Architectural alternatives for edge computing in the 5G ecosystem

宽带蜂窝网络的第五代（5G）技术标准承诺了基于无线电的蜂窝网络的前所未有的吞吐量，以及控制平面组件的高度灵活和以软件为中心的部署选项。这种灵活性有望减少端到端的通信延迟，并支持一种称为边缘计算的新型计算模式。5G网络可能会在传感器融合、工业自动化和自动驾驶汽车控制等许多有趣的应用领域取得重大进展。这个研究项目将有助于解决平台细节和应用程序设计之间的先有鸡还是先有蛋的问题，这是许多突破性技术开发所固有的。我们将游说应用程序开发人员了解应用程序设计中的模式，并系统地构建通用工作负载模型。这些工作负载模型可用于构建性能测试软件系统，该系统将用于评估开发平台技术，其结果将反馈到应用程序设计中。总的来说，在这个项目中获得的知识将帮助电信运营商、平台供应商和应用程序设计人员构建更高效和有效的服务。

The fifth generation (5G) technology standard for broadband cellular networks promises unprecedentedthroughput capacity of radio-based cellular networks coupled with highly flexible and software-focuseddeployment options for control plane components. This flexibility is expected to reduce end-to-endcommunication delays and support a novel compute pattern termed edge computing. There is a plethora ofinteresting application areas that will likely see significant advances resulting from 5G networking, such assensor fusion, industry automation, and autonomous vehicle control. This research project will help solve thechicken-and-egg problem between platform details and application design that is inherent to manygroundbreaking technology developments. We will canvass application developers to understand patterns inapplication design and systematically build general workload models. These workload models can be used tobuild performance testing software systems that will be used to assess development platform technologies, theresults of which will feed back into application design. Overall, the knowledge gained in this project will helptelecommunication operators, platform vendors, and application designers to build more efficient and effectiveservices.