基于协议分析的IM阻断策略及算法分析

On the Policy and Algorithm to Block Instant Messaging

本文根据即时通讯(IM)的协议模型,在TCP层和即时通讯协议层分析了阻断即时通讯的可能性,分别提出了基于TCP层和即时通讯协议层的协议分析的阻断策略.在TCP层,通过插入FIN报文或RST报文,以关闭TCP连接(特别是与登录服务器的连接);在即时通讯协议层,通过对相应字段的值进行设置,使客户端进入不活动状态.为使阻断策略能够有效地实施,本文也提出了一个消息驱动的实现算法.通过对Yahoo!Messenger的阻断实验,证明了本文提出的方法能够高效地完成阻断任务(平均数据包量不超过2.5个).用统计分析的方法对它们的性能进行比较,结果表明各种方法的效率在不同的背景流量下没有显著的差异,通过消息驱动算法实现的基于协议分析的阻断策略,其工作效率不受背景流量的影响.另外,即时通讯数据的触发方式,使得发送的阻断包的数量与背景流量无关,也不对网络负载造成明显的影响.本文的工作为实施即时通讯的阻断提供了有效的解决方案.

Depending on the general protocol model, this paper analyzed the possibility of blocking Instant Messaging （IM） at TCP layer and IM protocol layer respectively.At TCP layer, packets of FIN or RST were inserted to close a TCP connection;at IM protocol layer, certain commands with special flag were inserted to enforce a client to be idle. To implement these blocking policies effectively, a message-driven algorithm was offered, too. Some experiments to block Yahoo! Messenger and their statistical analyses were also presented in this paper. The results showed that all the policies implemented through the message-driven algorithm block IM effectively, their performances are independent of the back traffic, and thus, they don＇t impose much burden on the whole Internet. The works done in this paper provided practicable and effective solutions to block IM.