CI-EN: ORBIT GEN 3 - Enhancing the ORBIT Testbed with LTE and Cloud Radio Processing

该项目旨在对Orbit(下一代无线网络开放接入测试床)测试床进行第三代设备升级，该测试床自2005年以来一直由罗格斯大学作为社区资源运营。轨道试验台是研究人员通过互联网远程访问的，为进行无线网络实验提供了一个灵活、可扩展和可重复使用的平台。轨道降低了无线电和无线技术领域的实验障碍，从而提高了该领域的教育和研究生产率。该项目的目标是扩展试验台，以纳入两个关键的新功能：(1)LTE(长期演进)无线电接入，以支持对未来移动数据服务的现实评估，以及(2)“云无线电”处理，以实现对新兴的“5G”无线电接入技术的实验研究。拟议的试验台增强功能将通过促进对新兴无线电技术和网络架构(如动态频谱接入、协作MIMO(多输入多输出)和异类蜂窝网络)的评估，帮助加快无线/移动技术开发的步伐。更具体地说，该项目中将增加的LTE和无线电云能力将使提高无线系统容量的技术研究成为可能，有助于解决随着移动数据使用量持续指数增长而出现的频谱稀缺这一重要的社会问题。首先，无线电网仿真器和室外轨道园区网络都将进行升级，以在现有WiMax功能的基础上加入LTE。LTE正在全球范围内迅速部署在4G蜂窝系统中，使研究界能够使用这种接入技术进行现实的移动网络试验，这一点非常重要。LTE能力将通过改装商业基站以通过轨道管理框架(OMF)进行控制，从而在室外轨道网络中增加，而室内节点上的LTE将通过在现有SDR平台上运行的软件来实施。其次，无线电网格的后端将通过基于FPGA和CPU的“软件无线电云”的独特组合进行升级，这将使处理速度提高两个数量级。所提出的无线电云被设计为一个分层组织的高性能系统，它包括基于快速CPU的服务器、FPGA协处理器和瘦客户端软件定义的无线电节点，所有这些都通过快速和可编程的交换背板连接在一起。该系统将包括近16个计算服务器刀片(每个刀片的额定计算能力为700 GIPS)、一个基于FPGA的大型中央协处理器阵列、总共约48个软件定义无线电(SDR)客户端节点和用于连接的约128 x 10Gbps OpenFlow交换机端口。将启用的实验示例包括LTE-WiFi互通、基于LTE的移动云服务、宽带频谱感知和动态频谱接入算法、大规模MIMO(多输入多输出)和协作PHY、蜂窝云RAN(无线电接入网络)和虚拟无线网络。LTE功能将在项目的第一年发布，无线电云的第一个版本将在第二年发布，随后将在第三年发布更新版本。

This project is aimed at a third-generation equipment upgrade for the ORBIT (Open Access Testbed for Next Generation Wireless Networking) testbed which has been operated by Rutgers University as a community resource since 2005. The ORBIT testbed, which researchers access remotely over the Internet, provides a flexible, scalable and reproducible platform for conducting wireless network experiments. ORBIT lowers the barrier for experimentation in the area of radio and wireless technology and thus improves education and research productivity in the field. The goal of this project is to extend the testbed to incorporate two key new capabilities: (1) LTE (Long Term Evolution) radio access, to support realistic evaluation of future mobile data services, and (2) "cloud radio" processing to enable experimental studies of emerging "5G" radio access technologies. The proposed testbed enhancements will help accelerate the pace of wireless/mobile technology development by facilitating evaluation of emerging radio technologies and network architectures such as dynamic spectrum access, cooperative MIMO (multiple input multiple output) and heterogeneous cellular networks. More specifically, the LTE and radio cloud capabilities to be added in this project will enable the study of techniques for enhancing wireless system capacity, helping to address the important societal problem of spectrum scarcity as mobile data usage continues to grow exponentially.The ORBIT upgrade proposed here involves two major enhancements to the testbed. First, both the radio grid emulator and the outdoor ORBIT campus network will be upgraded to incorporate LTE in addition to the existing WiMax capability. LTE is rapidly being deployed in 4G cellular systems worldwide, and it is important to enable the research community to use this access technology for realistic mobile network experiments. LTE capability will be added to the outdoor ORBIT network by retrofitting a commercial base station to be controllable through the ORBIT management framework (OMF), while LTE on indoor nodes will be implemented in software running on available SDR platforms. Second, the radio grid's backend will be upgraded with a unique combination of FPGA and CPU based "software radio cloud" that will increase processing speeds by two orders-of-magnitude. The proposed radio cloud is designed as a hierarchically organized high-performance system which includes fast CPU-based servers, FPGA co-processors and "thin-client" software-defined radio nodes all connected together by a fast and programmable switching backplane. The system will include nearly 16 compute server blades (each with rated computing capacity of 700 GIPS), a large FPGA-based centralized co-processor array, a total of about 48 software defined radio (SDR) client nodes and about 128 x 10Gbps OpenFlow switch ports for connectivity. Examples of experiments that will be enabled include LTE-WiFi interworking, LTE-based mobile cloud services, wideband spectrum sensing and dynamic spectrum access algorithms, massive MIMO (multiple input multiple output) and cooperative PHY, cellular cloud RAN (radio access network) and virtual wireless networks. LTE capabilities will be released during year 1 of the project, and a first version of the radio cloud will be released in year 2, followed by an updated version in year 3.