SHF: EAGER: Addressing Resource Scarcity via Distributed Mobile Services

小型计算设备(例如智能手机、可穿戴设备等)具有一组硬件资源，程序员经常发现这些资源不足以实现应用程序的业务逻辑。例如，在没有GPS传感器的设备上运行时，无法确定移动应用程序的地理坐标。现有的软件工程方法无法解决移动设备资源稀缺的问题，除非求助于基于云的执行或要求附近的设备共享可执行代码。该项目将为分布式移动服务创建概念验证软件体系结构和编程抽象：附近的设备使用彼此的资源，而不必接受来自不可信各方的代码来执行。这些技术可以立即增强现有的移动计算生态系统。随着物联网(IoT)进入计算的主流，该项目也可以帮助解决该领域的资源稀缺问题。将代表人数不足的群体纳入拟议的研究活动将为扩大参与提供一个有效的途径，这与PI在这一领域以前的活动是一致的。该项目解决了为资源有限的设备开发软件的挑战。尽管附近的计算设备可能拥有它们愿意共享的缺失资源，但高度动态化、缺乏信任和极端的异构性阻碍了这种共享。预期的贡献有三个方面：(1)以移动服务市场(MSM)的概念为中心的软件体系结构，其利用附近设备的资源，而不使这些设备共享可执行代码；(2)用于静态选择所需服务、配置其设备约束并使运行时动态地选择附近设备来执行该服务的编程模型；该模型将被实现为一种特定于领域的语言，专为跨异类、邻近的移动设备的服务约束和通信领域量身定做；(3)对所创建的技术进行严格评估，这些技术通过将所提出的新概念应用于现实世界的移动应用程序和基准来实现这些技术。通过进行这项研究获得的见解将有助于确定如何组合静态和动态方法，以便能够静态地定义需求，同时以特定于环境的方式将它们与附近的特定设备动态匹配。

Small computing devices (e.g., smartphones, wearables, etc.) feature a set of hardware resources that programmers often find insufficient to implement an application's business logic. For example, a mobile application's geographic coordinates cannot be determined when running on a device without a GPS sensor. Existing software engineering approaches cannot address the resource scarcity of mobile devices, without either resorting to cloud-based execution or requiring nearby devices share executable code. This project will create a proof-of-concept software architecture and programming abstractions for distributed mobile services: nearby devices using each other's resources without having to accept code to execute from untrusted parties. These technologies can immediately enhance the extant mobile computing ecosystem. As the Internet of Things (IoT) is entering the mainstream of computing, this project can help address the resource scarcity in this domain as well. Involving underrepresented groups in the proposed research activities will provide an effective avenue for broadening participation, in line with the PI's prior activities in this realm.This project addresses the challenges of developing software for devices with limited resources. Although nearby computing devices may possess the missing resources they are willing to share, high dynamicity, lack of trust, and extreme heterogeneity stand on the way of such sharing. The expected contribution is three-fold: (1) a software architecture centered around the concept of the Mobile Service Market (MSM), which leverages the resources of nearby devices, without having the devices share executable code, (2) a programming model to statically select a needed service, configure its device constraints, and have the runtime dynamically select a nearby device to execute the service; the model will be realized as a domain-specific language tailored for the domain of service constraints and communication across heterogeneous, nearby, mobile devices, (3) a rigorous evaluation of the created technologies that realize the proposed new concepts by applying them to real-world mobile applications and benchmarks. Insights gained from conducting this research will help determine how static and dynamic approaches should be combined to be able to define requirements statically, while matching them dynamically with specific nearby devices in a context-specific fashion.