CAREER: Enabling Edge-hosted Private Services via Unikernel-based Lightweight Virtualization

职业：通过基于 Unikernel 的轻量级虚拟化启用边缘托管的私有服务

• 批准号: 1943269
• 负责人: Yifan Zhang
• 金额: $ 48.52万
• 依托单位: SUNY at Binghamton
• 依托单位国家: 美国
• 项目类别: Continuing Grant
• 财政年份: 2020
• 资助国家: 美国
• 起止时间: 2020-09-01 至 2025-08-31
• 项目状态: 未结题
• 来源: https://www.nsf.gov/awardsearch/showAward?AWD_ID=1943269&HistoricalAwards=false
• 关键词: CAREER; Enabling; Edge; hosted; Private

In the era of Internet of Things (IoT), the existing mobile-cloud computing frameworks suffer from significant limitations, such as long communication latencies, unfriendly service deployment and communication interfaces for end devices, difficulty to scale with exploding processing demands, and incapability to enable large-scale personalized services. This CAREER project proposes Edge-hosted Private Service (EPS) to address these limitations. The approach of EPS is to deploy computation/communication middle points, each of which performs a certain service or optimization for an individual mobile user, at network edge between mobile devices and the cloud. Enabling EPSes in practice is challenging because existing technologies cannot meet the runtime requirements of EPS. The overarching goal of this project is to enable efficient EPS runtime through the adoption of unikernel-based lightweight virtualization. This research will advance the state of the art of unikernel-based virtualization and lay the foundation for EPSes by (1) realizing a scalable and backward-compatible unikernel-based embedded virtualization runtime framework, (2) providing real-time support for unikernel-based embedded virtualization, (3) enabling intuitive and EPS-friendly embedded unikernel development, and (4) supporting simple and practical unikernel-based service management. This research has the potential to contribute to the advancement of unikernels, which is a promising technology for resource-constrained embedded devices, by closing the gap between the state of the art and the practical requirements of adopting unikernels in real world. The research will achieve a series of novel and effective techniques underpinning unikernel-based virtualization, which can greatly benefit applications and services in the fields of edge computing and IoT. This project has concrete plans to integrate the proposed research with curriculum development, recruiting and supporting underrepresented students, as well as K-12 outreach activities.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.

在物联网（IoT）时代，现有的移动云计算框架存在通信延迟长、终端设备的服务部署和通信接口不友好、难以满足爆炸式处理需求的扩展、无法实现大规模个性化服务等显著局限性。这个CAREER项目提出了边缘托管私有服务（EPS）来解决这些限制。EPS的方法是在移动设备和云之间的网络边缘部署计算/通信中点，每个中点为单个移动用户执行特定的服务或优化。由于现有技术无法满足EPS的运行时要求，因此在实践中启用EPS具有挑战性。这个项目的首要目标是通过采用基于unikernel的轻量级虚拟化来实现高效的EPS运行时。本研究将通过(1)实现可扩展且向后兼容的基于unikernel的嵌入式虚拟化运行时框架，(2)为基于unikernel的嵌入式虚拟化提供实时支持，(3)实现直观且对eps友好的嵌入式unikernel开发，以及(4)支持简单实用的基于unikernel的服务管理，推动基于unikernel的虚拟化技术的发展，并为EPSes奠定基础。这项研究有可能通过缩小技术状态与在现实世界中采用unikernels的实际需求之间的差距，为unkernels的进步做出贡献，这是一项有前途的技术，用于资源受限的嵌入式设备。该研究将实现一系列新颖有效的技术，支持基于内核的虚拟化，这将极大地促进边缘计算和物联网领域的应用和服务。该项目有具体的计划，将拟议的研究与课程开发、招募和支持代表性不足的学生以及K-12外展活动相结合。该奖项反映了美国国家科学基金会的法定使命，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。

项目成果

EDDL: A Distributed Deep Learning System for Resource-limited Edge Computing Environment

• DOI: 10.1145/3453142.3491286
• 发表时间: 2021-12
• 期刊: 2021 IEEE/ACM Symposium on Edge Computing (SEC)
• 作者: Pengzhan Hao;Yifan Zhang