Understanding Content Filtering

Applies to: Exchange Server 2010

The Content Filter agent evaluates inbound e-mail messages and assesses the probability that an inbound message is legitimate or spam. Unlike many other filtering technologies, the Content Filter agent uses characteristics from a statistically significant sample of e-mail messages. The inclusion of legitimate messages in this sample reduces the chance of mistakes. Because the Content Filter agent recognizes characteristics of legitimate messages and spam, its accuracy is increased. Updates to the Content Filter agent are available periodically through Microsoft Update.

Contents

Using the Content Filter Agent

Configuring the Content Filter Agent

Using SCL Value Stamped by the Content Filter Agent in Edge Transport Rules

Forefront Protection 2010 for Exchange Server

Using the Content Filter Agent

The Content Filter agent is one of several anti-spam agents. When you configure anti-spam agents on a computer that has the Edge Transport server role installed, the agents act on messages cumulatively to reduce the amount of spam that enters the organization. For more information about how to plan and deploy anti-spam agents, see Understanding Anti-Spam and Antivirus Functionality.

The Content Filter agent assigns a spam confidence level (SCL) rating to each message. The SCL rating is a number between 0 and 9. A higher SCL rating indicates that a message is more likely to be spam.

You can configure the Content Filter agent to take the following actions on messages according to their SCL rating:

• Delete message
• Reject message
• Quarantine message

For example, you may determine that messages that have an SCL rating of 7 or higher must be deleted, messages that have an SCL rating of 6 must be rejected, and messages that have an SCL rating of 5 must be quarantined.

You can adjust the SCL threshold behavior by assigning different SCL ratings to each of these actions. For more information about how to adjust the SCL threshold to suit your organization's requirements and about per-recipient SCL thresholds, see Understanding Spam Confidence Level Threshold.

Allow Phrases and Block Phrases

You can customize how the Content Filter agent assigns SCL values by configuring custom words. Custom words are individual words or phrases that the Content Filter agent uses to apply appropriate filter processing. You configure approved words or phrases with Allow phrases and unapproved words or phrases with Block phrases. When the Content Filter agent detects a preconfigured Allow phrase in an inbound message, the Content Filter agent automatically assigns an SCL value of 0 to the message. Alternatively, when the Content Filter agent detects a configured Block phrase in an inbound message, the Content Filter agent assigns an SCL rating of 9.

You can enter custom words or phrases in any combination of uppercase and lowercase letters. However, when the Content Filter agent evaluates message content, it ignores case. The maximum number of custom words or phrases that can be created is 800.

Outlook E-mail Postmark Validation

The Content Filter agent also includes Microsoft Office Outlook E-mail Postmark validation, a computational proof that Outlook applies to outgoing messages to help recipient messaging systems distinguish legitimate e-mail from junk e-mail. This feature helps reduce the chance of false positives. In the context of spam filtering, a false positive exists when a spam filter incorrectly identifies a message from a legitimate sender as spam. When Outlook E-mail Postmark validation is enabled, the Content Filter agent parses the inbound message for a computational postmark header. The presence of a valid, solved computational postmark header in the message indicates that the client computer that generated the message solved the computational postmark.

Computers don't require significant processing time to solve individual computational postmarks. However, processing postmarks for many messages may be prohibitive to a malicious sender. Anyone who sends millions of spam messages is unlikely to invest the processing power that is required to solve computational postmarks for all outbound spam. If a sender's e-mail contains a valid, solved computational postmark, it's unlikely that the sender is a malicious sender. In this case, the Content Filter agent would lower the SCL rating. If the postmark validation feature is enabled and an inbound message either doesn't contain a computational postmark header or the computational postmark header isn't valid, the Content Filter agent would not change the SCL rating.

Bypassing the Recipient, Sender, and Sender Domain

In some organizations, all e-mail to certain aliases must be accepted. This scenario can introduce problems if your organization is in an industry that manages significant volumes of spam.

