…In this subsection, we investigate the impact of quantization by evaluating the effective number of bits required to achieve a target BER. Following [32], the required electrical SNR per bit in an AWGN channel, E req b (elec) /N 0 , for the target BER is evaluated, using which the effective number of bits are calculated. A target BER of 10 −3 and uniform quantization is considered.…”

“…A target BER of 10 −3 and uniform quantization is considered. The quantization noise is modeled as an additive, uniformly distributed white noise [32].…”

Decision-directed iterative methods for PAPR reduction in optical wireless OFDM systems

“…The resolution, or step size µ, of the quantizer is the difference between two adjacent quantization levels, and is evaluated as µ = ξ/2 Q b , where ξ = ξ upper −ξ lower and is equal to DR considered in the previous sections. Quantization noise for clipped O-OFDM can be modeled as an additive, uniformly distributed white noise, with variance given by [34]…”

“…From ( 25), it can be observed that since the signal excursion of DCO-OFDM is twice the signal excursion of ACO-OFDM, the quantization noise variance for DCO-OFDM is four times greater than ACO-OFDM [34].…”

“…From (34), the complexity of O-OFDM with FDCDR in terms of the number of real operations per bit is evaluated as…”