CNS Core: Small: Core Scheduling Techniques and Programming Abstractions for Scalable Serverless Edge Computing Engine

CNS Core：小型：可扩展无服务器边缘计算引擎的核心调度技术和编程抽象

• 批准号: 2322919
• 负责人: Liting Hu
• 金额: $ 60万
• 依托单位: University of California-Santa Cruz
• 依托单位国家: 美国
• 项目类别: Standard Grant
• 财政年份: 2024
• 资助国家: 美国
• 起止时间: 2024-01-01 至 2026-12-31
• 项目状态: 未结题
• 来源: https://www.nsf.gov/awardsearch/showAward?AWD_ID=2322919&HistoricalAwards=false
• 关键词: CNS; Core; Small; Scheduling; Techniques

The proliferation of 5G and beyond facilitates the advancement of next-generation technologies, including smart cities, self-driving cars, online video gaming, virtual reality, and augmented reality. This necessitates a re-evaluation of how these services are characterized and deployed. Serverless computing is an emerging paradigm, referring to a software architecture where an application is decomposed into triggers (also called events) and actions (also called functions), and there is a platform that provides seamless hosting and execution environment, making it easy to develop, manage, scale, and operate them. This project aims to build a next-generation serverless edge computing engine that empowers a vast number of distributed edge applications, such as data analytics, edge AI, and media streaming, to run efficiently at the edge through the Function-as-a-Service model. This project breaks the traditional abstractions and redefines new abstractions in the scheduling layer and storage layer that collectively deliver a scalable serverless edge computing engine. First, a full decentralized scheduling architecture is proposed, which dramatically improves the scalability of the proposed serverless edge computing engine. Second, an active object store abstraction is proposed, which is used for storing and sharing application states in a user-customizable manner. Third, the proposed serverless edge computing engine is implemented on top of the open-source software stacks. The evaluation is multi-pronged and includes micro-benchmarks for component testing and real-world applications for overall system testing. The results of the research are integrated into the undergraduate and graduate systems courses. The source code, datasets, tools, techniques, and new course materials developed in this research will be made publicly available.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.

5G及以后的普及促进了下一代技术的发展，包括智能城市、自动驾驶汽车、在线视频游戏、虚拟现实和增强现实。这就需要重新评价这些服务的特点和部署方式。无服务器计算是一种新兴的范例，指的是一种软件架构，其中应用程序被分解为触发器（也称为事件）和动作（也称为函数），并且有一个平台提供无缝的托管和执行环境，使其易于开发，管理，扩展和操作它们。 该项目旨在构建下一代无服务器边缘计算引擎，使大量分布式边缘应用程序（如数据分析，边缘AI和媒体流）通过功能即服务模型在边缘高效运行。该项目打破了传统的抽象，并在调度层和存储层中重新定义了新的抽象，共同提供了一个可扩展的无服务器边缘计算引擎。首先，提出了一个完全分散的调度架构，这大大提高了所提出的无服务器边缘计算引擎的可扩展性。其次，主动对象存储抽象的建议，这是用于存储和共享应用程序的状态，在用户可定制的方式。第三，所提出的无服务器边缘计算引擎是在开源软件栈之上实现的。评估是多管齐下的，包括组件测试的微基准和整体系统测试的实际应用程序。研究的结果被整合到本科和研究生系统课程。本研究中开发的源代码、数据集、工具、技术和新课程材料将公开提供。该奖项反映了NSF的法定使命，并通过使用基金会的知识价值和更广泛的影响审查标准进行评估，被认为值得支持。