即時通訊（Instant Messenger）已成為現今網際網路中不可或缺的一項溝通工具，即時通訊軟體帶給人們便利卻也造成網路安全上的一大漏洞。隨著使用人數增加，公司已不能不正視即時通訊所帶來的各種問題。 過去公司為了避免使用者利用即時通訊洩漏公司機密，會利用一些阻擋方式來限制公司員工使用即時通訊軟體，如關閉通訊埠…等等。但隨著即時通訊軟體版本不斷地更新，越來越多即時通訊軟體已無法使用過去舊有的方式來阻擋使用者使用。本篇主要就是要討論不同於過去的阻擋方式，而是分析即時通訊的特徵及行為模式，並利用支援向量機來訓練即時通訊偵測系統，可作為未來在監測、側錄及阻擋即時通訊上一項依據。 Nowadays the instant messenger has become an indispensable Internet communication tool, the instant messenger software makes people feel convenient but also creates a big loophole in the network security. With the popularity of the instant messenger, the company has to face up to variaous questions which the instant messenger brings. In the past, many companies used some blocking ways to avoid staffs using the instant messenger to divulge the company secret, like closing communication port...etc. But with the evolution of the instant messenger software, some old methods cannot keep the users from the messengers. In this thesis, we discuss detection of instant messengers by different way. We analyze the characteristics of various instant messengers and their behavior patterns. We train the instant messenger detection system by using the support vector machine. In the future, our research results can be a basis for the monitoring, side recording and obstruction of the instant messenger.
