CAREER: a Disaggregated and Distributed Hardware and OS Framework for Next-Generation Datacenters

职业：下一代数据中心的分解和分布式硬件和操作系统框架

• 批准号: 1846101
• 负责人: Yiying Zhang
• 金额: $ 51.66万
• 依托单位: Purdue University
• 依托单位国家: 美国
• 项目类别: Continuing Grant
• 财政年份: 2019
• 资助国家: 美国
• 起止时间: 2019-06-01 至 2020-07-31
• 项目状态: 已结题
• 来源: https://www.nsf.gov/awardsearch/showAward?AWD_ID=1846101&HistoricalAwards=false
• 关键词: CAREER; Disaggregated; Distributed; Hardware; OS

We live in the big data era. The abundance of digital information enables us to have a better understanding of the world and to make intelligent decisions. However, the volume and variety of data and its generation speed raise many profound new challenges. Datacenters host and process data, and they play a key role in the success of all the applications. This project aims to answer a simple yet important question: how shall we build next-generation datacenters that are cost-effective and deliver good performance?For decades, the unit of deployment, operation, and failure in datacenters has been a computer server, that contains all the hardware resources that are needed to run a user program. However, this is too coarse for today's applications; it restricts the effectiveness of datacenter resource utilization, and limits the flexibility of hardware changes. As a result, there are huge energy and monetary wastes in today's datacenters and it is hard to incorporate new hardware innovations.This project intends to break the traditional server model into fine-grained hardware resources. Instead of packaging hardware devices in a server and connecting servers with a network, future datacenters will connect hardware devices directly to the network. Doing so makes it more flexible and cheaper to run applications and deploy new hardware. This project will build the software infrastructure for this new datacenter model. Through innovation, design, and development of the proposed infrastructure, this project will improve the flexibility and reduce the total cost of future datacenters.This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.

我们生活在大数据时代。丰富的数字信息使我们能够更好地了解世界并做出明智的决策。然而，数据的数量、种类及其生成速度提出了许多深刻的新挑战。数据中心托管和处理数据，它们在所有应用程序的成功中发挥着关键作用。该项目旨在回答一个简单但重要的问题：我们如何构建具有成本效益并提供良好性能的下一代数据中心？几十年来，数据中心的部署、操作和故障单元一直是计算机服务器，它包含运行用户程序所需的所有硬件资源。然而，这对于今天的应用来说太粗糙了；它限制了数据中心资源利用的有效性，并限制了硬件变更的灵活性。结果，当今的数据中心存在巨大的能源和金钱浪费，并且很难融入新的硬件创新。该项目旨在将传统的服务器模型打破为细粒度的硬件资源。未来的数据中心将不再将硬件设备封装在服务器中并将服务器与网络连接起来，而是将硬件设备直接连接到网络。这样做可以使运行应用程序和部署新硬件变得更加灵活、更加便宜。该项目将为这个新的数据中心模型构建软件基础设施。通过拟议基础设施的创新、设计和开发，该项目将提高灵活性并降低未来数据中心的总成本。该奖项反映了 NSF 的法定使命，并通过使用基金会的智力价值和更广泛的影响审查标准进行评估，被认为值得支持。