Abstract：
Cyber-physical systems (CPSs) based on space-ground integrated networks (SGINs) enable CPSs to break through geographical restrictions in space. Therefore, providing a test platform is necessary for new technical verification and network security strategy evaluations of SGINs. User behavior emulation technology can effectively support the construction of a test platform. Given the inherent dynamic changes, diverse behaviors, and large-scale characteristics of SGIN users, we propose user behavior emulation technology based on a cloud platform. First, the dynamic emulation architecture for user behavior for SGINs is designed. Then, normal user behavior emulation strategy driven by the group user behavior model in real time is proposed, which can improve the fidelity of emulation. Moreover, rogue user behavior emulation technology is adopted, based on traffic replay, to perform the security evaluation. Specifically, virtual Internet Protocol (IP) technology and the epoll model are effectively integrated in this investigation to resolve the contradiction between large-scale emulation and computational overhead. The experimental results demonstrate that the strategy meets the requirement of a diverse and high-fidelity dynamic user behavior emulation and reaches the emulation scale of 100,000-level concurrent communication for normal users and 100,000-level concurrent attacks for rogue users.
摘要：
基于空地集成网络（SGIN）的网络物理系统（CPS）使CPS能够突破空间中的地理限制。因此,为SGIN的新技术验证和网络安全策略评估提供测试平台是必要的。用户行为仿真技术能够有效支持测试平台的构建。鉴于内在的动态变化，不同的行为，和SGIN用户的大规模特征，提出了基于云平台的用户行为仿真技术。首先,设计了SGIN用户行为的动态仿真体系结构。然后，提出了群体用户行为模型实时驱动的正常用户行为仿真策略,这可以提高仿真的保真度。此外，采用流氓用户行为仿真技术，基于交通回放，执行安全评估。具体来说,本研究将虚拟互联网协议(IP)技术和epoll模型有效地集成在一起,以解决大规模仿真和计算开销之间的矛盾。实验结果表明，该策略满足多样化和高保真的动态用户行为仿真的要求，达到了正常用户10万级并发通信和流氓用户10万级并发攻击的仿真规模。
