The anonymous communication systems usually have special traffic patterns,which can be detected by censors for further disturbing or blocking.To solve this problem,we present a novel system,Deepflow,to hide anonymous communication traffic into the P2P streaming networks such as PPStream by using steganography.Each Deepflow node joins the PPStream network and performs like a normal PPStream client watching a live channel,while embedding communication data into video packets transferred in the network.The steganographed video packets are disseminated by innocuous PPStream clients and reach the target Deepflow node.It is not necessary for Deepflow nodes to link together in the network,which prevents malicious users from sensing IP address of other Deepflow nodes.Deepflow is competent for secret chatting,transferring text documents,or publishing secret bootstrapping information for other anonymous communication systems.Comprehensive experiments in real network environments are conducted in this paper to show the security and efficiency of Deepflow. The anonymous communication systems usually have special traffic patterns, which can be detected by censors for further disturbing or blocking. To solve this problem, we present a novel system, Deepflow, to hide anonymous communication traffic into the P2P streaming networks such as PPStream by using steganography.Each Deepflow node joins the PPStream network and performs like a normal PPStream client watching a live channel, while embedding communication data into video packets transferred in the network. steganographed video packets are disseminated by innocuous PPStream clients and reach the target Deepflow node. It is not necessary for Deepflow nodes to link together in the network, which prevents you from sensing IP address of other Deepflow nodes. Depress is competent for secret chatting, transferring text documents, or publishing secret bootstrapping information for other anonymous communication systems. Comprehensive experiments in real network environments are conducted in this paper to show the security and efficiency of Deepflow.