NeTS: Small: Beyond Separate-Then-Centralize: A Cellular Operating System to Optimize Software-Defined 5G Wireless Networks

现有的无线网络依赖于封闭且僵化的架构，这往往会限制和推迟新无线网络技术的采用。因此，最近引入了软件定义联网(SDN)的概念，以简化网络控制，并使引入和部署新的应用和服务变得更容易。然而，到目前为止，现有的SDN架构远远不能满足下一代无线网络的要求。该项目为下一代无线蜂窝网络提出了一种完全不同的SDN方法。其核心是基于跨层优化理论，为下一代蜂窝网络开发软件定义的无线组网的原则性理论和实践。该项目有望弥合SDN和分布式网络优化/控制之间的差距；导致不依赖任何集中控制实体的简化控制平面；最后，大大减轻网络工程师的负担，他们将能够在不深入了解分布式优化概念的情况下控制和管理网络运营。本项目将研究无线蜂窝操作系统(WiCOS)的核心构建原则。WiCOS将为网络设计人员提供一组抽象，隐藏网络操作的低层细节及其分布式实施的细节，从而为网络设计人员提供在高级别抽象网络功能的集中视图。基于这种抽象表示，WiCOS将采用编写在网络高层视图上的集中控制程序，并基于优化理论自动生成分布式跨层控制程序，这些程序由每个单独的网元在无线硬件的抽象、公共表示上在网络边缘执行。这些目标将通过执行四项研究任务来实现：WiCOS架构设计；自动网络控制问题分解；可编程协议栈(PPS)设计；以及原型和演示。该项目将通过集中于协调网络、半协调网络和非协调网络三种情况，提供拟议的WiCOS框架的演示，并为每种情况制定详细的演示。

Existing wireless networks rely on closed and inflexible architectures that often limit and delay the adoption of new wireless networking technologies. The notion of software defined networking (SDN) has therefore been recently introduced to simplify network control and to make it easier to introduce and deploy new applications and services. However, to date, existing SDN architectures are far from fulfilling the requirements of next-generation wireless networks. This project proposes a radically different approach to SDN for next-generation wireless cellular networks. At the core, it will try to develop a principled theory and practice of software-defined wireless networking for next generation cellular networks based on cross-layer optimization theory. The project is expected to bridge the gap between SDN and distributed network optimization/control; lead to simplified control plane that does not rely on any centralized control entity; finally, considerably ease the burden of network engineers that will be able to control and manage the network operations without in-depth understanding of distributed optimization concepts.This project will study the core building principles of a Wireless Cellular Operating System (WiCOS). WiCOS will provide the network designer with a set of abstractions hiding the low-level details of the network operations as well as details of their distributed implementation, thus providing the network designer with a centralized view abstracting the network functionalities at a high level. Based on this abstract representation, WiCOS will take centralized control programs written on a high-level view of the network and automatically generate distributed cross-layer control programs based on optimization theory that are executed at the network edge by each individual network element on an abstract, common representation of the radio hardware. The goals will be accomplished by executing four research tasks: WiCOS architecture design; automated network control problem decomposition; programmable protocol stack (PPS) design; and prototyping and demonstration. The project will provide demonstrations of the proposed WiCOS framework by concentrating on three scenarios, i.e., coordinated networks, semi-coordinated networks and uncoordinated networks, and develop detailed demonstrations for each scenario.