Servlerless Mobile Edge Computing

当前的移动网络不能支持需要低延迟或产生大量数据的下一代应用，这些数据可能会使运营商网络中的网络基础设施不堪重负。例如，帮助个人认知障碍的可穿戴应用程序，以及安全关键应用程序，如用于机场安检和监控的人脸识别应用程序。与移动或嵌入式设备上当前可用的计算资源相比，这些应用程序需要更多的计算资源。然而，在广域云上使用服务器也不是一种选择，因为这些应用程序需要较低的响应时间，或者涉及处理来自多个传感器和摄像头的大量数据。为了应对这些挑战，研究界和电信业正在探索将计算和存储能力添加到网络边缘的方法。这些方法被不同地称为云小程序、微型数据中心或雾，它们在传统的云体系结构中增加了一层服务器，这些服务器位于更靠近最终用户的位置，通常距离最终用户只有一跳。 这一提议提出了一种新的愿景，将边缘计算概括为在网络的地理范围内部署的分层云体系结构。这一愿景将支持可扩展处理，方法是沿着位于以下位置的一系列数据中心(规模不断增大)提供存储和计算 终端设备(例如智能手机、物联网设备)和传统的广域云数据中心。我们探索了使用一种新的云计算模型，即功能即服务(FAAS)作为部署机制，以便在网络地理跨度的分层云架构上运行应用程序。FAAS，也称为无服务器计算，是一种云计算方法，其中云提供商完全管理用于服务请求的基础设施，包括底层虚拟机或容器、主机操作系统和应用程序运行时。FAAS应用程序由一组轻量级无状态事件处理程序组成，这些处理程序在由云提供商管理的运行时环境中执行。FAAS处理程序的小尺寸和无状态特性使它们成为我们分层云计算愿景的理想候选者。

Current mobile networks are not able to support next generation applications that require low latency or that produce large volumes of data that can overwhelm the network infrastructure in a carrier network. Examples include wearable applications that assist individuals cognitive impairments, as well as safety critical applications, such as face recognition applications for airport security and surveillance. These applications require more computational resources than those that are currently available on mobile or embedded devices. The use of servers on the wide-area cloud, however, is also not an option as these applications require low response times, or involve processing of large volumes of data from multiple sensors and cameras. To address these challenges the research community and the telecommunications industry are exploring ways to add computation and storage capabilities to the edge of the network. These approaches, variously referred to as cloudlets, micro data centers, or fog, augment the traditional cloud architecture with an additional layer of servers that are located closer to the end user, typically one-hop away. This proposal puts forward a new vision that generalizes edge computing into a hierarchical cloud architecture deployed over the geographic span of a network. This vision will supports scalable processing by providing storage and computation along a succession of datacenters (of increasing sizes) positioned between the end device (e.g., smartphone, IoT appliance) and the traditional wide area cloud datacenter. We explore the use a new cloud computing model know as Function as a Service (FaaS) as the deployment mechanism for running applications on a hierarchical cloud architecture along the the geographic span of the network. FaaS, also known as Serverless computing, is a cloud computing approach in which the cloud provider fully manages the infrastructure used to serve request, including the underlying virtual machines or containers, the host operating system, and the application run time. FaaS applications are composed of a collection of light-weight stateless event handlers that execute on a run time environment that is managed by the cloud provider. The small size and stateless nature of FaaS handlers makes them ideal candidates for our hierarchical cloud computing vision.