移动应用特征提取算法研究

随着移动互联网的普及与发展,新的应用不断出现,移动互联网的网络安全和流量管理变得越来越重要,其应用监管与流量控制以移动应用识别为基础.深度包检测技术是准确性最高的一种应用层协议识别方法,其需要有准确的应用协议特征,应用特征产生的效率和准确度决定了应用识别系统的优劣.因此,高效、准确的自动提取应用特征具有十分重要的意义.文章从算法复杂性及敏感性两个方面,对当前主要特征提取算法进行比较和分析;并通过进行实验研究,分别给出了不同算法的性能仿真结果,为基于载荷的移动应用特征提取算法的选取提供了一个参考,具有一定的研究和应用指导价值.

Research on Signature Extraction Algorithms for Mobile Applications

With the popularization and development of mobile Intemet, new applications continue to emerge with wide range of application. Network security and traffic management of the mobile Intemet is becoming increasingly important, and its application regulation and flow control is based on mobile application identification. Deep Packet Inspection technology is one of the highest accuracy application layer identification method, which requires accurate application protocol significances, and efficiency and accuracy of generating signatures determine the merits of identification system. Therefore, efficient and accurate automated application signatures extraction has a very important significance. In this paper, the algorithm complexity and sensitivity of the current main signatures extraction algorithms are compared and analyzed; And through experimental studies, the performance simulation results of different algorithms are given respectively, provides a reference for the selection of mobile applications signatures extraction algorithms based on the payload, and has certain guidance value of research and application.