文档详情

使用IMF过滤垃圾邮件.doc

发布:2017-12-13约1.02万字共10页下载文档
文本预览下载声明
使用Exchange IMF引擎过滤垃圾邮件 文档主要包含以下几个部分: IMF技术简介 IMF配置和微调步骤 IMF归档邮件察看和处理步骤 IMF日常维护和排错建议 工作步骤Check List 参考资料 IMF技术简介 1)IMF Overview Intelligent Message Filter (IMF) is a content filter designed specifically for Exchange. It is based on patented machine-learning technology from Microsoft Research known as Microsoft SmartScreen? technology. SmartScreen is currently used by MSN, Microsoft Hotmail?, Microsoft Office Outlook? 2003, and Exchange. IMF was designed to distinguish between characteristics of legitimate e-mail messages and spam, based on millions of messages. IMF can make an accurate assessment of the probability that an incoming e-mail message is either a legitimate message or spam. Unlike many other filtering technologies, IMF uses characteristics from a statistically sound sample of e-mail messages. In addition to spam, the inclusion of legitimate messages in this sample reduces the likelihood of mistakes. Because IMF recognizes characteristics of both legitimate and UCE messages, the accuracy of IMF is increased. Intelligent Message Filter is installed on Exchange servers that accept inbound SMTP messages from the Internet. When an external user sends e-mail messages to an Exchange server with IMF installed, the IMF evaluates the textual content of the messages and assigns each message a rating based on the probability that the message is spam. This rating ranges from 1 to 9 and is stored as a message property known as the spam confidence level (SCL) rating. This rating is persisted with the message when the message is sent to other Exchange servers. After Intelligent Message Filter has assigned an SCL to the message, it is evaluated against two thresholds configured by the administrator as follows: Gateway blocking configuration: Block messages with an SCL rating greater than or equal to. If the SCL of a message is greater than or equal to the value set here, one of the following actions can be performed on the message : Archive Delete No action Reject Store
显示全部
相似文档