Abstract
The traffic model and application identification of Wechat has gained a great concern. First of all, Wechat traffic model was deep analyzed and concluded. It could be concluded that the traffic was a pulsed long-live connection. Furthermore, the applicability and economy of this pattern on Wechat were analyzed. Secondly, a method was proposed to accurately identify the Wechat traffic using deep packet inspection, and a payload-based signature was given. Finally, the sub-functions of Wechat in a fine-grained fashion was classified. Several algorithms were proposed to assist different sub-functions to improve their accuracy of identification. Experimental results showed that our proposed approaches could achieve the accuracy of 98% in identification and the accuracy of 52% in classification of sub-functions.
| Original language | English |
|---|---|
| Pages (from-to) | 257-263 |
| Number of pages | 7 |
| Journal | Dongbei Daxue Xuebao/Journal of Northeastern University |
| Volume | 35 |
| Publication status | Published - 1 Oct 2014 |
Keywords
- Application identification
- Deep packet inspection
- Fine-grained classification
- Traffic classification