Circuits and Systems Enabling Silicon-Photonics Signaling

计算和数据通信改变了我们工作、交流、社交、学习和教育的方式。维持和改善我们的生活方式，并解决我们作为一个社会面临的大多数具有挑战性的问题，例如开发个性化医疗或可持续管理能源资源，我们对能够以可管理的能力处理大量数据的超级计算机和数据中心的需求持续增长。在过去的十年中，这种需求推动了 CMOS 技术的指数级扩展以及多核和互连数据处理和通信的兴起。然而，这种激进的扩展现在正受到输入/输出 (I/O) 数据链路扩展速度放缓的威胁。例如，未来十年的超高性能计算系统预计需要通过 I/O 链路实现每秒 TB 级的聚合数据吞吐量。根据现有趋势，这意味着在未来十年内以 1/10 的功率提供 10 倍的 I/O 带宽，并提出了标准铜互连无法解决的真正挑战。****硅光子链路是一种有前途的替代方案，其中铜互连被光纤取代，硅兼容的光子器件由电子电路驱动进行数据传输。我们提出的研究的最终目标是用支持波分复用 (WDM) 的完全集成硅光子电光端到端链路取代传统的铜基电气链路。光学界的最新研究已经实现了硅光子器件的可行实现。我们提出的研究现在将通过在一个完整的 WDM 系统中共同设计电路和光子学来解决更大的问题。我们计划努力实现两项突破，我们认为这两项突破对于该技术成功取代纯电气 I/O 至关重要 - (1) 卓越的电源效率，以及 (2) 对 WDM 的支持。 WDM 具有在单根光纤中并行化多个通道的优点 - 这在铜缆中是不容易实现的。数字密集型集成电路将被设计用于驱动 CMOS 兼容的硅光子器件，以促进电光发射器和接收器的发展，同时解决变化、封装、功耗等问题。****世界各地的许多公司和大学都致力于传统的电气和光学互连技术。完整的高能效 WDM 硅光子集成系统将是一项颠覆性技术，有可能催生新公司、就业机会和技术，并可能为加拿大重建其在数据通信领域的地位带来变革。**

Computing and data communications have changed the way we work, communicate, socialize, learn and educate. To maintain and improve our lifestyle, and to solve most of the challenging problems we face as a society, e.g. developing personalized medicine or a sustainable management of energy resources, our need for super-computers and data-centers that can process massive amounts of data with manageable power continues to grow. This need has driven the exponential scaling of CMOS technology and the rise of multi-core and inter-linked data processing and communication in the last decade. However, this aggressive scaling is now being imperiled by the slower pace of scaling of Input/Output (I/O) data links. For example, ultra-high performance computing systems over the next decade are expected to require Terabytes per second of aggregate data throughput over I/O links. Based on the existing trend, this means providing 10 times the I/O bandwidth for 1/10th the power over the next decade, and presents a real challenge that cannot be addressed by standard copper interconnects.****Silicon-photonics link, where the copper interconnect is replaced with an optical fiber and silicon-compatible photonic devices are driven by electronic circuits for data transmission, is a promising alternative. The end goal of our proposed research is the replacement of conventional copper-based electrical links with fully-integrated silicon-photonics electro-optical end-to-end link supporting wavelength-division multiplexing (WDM). Recent research in the optical community has enabled viable implementations of silicon-photonics devices. Our proposed research will now address the bigger picture - by co-designing circuits and photonics in one complete WDM system. We plan to work on achieving two breakthroughs that we believe are critical to the success of this technology in replacing purely electrical I/Os - (1) superior power efficiency, and (2) support for WDM. WDM provides the benefit of parallelizing numerous channels in a single optical fiber - something that is not easily possible in a copper cable. Digital-intensive integrated circuits will be designed to drive CMOS-compatible silicon-photonic devices to facilitate the electro-optical transmitter and receiver, while addressing concerns like variations, packaging, power consumption, etc.****Numerous companies and universities across the world are engaged in conventional electrical and optical interconnect technology. A complete power-efficient WDM silicon-photonics integrated system will be a disruptive technology, and have the potential to herald new companies, jobs and technologies and be potentially transformative for Canada in re-establishing its position in data communications.**