支持多租户的网络测试床模拟流量标记和溯源模型  

Multi-tenant Network Testbed Flow Watermarking and Provenance Tracing Model

作　　者：丁嘉宁 张鹏[1,2] 杨嵘 刘俊朋[1,2] 刘庆云[1,2] 熊刚[1,2] 

机构地区：[1]中国科学院信息工程研究所信息内容安全技术国家工程实验室,北京100093 [2]中国科学院大学网络空间安全学院,北京100049

出　　处：《电子学报》2017年第8期1947-1956,共10页Acta Electronica Sinica

基　　金：国家自然科学基金(No.61402464);国家重点研发计划(No.2016YFB0801304)

摘　　要：为了在支持多租户的网络测试床中对模拟流量进行标记和溯源,提出了一种基于时间间隔的网络流水印模型,当生成模拟流量时,该模型首先把水印内容转换成0~1比特序列,然后将0~1比特序列转换成流中数据包发送的时间间隔从而实现对模拟流量的标记.当接收模拟流量时,该模型通过将流中数据包的时间间隔转换成0~1比特序列,进而获取对应的水印内容,从而实现模拟流量的溯源.理论分析表明,该模型能够抵御多种攻击手段,同时大量实验证明了该模型在不丢包情况和丢包情况下对模拟流量进行溯源的有效性.In order to label and trace the provenance of any simulated flowin multi-tenant network testbed,an interval-based flowwatermarking and provenance tracing model was proposed. When a simulation flowwas generated,this model first transformed the user’s watermarking content into 0-1 bit sequence and then sent packets of the flowat particular intervals according to the 0-1 bit sequence to label the flow. When the simulation flowwas captured by the model,the time intervals between packets in the flowwere transformed into the 0-1 bit sequence so that the watermarking content could be extracted to trace the provenance of this simulation flow. The resilience against various known attack techniques is illustrated through theoretical analysis. Moreover,a large number of experiments prove the validity of this model in tracing simulation flows under both normal and abnormal circumstance.

关 键 词：网络测试床 云计算 水印 溯源 模拟流量 

分 类 号：TP393[自动化与计算机技术—计算机应用技术]