Collaborative Research: IRES Track I: Wireless Federated Fog Computing for Remote Industry 4.0 Applications

该项目的总体目标是通过与位于西班牙马德里的世界级研究机构IMDEA网络研究所的国际研究合作，为美国学生创造独特的培训和研究机会。该项目的研究计划让学生参与无线网络、边缘/雾计算和机器学习（ML）交叉领域的跨学科前沿研究。拟议的项目将使美国团队能够建立一个无线联合雾（WFF）平台，该平台具有鲁强性和可扩展性，可适应偏远地区不同类型的工业4.0应用，这些地区缺乏云资源的访问，即它提供了访问无限云资源的错觉。为此，在项目期间将开发一个跨层通信计算框架，并为以下方面提供新颖的解决方案：实时容器迁移；深度学习（DL）划分与逼近；以及无服务器工作流的开发和部署。该项目的预期结果将是以下工件：(1)跨自治雾系统无缝迁移集装箱化服务的方法；(2)跨WFF的微服务工作流延迟感知分区方法；(3) WFF上用于声明性和可靠工作流开发（工作流即代码）的无服务器框架；(4)基于不确定WFF的鲁棒延迟感知分布式深度学习推理方法；(5)使用近似推理在本地执行敏感DL服务的方法。该项目将为美国学生提供在世界一流研究机构获得国际跨学科经验的独特机会。每年，研究生和本科生将通过明确的程序选择访问IMDEA的全球计算研究小组，为期8周。该项目将使学生在一所著名的研究机构接触通信计算系统的分析和实践挑战，并培养他们成为具有世界一流技能的全球参与和文化劳动力。项目成果将由一名专业评价员每年进行评价，并在随后几年中予以加强。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。

The overarching goal of this project is to create unique training and research opportunities for U.S. students through an international research collaboration with the IMDEA Networks Institute, a world-class research institution in Madrid, Spain. The project's research plan engages students in an interdisciplinary cutting-edge research at the intersections of wireless networking, edge/fog computing, and machine learning (ML). The proposed project will enable the U.S. team to build a wireless federated fog (WFF) platform that is robust and scalable in accommodating different types of Industry 4.0 applications in remote areas where there is shortage of access to cloud resources, i.e., it offers the illusion of accessing infinite cloud-like resources. To this end, a cross-layer communications-computing framework will be developed during the project years with novel solutions for: live container migration; deep learning (DL) partitioning and approximation; and serverless workflow development and deployment. The anticipated outcomes of the project will be the following artifacts: (1) Methods for seamless migration of containerized services across autonomous fog systems; (2) Methods for latency-aware partitioning of micro-service workflows across WFF; (3) A serverless framework for declarative and reliable workflow development (workflow as a code) on WFF; (4) Methods for robust and latency-aware distributed DL inference atop uncertain WFF; (5) Methods using approximate inference for on-premises execution of sensitive DL services. This project will provide a unique opportunity for U.S. students to gain international cross-disciplinary experience at a world-class research institute. Every year, graduate and undergraduate students will be selected via a well-defined procedure to visit the Global Computing research group at IMDEA for 8 summer weeks. The project will expose the students to both analytical and practical challenges of the communication-computing systems at a reputed research institution, and train them to become a globally-engaged and cultured workforce with world-class skills. The project outcomes will be evaluated annually by a professional evaluator, and enhanced during the subsequent years.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.