CRII: NeTS: A Proactive Perspective on Preventing Network Inference: Shifting from Optimized to Dynamic Wireless Network Design

该项目将设计策略，以挫败针对从无线网络推断信息的恶意努力。新的主动反推理策略将被开发和测试，以使网络更具动态性和更少的推理敏感性。该项目涉及对本科生和研究生进行网络推理/反推理理论和实验活动的培训，预计将对关键任务无线网络的安全设计产生重大影响。网络推理是从网络测量中推断网络统计和属性信息的有效方法。网络推理已经实现了广泛的应用，例如网络监视、管理和诊断;然而，在某些情况下，例如用于军事应用的无线网络，重要的是屏蔽这些信息以防止将推理用于恶意意图。到目前为止，无线网络中的反推理策略一直被忽视和探索不足。这项工作的中心目标是通过探索主动策略的基本方面来解决当前的理解差距，这些策略将使广泛的反推理应用成为可能。具体目标是：（1）建立一个理论框架，分析增强网络动力学以实现反推理的主动策略的行为;（2）揭示基于网络动力学的反推理策略的影响和成本之间的基本关系;以及（3）为网络应用实现有效的反推理设计。为了实现这些目标，将利用欺骗流量和路由改变策略来诱导网络中更高的动态性，以使推理不准确。此外，这些战略的成本效益影响将通过从优化转向动态网络设计进行检查。

This project will design strategies to thwart malicious efforts targeting the inference of information from wireless networks. New proactive anti-inference strategies will be developed and tested to make networks more dynamic and less inference-susceptible. The project involves the training of undergraduate and graduate students in network inference/anti-inference theoretical and experimental activities, and is expected to significantly impact the security design of mission-critical wireless networks.Network inference is an effective way to infer the information of network statistics and properties from network measurements. Network inference has enabled a wide range of applications, such as network surveillance, management and diagnosis; however, in certain circumstances, such as wireless networks for military applications, it is important to mask this information to prevent the use of inference for malicious intent. To date, strategies for anti-inference in wireless networks have been overlooked and under-explored. The central goal of this effort is to address the current gap in understanding by exploring the fundamental aspects of proactive strategies that will enable a wide range of anti-inference applications. Specific objectives are to (1) develop a theoretical framework to analyze the behaviors of proactive strategies that enhance network dynamics to achieve anti-inference; (2) reveal the fundamental relationship between the impacts and costs of network dynamics based anti-inference strategies; and (3) implement efficient anti-inference design for network applications. To accomplish these objectives, both deception traffic and routing changing strategies will be utilized to induce higher dynamics in the network to make inference inaccurate. In addition, the cost-benefit impact of these strategies will be examined by shifting from optimized to dynamic network design.