Collaborative Research: WiFiUS: Heterogeneous Resource Allocation for Hierarchical Software-Defined Radio Access Networks at the Edge

无线蜂窝基础设施正变得密集和异构。软件定义网络为加强网络规划和管理带来了新的机遇。当应用于地理上大规模的无线接入网络时，由于高度集中化，软件定义网络存在可扩展性和生存性问题。本课题设计了一种分层的软件定义无线接入网体系结构。相对于软件定义无线接入网络的平面版本，该方案通过探索局部性降低了控制开销和系统延迟，从而增强了系统的可扩展性。此外，分层设计的树形结构可以方便地处理单点故障。该计划首先研究分层软件定义无线接入网络的体系结构设计、控制网络的协议设计、生存性、局部性和优化，然后是大规模弹性方法、基于机器学习的方法和基于匹配理论的方法，用于全网资源分配。智力上的优势源于不同技术的跨学科融合，包括软件定义网络、无线通信、软件定义无线电、机器学习和博弈论。该研究计划解决了大规模无线接入网络中干扰集群共存和弹性资源分配的挑战性问题，同时开发了理论上新颖的框架来解决这些挑战。本研究将通过探索软件定义网络的力量，为无线网络提供更简单、更有效的资源分配方案。这个变革性和跨学科的项目涉及网络架构设计、理论建模和分析以及量化性能效益的实验模拟的互补组合。研究成果将通过高质量的期刊文章和会议报告提供给研究界，这些成果可能被工业实体用于网络发展和未来的工业标准化。该项目将加强美国和芬兰在无线通信研究领域的合作。拟议的研究活动将通过课程开发和专题研讨会补充和丰富休斯顿大学和亚利桑那州立大学不断增长的博弈论和优化课程。将对有关领域的高技能人员进行培训，以执行拟议的研究任务。将特别努力吸引少数民族和代表性不足的群体，包括非洲裔美国人、西班牙裔和女研究生。此外，还计划开展K-12外展活动，以激励高中生学习科学和工程。

Wireless cellular infrastructure is becoming dense and heterogeneous. Software-defined networking has brought new opportunities for enhancing network planning and administration. When applied in geographically large-scale radio access networks, software-defined networking suffers from scalability and survivability problems due to high centralization. In this project, a hierarchical software-defined radio access network architecture is designed. In contrast to a flat version of software defined radio access networks, which have large control overhead and system latency for large scaled networks, the proposed scheme offers reduced control overhead and system latency by exploration locality, thereby enhancing system scalability. In addition, the tree structure of the hierarchical design can easily deal with single points of failure. The plan is to first investigate the architecture design of the hierarchical software-defined radio access network, the protocol design for the control network, survivability, and locality and optimization, followed by large-scale elastic approaches, machine learning based approaches, and matching theory based approaches for network-wide resource allocation. The intellectual merit originates from the interdisciplinary fusion of different technologies including software-defined networking, wireless communications, software-defined radio, machine learning, and game theory. The research plan addresses challenging issues of coexistence of interfering clusters and elastic resource allocation for a large-scale radio access networks, while developing theoretically novel frameworks to tackle these challenges. This research will lead to simpler and more efficient resource allocation schemes for wireless networks by exploring the power of software-defined networks. The transformative and interdisciplinary project involves a complementary mix of network architecture design, theoretical modeling and analysis, and experimental simulations quantifying performance benefits. The outcomes will be made available to the research community through high quality journal articles and conference presentations which may be used by industrial entities for network development and future industrial standardization. This project will strengthen collaboration in the research field of wireless communication between the United States and Finland. The proposed research activities will complement and enrich the growing curriculum on game theory and optimization at the University of Houston and Arizona State University through course development and special topic seminars. Highly skilled personnel in related areas will be trained in carrying out the proposed research tasks. Special efforts will be made to engage minority and underrepresented groups, including African-American, Hispanic and female graduate students. K-12 outreach is also planned so as to motivate high school students in science and engineering.