RINGS: Resilient Delivery of Real-Time Interactive Services Over NextG Compute-Dense Mobile Networks

实时交互（RTI）服务有望成为信息技术领域的“下一件大事”，它融合了传感、计算和通信领域，这些领域直到最近都是分开处理的。在传统系统中，这些构建块没有经过优化以相互协作，特别是没有进行实时优化，从而限制了可以提供的服务类型。rti需要将分布式数据流实时聚合到边缘/云计算服务器上，这些服务器可以在数据生成后立即处理数据。该项目的目标是开发一个通用的数学框架以及具体的算法，以提供具有保证延迟的RTI服务。研究结果使美国经济受益，使其能够实现更高效、更可靠、更有弹性的自动化方案，例如智能工厂和农场，以及改进的增强/虚拟现实。该项目的跨学科性质使从事该项目的学生受益；一项详细的扩大和增加代表性不足群体参与的计划加强了这项工作的影响。该项目为可靠和弹性实时交互（RTI）系统设计了通用数学框架和具体算法。首先，该项目探索了可靠地交付延迟关键服务的限制。可靠网络稳定区域的建立提供了在所有算法中最好的情况下网络可以支持多少RTI流量的界限。基于包生命周期的排队模型是李雅普诺夫漂移启发控制算法的基础，该算法不仅考虑了包延迟，而且考虑了能量消耗等成本。该项目的第二个重点是开发鲁棒的NextG网络控制算法，该算法可以适应动态网络状态。在通用Lyapunov漂移控制框架下，将前两个推力中的方法集成为实现可靠的RTI服务交付的途径，这是通过在高级无线研究平台（PAWR）社区测试平台上实现算法来验证的。该奖项反映了美国国家科学基金会的法定使命，并通过使用基金会的知识价值和更广泛的影响审查标准进行评估，被认为值得支持。

Real-time interactive (RTI) services are anticipated to be the ‘’next big thing” in information technology, merging the areas of sensing, computation, and communications that until recently have been treated separately. In traditional systems, these building blocks are not optimized to work with each other, and particularly not in real-time, constraining the type of services that can be offered. RTIs require real-time aggregation of distributed data streams onto edge/cloud compute servers that can process data as soon as it is generated. The goal of this project is to develop a general mathematical framework as well as concrete algorithms to provide such RTI services with guaranteed latencies. The results benefit the US economy by enabling more efficient, more reliable, and more resilient automation schemes, e.g., for smart factories and farms, as well as improved augmented/virtual reality. The interdisciplinary nature of the project benefits the students working on this project; a detailed plan for outreach and increasing participation of underrepresented groups strengthens the impact of the work.The project designs a general mathematical framework and concrete algorithms for reliable and resilient Real-time interactive (RTI) systems. In a first thrust, the project explores the limits of delivering latency-critical services reliably. The establishment of the reliable network stability region provides bounds on how much offered RTI traffic a network can support under the best of all algorithms. A packet-lifetime-based queuing model serves as the basis of Lyapunov-drift-inspired control algorithms that not only account for the packet delay but also takes cost such as energy consumption into account. A second thrust of the project develops robust NextG network control algorithms that can adapt to dynamic network states. Integration of the approaches in the first two thrusts under a common Lyapunov drift control framework is an avenue to enable reliable RTI service delivery, which is verified by the implementation of the algorithm on a Platform for Advanced Wireless Research (PAWR) community testbed.This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.