CNS: Core: Small: Elevating the Edge to be a Peer of the Cloud

CNS：核心：小型：将边缘提升为云的对等点

• 批准号: 1909346
• 负责人: Umakishore Ramachandran
• 金额: $ 50万
• 依托单位: Georgia Tech Research Corporation
• 依托单位国家: 美国
• 项目类别: Standard Grant
• 财政年份: 2019
• 资助国家: 美国
• 起止时间: 2019-10-01 至 2022-09-30
• 项目状态: 已结题
• 来源: https://www.nsf.gov/awardsearch/showAward?AWD_ID=1909346&HistoricalAwards=false
• 关键词: CNS; Core; Small; Elevating; Edge

Today, while mobile devices give users access to vast amounts of information, all such information flows from the Cloud. The Cloud denotes a data center endowed with hundreds-of-thousands of servers centrally located to achieve economies of scale. Increasingly, sensor devices (such as those found in the Internet of Things) are being deployed everywhere in the environment. These sensors produce data 24x7, which must quickly be transformed into actionable knowledge to better serve society. Unfortunately, sending these sensor streams to the Cloud for processing is suboptimal for several reasons including latency (i.e., time elapsed between sensing and actuation), saturation of the available network bandwidth (due to the enormity of the sensor streams), lack of actionable content (e.g., no action in front of a camera), and privacy concerns for the collected data. Instead of sending the sensor data to the Cloud, there is a need for having computational resources close to the sensors to quickly process the sensor streams and quench the uninteresting sensor streams at the source itself. Edge/Fog computing is a new paradigm that signifies such geo-distributed computational resources (e.g., compute capability located in an AT&T cell tower) to extend the Cloud's centralized utility computing model across the network edge, closer to where end-user devices and sensors operate. To fully unlock the power of Edge computing, this project proposes research to elevate the Edge to be a peer of the Cloud. To achieve this goal, we must develop new technologies that enable Cloud-like functionality across a large number of geo-distributed micro data centers instead of one large, centrally located data center. The facets of the proposed research include geo-distributed programming frameworks for facilitating the development of complex latency sensitive applications that span the Edge-Cloud computational continuum; geo-distributed data replication and consistency models that are resilient to coordinated power failures; and support for rapid dynamic deployment of application components while recognizing that both computational, networking, and storage resources are limited at the Edge. The proposed research will enable the rapid prototyping of applications of societal importance spanning healthcare, transportation, disaster recovery, and the like that utilize sensor platforms in the Internet of Things (IoT) to convert sensed information into actionable knowledge. When coupled with the proliferation of IoT platforms, the Edge has the potential to be the next disruptive technology, following in the footsteps of Cloud computing. In the Cloud computing revolution, the availability of powerful software frameworks enabled the development of complex applications such as search engines (e.g., Google, Bing), e-Commerce (e.g., eBay), and social networking (e.g., Facebook, Twitter). This proposal aims to spur the same revolution for Edge computing, thus unburdening application developers from the complexities of managing geo-distributed computational resources and freeing them to create next generation software. Planned activities involve in situ studies via collaboration with the Georgia Tech Police Department, which creates the potential for technology transfer and direct societal benefits beyond technical accomplishments.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.

如今，虽然移动设备让用户能够访问海量信息，但所有这些信息都来自云。云是指拥有数十万台服务器的数据中心，这些服务器集中分布在一起，以实现规模经济。传感器设备(如物联网中的传感器设备)正越来越多地部署在环境中的任何地方。这些传感器全天候产生数据，这些数据必须迅速转化为可操作的知识，以更好地服务于社会。遗憾的是，将这些传感器流发送到云端进行处理并不理想，原因包括延迟(即，从检测到启动之间经过的时间)、可用网络带宽饱和(由于传感器流的巨大容量)、缺乏可操作的内容(例如，在摄像头前没有操作)以及收集的数据的隐私问题。不是将传感器数据发送到云，而是需要有接近传感器的计算资源来快速处理传感器流，并在源本身抑制不感兴趣的传感器流。边缘/雾计算是一种新的范例，它意味着这种地理分布的计算资源(例如，位于AT&amp；T信号塔的计算能力)将云的集中式效用计算模型扩展到网络边缘，更接近最终用户设备和传感器的运行位置。为了充分释放边缘计算的力量，本项目提出了将边缘提升为云的同行的研究。为了实现这一目标，我们必须开发新技术，在大量地理位置分散的微型数据中心而不是一个位于中心位置的大型数据中心实现类似云的功能。建议研究的方面包括：地理分布式编程框架，用于促进跨边缘-云计算连续体的复杂延迟敏感型应用程序的开发；地理分布式数据复制和一致性模型，可恢复协调电源故障；以及支持快速动态部署应用程序组件，同时认识到边缘的计算、网络和存储资源都是有限的。拟议的研究将使医疗保健、交通运输、灾难恢复等领域具有社会重要性的应用程序能够快速成型，这些应用程序利用物联网(IoT)中的传感器平台将感知信息转换为可操作的知识。再加上物联网平台的激增，Edge有可能成为继云计算之后的下一项颠覆性技术。在云计算革命中，强大的软件框架的可用性使得复杂应用的开发成为可能，例如搜索引擎(例如Google、Bing)、电子商务(例如eBay)和社交网络(例如Facebook、Twitter)。这项提议旨在推动边缘计算的同样革命，从而使应用程序开发人员从管理地理分布的计算资源的复杂性中解脱出来，并将他们解放出来创建下一代软件。计划中的活动包括通过与佐治亚理工大学警察局的合作进行现场研究，这创造了技术转让和技术成就以外的直接社会效益的潜力。该奖项反映了NSF的法定使命，并通过使用基金会的智力优势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

eCloud: A Vision for the Evolution of the Edge-Cloud Continuum

eCloud：边缘云连续体演进的愿景

• DOI: 10.1109/mc.2021.3059737
• 发表时间: 2021
• 期刊: Computer
• 影响因子: 2.2
• 作者: Arulraj, Joy;Chatterjee, Abhijit;Daglis, Alexandros;Dhekne, Ashutosh;Ramachandran, Umakishore
• 通讯作者: Ramachandran, Umakishore

Creating a revolutionary academic program

创建革命性的学术项目

• DOI: 10.1145/3503915
• 发表时间: 2022
• 期刊: Communications of the ACM
• 影响因子: 22.7
• 作者: Ramachandran, Umakishore;Galil, Zvi
• 通讯作者: Galil, Zvi

MicroEdge: a multi-tenant edge cluster system architecture for scalable camera processing

• DOI: 10.1145/3528535.3565254
• 发表时间: 2022-11
• 期刊: Proceedings of the 23rd ACM/IFIP International Middleware Conference
• 作者: Difei Cao;Jinsun Yoo;Zhuangdi Xu;Enrique Saurez;Harshit Gupta;T. Krishna;U. Ramachandran

Where did my 256 GB go? A Measurement Analysis of Storage Consumption on Smart Mobile Devices

• DOI: 10.1145/3460095
• 发表时间: 2021-05
• 期刊: Proceedings of the ACM on Measurement and Analysis of Computing Systems
• 作者: Ashish Bijlani;U. Ramachandran;R. Campbell

ePulsar: Control Plane for Publish-Subscribe Systems on Geo-Distributed Edge Infrastructure

• DOI: 10.1145/3453142.3491271
• 发表时间: 2021-12
• 期刊: 2021 IEEE/ACM Symposium on Edge Computing (SEC)
• 作者: Harshit Gupta;Tyler C. Landle;U. Ramachandran