EAGER: Techniques for Deploying Mission Critical IoT Applications

传感、计算和通信设备的快速部署导致物联网（IoT）基础设施的显著增长，并且在诸如医疗、农业、环境、商业和国土安全等许多领域中的应用中越来越多地使用该基础设施。传感器网络、云计算和边缘计算领域为物联网的发展做出了重大贡献。然而，与边缘计算和传感器网络相比，物联网提出了一个更具挑战性的环境：特别具有挑战性的是任务关键型应用程序的部署，这些应用程序的性能可能会深刻影响业务/组织运营。该提案的目标是开发技术，以在物联网环境中部署具有严格要求的使命关键应用程序。这项建议将通过集中注意下列任务来解决部署关键任务系统的问题：（a）第一个重点是发展技术，使部署工作符合具体应用的需要。人们希望将物联网软件基础设计为通用的，以便它可以被广泛的应用程序使用。然而，通用软件可能不适合于满足使命关键应用的要求。这可能会迫使设计者手动开发专门用于特定应用程序的服务。该项目将开发能够自动优化通用软件的技术，以获得满足具体应用需要的定制版本。(b)第二个重点是开发组合技术，系统地利用应用程序之间的共性。由于可能有多个应用程序共享底层资源，因此利用这些共性对于满足严格的性能要求至关重要。将进行广泛的性能评估，以评估所提出的技术相对于手动开发的实现所产生的实现的性能。

The rapid deployment of sensing, computing and communication devices is resulting in significant growth in Internet of Things (IoT) infrastructure, and in increased use of this infrastructure for applications in many domains such as medical, agricultural, environmental, commercial, and homeland security. The area of sensor networks, and cloud and edge computing has contributed significantly to the development of IoT. However, IoT presents a more challenging environment as compared to edge computing and sensor networks: Specially challenging is the deployment of mission-critical applications - applications whose performance may deeply impact business/organization operations. The goal of this proposal is to develop techniques to deploy mission critical applications with stringent requirements in IoT environments. This proposal will address the problem of deploying mission-critical systems by focusing on the following tasks: (a) The first thrust will be development of techniques to customize deployments to meet the needs of specific applications. One would like to design the IoT software base to be generic so that it can be used by a wide range of applications. However, generic software may not be suitable to to meet the requirements of mission critical applications. This may force the designers to manually develop services specialized for specific applications. The project will develop techniques that can automatically optimize generic software to obtain customized versions to meet the needs of specific applications. (b) The second thrust is the development of compositional techniques that will systematically exploit commonalities between the applications. As there may be multiple applications that share the underlying resources, exploiting such commonalities will be essential to meet stringent performance requirements. Extensive performance evaluation will be conducted to evaluate the performance of implementations generated by the proposed techniques with respect to manually developed implementations.