Nonlinear dynamic behaviour in optical data networks: experimentation, modeling and numerical techniques

点击翻译按钮获取中文摘要

Future passive optical networks (PONs) for fiber-to-the-home must increase the bandwidth available on a per client basis to meet rising demand. To keep costs low, incoherent sources have been proposed for PONs using spectrum sliced wavelength division multiplexing (SS-WDM). These sources suffer from severe intensity noise, limiting the spectral efficiency and/or bit rate of SS-WDM systems. Semiconductor optical amplifiers (SOAs) have nonlinear dynamic responses that can be exploited to reduce the intensity noise of incoherent sources, albeit with mixed results. In order to explore the SOA behavior, we use Monte Carlo techniques assessing the most important criteria in physical layer data communications - the bit error rate. Given the signal dependent nature of noise in optical systems, nonlinear interactions, and high bit rates, accelerated Monte Carlo techniques such as multi-canonical Monte Carlo (MMC) are required. Armed with accurate numerical models of SOAs and the MMC technique for simulating bit error rate, we will investigate noise cleaning for fiber to the home applications. These techniques are also useful for other dynamic, nonlinear optical systems.