基于特征分析的流媒体数据包检测过滤方法

Streaming Media Data Packet Detection and Filtration Approach Based on Feature Analysis

针对实时网络流媒体数据传输过程中安全性和可靠性的需求,提出一种基于特征分析的检测过滤方法。通过对流媒体传输协议和数据流特征的分析,以及对报文头部特征和呼叫信令内容的解析,采用基于专有规则集的智能模式匹配方法对流媒体数据包进行检测识别。实验结果表明,与五元组数据包检测过滤方法相比,该方法在误判率、漏判率等方面都具有较为明显的优势,能够更有效地实现网络流媒体数据包的安全性过滤。

According to security and stability requirements of real-time streaming media transmission, this paper proposes a feature analysis-based detection and filtration approach, which parses feature of message head and calf signaling content by the analysis of streaming media transmission protocol and dam＇s streaming features, and then establishes an intelligent model matching method on the basis of special rules set, to matchingly detect the data packets. Experimental results show that this approach gets comparative advantage in the false positive and false negative rate, and comparied with the 5-tuple data packets detection filtration approach, it provides more effective security filtration of streaming media data packets.