From Scaling Disparities to Integrated Parallelism: A Decathlon for a Decade
From Scaling Disparities to Integrated Parallelism: A Decathlon for a Decade
复制标题
DOI：
10.1109/jlt.2017.2662082
复制
发表时间：
2017-03
影响因子：
4.7
通讯作者：
P. Winzer;D. Neilson
中科院分区：
文献类型：
--
作者：
P. Winzer;D. Neilson
基于来自不同地区和应用程序的各种长期网络流量数据，除了关键信息和通信技术的长期扩展趋势之外，我们还确定了用于生成和处理数据的技术与用于传输数据的技术之间的根本扩展差异。这些差异可能导致数据传输网络每五年落后其所需能力大约 4 倍。到 2024 年，我们预测在 1-Pb/s 光传输系统中需要 10-Tb/s 光接口。为了满足这些需求，需要以波分复用×空分复用矩阵的形式在波长和空间上进行复用。我们估计了此类系统的特征并概述了它们的目标规格，这表明需要在从系统和网络架构到数字信号处理再到集成阵列设备设计的多个领域取得非常重大的研究进展，以避免光网络容量紧缩。
Based on a variety of long-term network traffic data from different geographies and applications, in addition to long-term scaling trends of key information and communication technologies, we identify fundamental scaling disparities between the technologies used to generate and process data and those used to transport data. These disparities could lead to the data transport network falling behind its required capabilities by a factor of approximately 4 every five years. By 2024, we predict the need for 10-Tb/s optical interfaces working in 1-Pb/s optical transport systems. To satisfy these needs, multiplexing in both wavelength and space in the form of a wavelength-division multiplexing × space-division multiplexing matrix will be required. We estimate the characteristics of such systems and outline their target specifications, which reveals the need for very significant research progress in multiple areas, from system and network architectures to digital signal processing to integrated arrayed device designs, in order to avoid an optical network capacity crunch.