Collaborative Research: NeTS: JUNO3: Leveraging Heterogeneous Programmable Data Planes for Security and Privacy of Cellular Networks, 5G & Beyond

保护“5G及以后”蜂窝网络对于支持来自移动的和物联网设备的不断增长的流量至关重要。蜂窝网络基础设施的重要部分正在基于软件的环境中实施。向分散的虚拟化蜂窝核心网络的转变可能导致攻击面增加和更大的脆弱性。 试图避免注意的缓慢攻击可能具有破坏性，因为它们不容易被检测到，并且通常需要终端主机安全中间盒的内存和计算能力来检测或防止它们。同样，攻击者试图通过窃听通信来侵犯用户隐私，这是不容易防止的，特别是在大规模的情况下。这些威胁使蜂窝用户和运营商容易受到攻击。这个美日联合项目旨在提供强大的安全监控和隐私保护解决方案，利用可编程交换机的高速、可编程网络接口卡的增强功能以及终端主机服务器的内存/计算能力。通过利用这些数据平面组件的优势，该项目将开发一个高效和高性能的蜂窝网络安全解决方案。为了实现这一目标，这一美日联合项目将追求由美国和日本的PI合作追求的技术任务。首先，该团队将设计一个异构数据平面框架，该框架将多个数据平面设备紧密结合在一起，用于网络功能处理。该方法将使用协同过滤系统，其中大部分流量仅由高速可编程交换机处理，这些交换机可以轻松提取聚合的粗粒度指标。可疑流量将被重定向到可编程网络接口卡，或根据需要重定向到主机，以进行进一步检查和指标收集。其次，该项目将开发蜂窝流量的实时监控，利用蜂窝核心网络作为关键Vantage。在蜂窝核心进行监控不仅可以有效地检测和阻止基于数据平面的攻击，还可以检测和阻止控制平面上的攻击。它处于能够在数据和控制平面状态之间进行关联的独特位置，以进一步改进检测安全攻击的现有方法。最后，该项目将设计隐私保护机制，确保用户在面对指纹攻击时的匿名性。该方法将利用流量变形技术，利用多层，可编程，异构数据平面框架的整个功能范围，以实现高速操作。通过为蜂窝网络用户及其流量提供强大的威胁预防和隐私保护，拟议的技术将产生重大的社会影响。该奖项反映了NSF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。

Securing “5G and beyond” cellular networks is critical to support the growing traffic from mobile and IoT devices. Significant parts of the cellular network infrastructure are being implemented on software-based environments. The shift to a disaggregated, virtualized cellular core network may result in an increased attack surface and greater vulnerability. Slow attacks, which attempt to avoid notice, can be damaging as they cannot be easily detected, and generally require the memory and computational capacity of end-host security middleboxes to detect or prevent them. Likewise, attackers seeking to violate user-privacy by eavesdropping on communication, cannot be easily prevented, especially at large scale. These threats leave both cellular users and operators vulnerable to attacks. This joint US-Japan project seeks to provide strong security monitoring and privacy protection solutions that exploit the high speed of programmable switches, the increased capabilities of programmable network interface cards, and the memory/computational capacity of end-host servers. By leveraging the strengths of each of these data plane components, the project will develop an efficient and performant cellular network security solution. To achieve this goal, this joint US-Japan project will pursue technical tasks that will be collaboratively pursued by the PIs based in the US and Japan. First, the team will design a heterogeneous data plane framework that cohesively combines multiple data plane devices for network function processing. The approach will use a collaborative filtering system, where most of the traffic is processed only by high-speed programmable switches that can easily extract aggregated, coarse-grained metrics. Suspicious traffic will be redirected to programmable network interface cards, or the host as necessary, for further inspection and metrics collection. Second, the project will develop real-time monitoring of cellular traffic, leveraging the cellular core network as a key vantage point. Monitoring at the cellular core can not only effectively detect and thwart data plane-based attacks, but also those on the control plane. It is in the unique position of being able to correlate between data and control plane state to further improve upon existing approaches to detect security attacks. Finally, the project will design privacy protection mechanisms that ensure anonymity of users in the face of fingerprinting attacks. The approach will leverage traffic morphing techniques that leverage the entire range of capabilities of a multi-tier, programmable, heterogeneous data plane framework, to enable high-speed operation. The proposed techniques will have significant societal impact by providing strong threat prevention and privacy preservation for cellular network users and their traffic.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.