Space Division Multiplexing

本提案的研究目标是建立一个使用多核光纤的世界级空分复用（SDM）计划，以形成下一代，每秒太比特（10e12比特/秒）的超密集光互连。我们将设计、制造和测试与高速垂直腔面发射激光器（VCSELs）二维阵列技术兼容的多芯光纤和耦合器件，用于开发容量大幅扩大的有源光缆。**近几十年来，我们见证了一场令人难以置信的通信和计算革命，这是基于光纤、光子器件和电子电路的进步。虽然单模光纤和波分复用（WDM）在历史上为长途和城域网络提供了大量的传输容量，但光网络的新应用，如机器互连和局域网，需要每根光纤芯的大容量和增加的芯空间密度。电子工业目前正在经历集成电路的通信瓶颈，这在芯片间通信（例如，微处理器和存储器之间，以及多个微处理器之间）中尤为明显。短距离光互连，通常跨越几百米，克服了由散热和尺寸拥挤引起的电气互连瓶颈：2013年制造的高性能超级计算机通常使用超过10e5个激光器；IBM蓝水超级计算机使用超过100万个光通道！在计算系统中，空间是一个主要问题。然而，极短距离光互连（小于1米长）的好处仍然需要建立芯片间数据通信和核心对核心片上通信。与使用传统的单芯光纤相比，能否设计出更节省空间的光链路？**光纤制造技术的最新发展引发了人们对多芯光纤的兴趣，通过在相同直径的单个光纤中构建几个引导芯来增加空间密度。2012年，日本NTT的研究人员使用这种12芯光纤展示了每秒1.01拍比特（10e15比特/秒）的创纪录传输容量。该研究计划将通过开发定制的多芯光纤、有源和无源耦合设备、电路和架构，将这一成果转化为光互连，以取代目前行业中使用的空间效率较低的光纤带。新的光互连将能够在几十个引导核心中承载独立的通道，从而在使用25gb /s vcsel的单个光纤上实现近太比特/秒的容量。我们的团队已经配备了必要的设计工具，为光互连的SDM概念做出了贡献，为多芯光纤采购建立了关键的加拿大和国际合作，并从世界一流的实验室设施中受益。**拟议的研究计划将有助于电子和光子集成技术之间协同作用的发展，这将导致光电子制造的革命性变化，这将极大地改变设计原则并从根本上提高性能。加拿大在信息和通信技术（ICT）领域处于世界领先地位。拟议的研究直接支持加拿大在ICT设备和系统方面的战略重点（如NSERC选择的目标领域和研究主题所述），通过开发将大大增加光链路空间密度的技术。

The research objective of this proposal is to build a world-class program on space-division multiplexing (SDM) using multicore optical fibers, in order to shape next-generation, terabit-per-second capacity (10e12 bits/s) ultra-dense optical interconnects. We will design, make and test multicore optical fibers and coupling devices compatible with the established technology of two dimensional arrays of high-speed vertical-cavity surface-emission lasers (VCSELs) for the development of active optical cables with dramatically expanded capacity. **In recent decades, we have witnessed an incredible revolution in communications and computing based on the advancement of optical fibers, photonic devices and electronic circuits. While single-mode fibers and wavelength division multiplexing (WDM) historically offered plenty of transmission capacity for long-haul and metro networks, the new applications for optical networks, such as machine interconnectivity and local area networks, require both large capacity per fiber core and increased core spatial density. The electronics industry is presently experiencing communications bottlenecks in integrated circuits, and this is particularly evident in inter-chip communications (e.g., between microprocessor and memory, and between multiple microprocessors). Short-reach optical interconnects, which typically span up to a few hundred meters, overcome the bottleneck of electrical interconnects caused by heat dissipation and size congestion: high-performance super computers manufactured in 2013 use more than 10e5 lasers typically; the IBM Blue Water super-computer uses over 1 million optical channels! In computing systems, space is a major concern. However, the benefits of very short reach optical interconnects (less than 1 meter long) still have to be established for inter-chip data communications and core-to-core on-chip communications. Can more space-efficient optical links be engineered, versus using the conventional single-core optical fiber?**Recent developments in fiber-optic fabrication techniques have triggered interest in multicore optical fibers to increase spatial density by building several guiding cores in a single fiber of the same diameter. Record transmission capacity of 1.01 petabit-per-second (10e15 bits/s) was demonstrated in 2012 by Japanese researchers at NTT using such a 12-core fiber. The research program will transpose this achievement to optical interconnects by developing customized multicore fiber, active and passive coupling devices, circuits, and architectures to replace the less space-efficient fiber ribbons currently used in the industry. The new optical interconnects will be able to carry independent channels in several tens of guiding cores, thus permitting near terabit/s capacity over a single fiber using 25-Gb/s VCSELs. Our group is already equipped with the necessary design tools, has contributed to the concept of SDM for optical interconnects, has set key Canadian and international collaborations for multicore fiber sourcing, and benefits from world-class laboratory facilities. **The proposed research program will contribute to the development of synergies between electronic and photonic integration technologies, which will result in revolutionary changes in photonics manufacturing that will dramatically change design principles and fundamentally improve performance. Canada is a world leader in Information and Communications Technologies (ICT). The proposed research directly supports Canada's strategic priorities in ICT Device and Systems (as stated in NSERC selected Target Areas and Research Topics), by the development of technologies that will dramatically increase the spatial density of optical links.