NeTS NEDG: Dynamic Spectrum Access for Availability and Reliability

历史上的静态频谱分配导致人为的频谱稀缺，使得未来的无线网络没有可用的频谱。动态频谱接入是打破这种稀缺性的理想解决方案，可以为新的无线网络提供频谱。但是，如果不提供适当的可靠性保证，动态频谱接入对于许多网络和业务来说是不可接受的。因此，动态频谱接入不应仅仅着眼于提高频谱利用率，而应提供可靠的频谱使用，满足网络的个性化需求。该研究开发了SAFIRE，这是一种用于动态频谱访问的健壮架构，可为大型无线网络提供可靠和高效的频谱使用。本研究对无线网络设计者和服务提供商来说具有重要的实用价值，他们需要依赖可用和可靠的频谱接入来部署和推进他们的网络。认识到频谱利用率和可靠性之间的基本权衡，本研究的重点是有效的算法，以满足个别网络？可靠性要求，同时提高频谱利用率。研究人员将开发统计机制，根据可靠性要求主动调节频谱需求，并应用分布式协调和跨层设计，快速调整频谱分配和应用模式，以适应网络和频谱动态。为了支持大规模动态网络，本研究侧重于以最小管理开销的计算效率机制。这些进步将有助于吸引大型商业利益进入动态频谱系统，促进其广泛采用。最后，研究人员将把这项研究作为开发跨软件和硬件层的跨学科计算机科学培训的基础。

Historical static spectrum allocation has led to an artificial spectrum scarcity, leaving no usable spectrum for future wireless networks. Dynamic spectrum access is the ideal solution to break such scarcity and make spectrum available to new wireless networks. However, without providing proper reliability guarantees, dynamic spectrum access is unacceptable to many networks and services. Thus, instead of focusing solely on improving spectrum utilization, dynamic spectrum access should provide reliable spectrum usage that meets network's individual needs. This research develops SAFIRE, a robust architecture for dynamic spectrum access that provides reliable and efficient spectrum usage to large wireless networks. This research holds great practical values for wireless network designers and service providers who rely on available and reliable spectrum access to deploy and advance their networks.Recognizing the fundamental tradeoffs between spectrum utilization and reliability, this research focuses on efficient algorithms to meet individual networks? reliability requirements while improving spectrum utilization. The investigators will develop statistical mechanisms to proactively regulate spectrum demands based on their reliability requirements, and apply distributed coordination and cross-layer design to quickly adapt spectrum allocations and application patterns to network and spectrum dynamics. To support large-scale dynamic networks, this research focuses on computational-efficient mechanisms with minimum management overhead. These advancements will help attract large commercial interests into dynamic spectrum systems, facilitating their wide adoption. Finally, the investigators will use this research as a basis to develop interdisciplinary computer science training across both software and hardware layers.