Overcoming Capacity and Energy Limits in Optical Communications

世界互联网基础设施对国家和全球经济的重要性日益增加，使国际贸易更加有效，并带来许多新的商业机会。因此，在过去十年中，全球网络上的数据流量一直以每年约40%的速度稳步增长，预计在可预见的未来，这一趋势不会有任何缓解。然而，通过明显和传统的措施来满足这种增加的需求变得不切实际，例如提高频谱效率和利用率，安装更多的光纤，以及增加数据中心的数量或规模，因为今天的电信基础设施已经估计占世界碳排放量的约2%（超过航空业！）。因此，使用现有技术进一步扩大产能可能会对环境产生重大影响。另一个关键问题是，数据中心的散热已达到单位体积的极限，这实际上意味着，如果使用当前的技术平台来升级容量，那么这将需要新的热管理策略，这可能会进一步增加功耗，成本和复杂性。显然，从长远来看，这种情况是不可持续的。目前的网络在容量和能源效率方面大多受到所使用的架构/技术的限制。传统上，这些又受到设计网络的理念的驱动，以最大限度地减少昂贵和相对不可靠的光学组件的使用，并利用电子产品的成熟度-在大规模生产时具有降低成本的潜力-来完成大部分信号路由，处理和传输损害缓解。然而，这种方法-考虑到前面讨论的能源规模限制-将无法持续向前发展，这表明这种方法需要改变。具体而言，将有必要增加网络中的光学器件的数量，以减少施加在电子器件上的负担。我坚信，过渡到更多的光学授权系统是不可避免的。这个建议的目的是调查采用我目前正在研究的新光学元件和子系统阵列的可能性（包括光梳发生器、注入锁定激光器和根据相位不同放大信号的设备）进入光网络，以允许缩放到更大的数据传输容量的方式，同时降低功耗和/或或更好的热管理特性。今天，存在与未来光通信相关的几个主要概念，它们都依赖于具有更高质量的光信号（例如，信噪比），或者严格控制携带独立数据流的光信号的相干特性（称为“超级信道”技术）。我目前的研究涉及开发高纯度（低噪声）光梳，允许严格控制相干性-有望提供必要的关键参数。在奖学金中，我计划将这项技术应用到光链路的产生和接收部分。较高信噪比的发射机以及检测侧的较低噪声和较高速度的数据解复用应允许扩展范围而无需额外的在线放大，并允许使用在相同频谱带宽内携带更多信息的调制格式。除了其他优点之外，在这两种情况下，每发送比特的能量都可以减少。然而，我不会把我的研究局限于光通信，我会研究其他领域，这些领域的研究结果可能也会有所帮助，例如，超精确的时间和频率传输我将与学术界和非学术界的合作伙伴密切合作。

The world's Internet infrastructure is of ever increasing importance to both the national and global economy, enabling ever more efficient international trade and a host of new business opportunities. As a result data traffic on the world's networks has steadily been growing at around 40% per year over the past decade and no respite in this trend is anticipated for the foreseeable future. However, it is becoming impractical to satisfy this increased demand through the obvious and traditional measures such as improving spectral efficiency and utilization, installing more optical fibres, and increasing the number or size of the data centres, since today's telecommunications infrastructure is already estimated to be responsible for about 2% of world carbon emission (more than the aviation industry!). Consequently further capacity scaling using the existing technology is likely to have a significant impact on the environment. Another key issue is that the heat dissipation in data centres has reached its limit per unit volume, which effectively means that if current technological platforms are used to upgrade capacity then this will require new strategies for thermal management, likely further increasing the power consumption, cost and complexity. Clearly such a situation is unsustainable in the longer term. Current networks are mostly limited - both in terms of capacity and energy efficiency - by the architectures/technologies used. These in turn have traditionally been driven by the philosophy of designing the network so as to minimize the usage of expensive and relatively unreliable optical components and exploiting the maturity of electronics - with its potential for cost reduction when mass produced - to do the majority of the signal routing, processing and transmission impairment mitigation. However, this approach - given the previously discussed energy scaling constraints - will not be sustainable moving forward, dictating a change in this approach. Specifically it will be necessary to increase the amount of optics in the network to reduce the burden placed upon the electronics. I strongly believe that the transition to more optically empowered systems is simply unavoidable. The aim of this proposal is to investigate the possibility of employing the array of new optical components and subsystems that I am currently researching (which includes Optical Comb generators, injection locked lasers and devices that amplify signals differently depending on their phase) into optical networks, in a manner that allows for scaling to larger data transmission capacities with a simultaneous reduction in power consumption and/or better thermal management characteristics. Today, there are several main concepts relevant to future optical communications, all of them relying on having higher quality optical signals (e.g., signal-to-noise ratio), or tight control of the coherence properties of optical signals carrying independent data streams (called 'superchannel' technologies). My current research deals with development of high purity (low noise) Optical Combs that allow for tight control of coherence - promising to give both the key parameters necessary. Within the Fellowship, I plan to implement this technology into the generation and reception part of optical links. Higher signal-to-noise ratio transmitters and lower-noise and higher speed demultiplexing of data at the detection side should allow for extended reach without the need for additional in-line amplification and allow the use of modulation formats carrying more information inside the same spectral bandwidth. Among other advantages, the energy-per-transmitted-bit can be reduced in both these cases. However, I will not limit my research to optical communications and will investigate other fields where the results might also be helpful - e.g., ultraprecise transfer of time and frequency. I will work in close collaboration with academic as well as non-academic partners.