NeTS: JUNO3: Cloud-Carrier Cooperation for Efficient and Ultra-Reliable Programmable Backbone Networks

随着下一代5G/6G网络的出现，主干网络需要发展以适应在延迟、带宽和可靠性方面有严格要求的异类云服务。基于僵化/单一架构的传统设计对单个运营商/提供商网络实现超可靠的端到端通信具有挑战性，尤其是在资源紧张(由于网络拥塞或地震、海啸等大规模灾难)期间，当云服务可能因运营商网络和云基础设施故障而严重中断时。在这种情况下，多个网络实体之间的合作，即。与依赖单个网络相比，运营商、数据中心提供商(DCP)、互联网服务提供商(ISP)等可以在快速高效的服务提供方面带来更大的好处。然而，当网络和云基础设施由不同的实体拥有时，由于监管政策，诸如详细的网络拓扑、故障/损坏情况、资源可用性等机密信息不会被披露。这需要网络运营商和服务提供商等多个利益相关者之间建立一个新的合作框架，以创建一个平台，在集成的网络-云生态系统中灵活利用可用的和分布式的网络和计算资源。这种创新将利用名为提供商中立交换(PNE)(类似于互联网交换点(IXP)或托管中心)的中立第三方调解人/经纪人的服务，在保护信息机密性的同时促进多个实体之间的合作。该项目将研究新的网络架构，并借助先进的网络可编程性，在可编程主干网络中实现可靠的云-运营商合作。这项研究将开发新的见解来改善美国的国家网络基础设施，并在资源紧张的时候为互联网用户提供关键通信和服务的高可用性方面的好处。这项国家科学基金会和日本国立信息通信技术研究所(NICT)的联合项目将在网络运营商和服务提供商之间建立一个新的合作框架，以实现高效和可靠的服务提供。云-运营商合作的研究主要有三个方面：一)云-运营商-运营商的合作策略。将研究网络运营商、互联网服务提供商和数据中心之间的新合作战略，以开发一个有效的框架，在第三方中介/实体PNE的帮助下，实现灵活的资源共享和可靠的端到端服务提供，同时确保不同实体之间的信息保密。Ii)物理层感知故障管理，实现可靠协作。将开发考虑物理层降级/断开的抢占式故障检测机制和针对级联故障的健壮性策略，以确保骨干网络中的可靠协作。Iii)云-运营商-运营商合作的可靠运营商/管理(C/M)平面设计。将为第三方实体PNE开发一个可靠的C/M平面架构，以促进各实体/域之间的合作，并将开发一个分散的端到端C/M平面，以实现网络和云服务的广泛整合。对建议方法的评估将使用仙台的NICT试验台进行。该奖项反映了NSF的法定使命，并通过使用基金会的智力优势和更广泛的影响审查标准进行评估，被认为值得支持。

With the emergence of next-generation 5G/6G networks, backbone networks need to evolve to accommodate heterogeneous cloud services with stringent requirements in latency, bandwidth, and reliability. Traditional designs, based on inflexible/monolithic architectures, are challenging for individual carrier/provider networks to achieve ultra-reliable end-to-end communication, especially during resource crunch (due to network congestion or large-scale disasters such as earthquake, tsunami, etc.), when cloud services can be severely disrupted due to failures in both carrier network and cloud infrastructures. In such cases, cooperation among multiple network entities, viz. carriers, Datacenter Providers (DCPs), Internet Service Providers (ISPs), etc., can bring greater benefit in terms of fast and efficient service provisioning vs. relying on individual networks. However, when network and cloud infrastructures are owned by different entities, confidential information, such as detailed network topologies, failure/damage scenario, resource availability, etc., is not disclosed, due to regulatory policies. This calls for a new cooperative framework among multiple stakeholders, such as network operators and service providers, to create a platform for flexible utilization of available and distributed network and compute resources within an integrated network-cloud ecosystem. Such innovation will exploit services of a neutral third-party mediator/broker, called Provider Neutral Exchange (PNE) (similar to Internet exchange points (IXPs) or colocation centers) to facilitate cooperation among multiple entities while preserving information confidentiality. This project will investigate novel network architectures, aided by advanced network programmability, to enable reliable cloud-carrier cooperation in programmable backbone networks. This research will develop new insights to improve US national network infrastructures and provide benefits for Internet users in terms of high availability of critical communications and services at times of resource crunch.This joint National Science Foundation and Japan National Institute of Information and Communications Technology (NICT) project will develop a novel cooperation framework among network carriers and service providers for efficient and reliable service provisioning. There are three main research areas of cloud-carrier cooperation: i) Cloud-Carrier-ISP cooperation strategies. Novel cooperation strategies among network carriers, ISPs, and DCPs will be studied to develop an efficient framework for flexible resource sharing and reliable end-to-end service provisioning with the aid of a third-party broker/entity, PNE, while ensuring information confidentiality among different entities. ii) Physical-layer-aware failure management for reliable cooperation. Pre-emptive failure detection mechanisms considering degradation/disconnection at the physical layer and robustness strategies against cascading failures will be developed to ensure reliable cooperation in backbone networks. iii) Reliable Carrier/Management (C/M) plane design for cloud-carrier-ISP cooperation. A reliable C/M-plane architecture for the third-party entity, PNE, will be developed to facilitate cooperation across various entities/domains and a decentralized end-to-end C/M plane will be developed to enable extended integration of network and cloud services. Evaluation of the proposed approaches will be performed using the NICT testbed in Sendai.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.