For example, a company named Woodgrove Bank has an alias named customerloans@woodgrovebank.com that provides e-mail-based support to external loan customers. The Exchange administrators configure the Content Filter agent to set Block phrases that filter out words or phrases that are typically used in spam that is sent by unscrupulous loan agencies. To prevent potentially legitimate messages from being rejected, the administrators set exceptions to content filtering by entering a list of SMTP e-mail recipient addresses in the Content Filter agent configuration.

You can also specify senders and sender domains that you do not want the Content Filter agent to block.

Safelist Aggregation

In Exchange 2010, the Content Filter agent on the Edge Transport server uses the Outlook Safe Senders Lists, Blocked Sender List, Safe Recipients Lists, and trusted contacts from Outlook to optimize spam filtering. Safelist aggregation is a set of anti-spam functionality that is shared across Outlook and Exchange 2010. As its name suggests, this functionality collects data from the anti-spam safe lists that Outlook users configure and makes this data available to the anti-spam agents on the Edge Transport server. E-mail messages that Outlook users receive from contacts that those users have added to their Outlook Safe Recipients List, Safe Senders List, or trusted contacts list are identified by the Content Filter agent as safe. The Sender Filter agent also performs per-recipient sender filtering using the Blocked Senders list that users configure. For more information, see Understanding Safelist Aggregation.

Configuring the Content Filter Agent

You configure the Content Filter agent by using the Exchange Management Console or the Exchange Management Shell.

For more information about how to configure content filtering, see Configure Content Filtering Properties.

Using SCL Value Stamped by the Content Filter Agent in Edge Transport Rules

In Exchange 2010, transport rules that run on Edge Transport servers are applied to messages by the Edge Rule agent on the OnEndOfData SMTP transport event. One of the transport rule conditions available on Edge Transport servers is the with a spam confidence (SCL) rating that is greater than or equal to limit transport rule condition. By using this transport rule condition, you can apply a transport rule action to a message based on the SCL value stamped on the message. The Content Filter agent stamps an SCL value on the message based on an analysis of the message content and is used to determine whether the message is spam. The Content Filter agent also runs on the OnEndOfData SMTP transport event.

Because both the Edge Rule agent and the Content Filter agent run on the OnEndOfData SMTP transport event, the priority value applied to each transport agent is used to determine which transport agent runs first. By default, the Edge Rule agent runs before the Content Filter agent to reduce the cost of processing messages that may be blocked by the Edge Rule agent. However, because the Edge Rule agent runs before the Content Filter agent and therefore the SCL value has not yet been stamped on the message, you can't use the with a spam confidence (SCL) rating that is greater than or equal to limit transport rule condition in the default configuration.

For details about how to configure the Content Filter agent to run before the Edge Rule agent on the OnEndOfData SMTP transport event, see Make the SCL Value Available to Edge Transport Rules. This enables the Content Filter agent to stamp an SCL value on a message that can then be read by the with a spam confidence (SCL) rating that is greater than or equal to limit transport rule condition.

For more information about transport agents and transport agent priority, see Understanding Transport Agents.

If you configure the Content Filter agent with a higher priority value than the Edge Rule agent, the Edge Transport server may incur additional processing costs because all the messages that are received by the Edge Transport server will be evaluated by the Content Filter agent. This is true even if the message is later rejected by a transport rule that is configured on the Edge Rule agent. Also, you will no longer be able to configure a transport rule on the Edge Transport server to stamp a message that has an SCL value of -1. This value indicates to the Content Filter agent that the message should not be evaluated.

Forefront Protection 2010 for Exchange Server

Microsoft Forefront Protection 2010 for Exchange Server (FPE) integrates multiple scan engines into a comprehensive, layered solution that helps you protect your Microsoft Exchange server messaging environment from malware, spam, and inappropriate content. FPE prevents the spread of malicious content by scanning all messages in real time with minimal impact on Exchange server performance or message delivery time.

You can enable FPE anti-spam technology in both the Exchange Edge Transport and Exchange Hub Transport roles. However, the Edge Transport role is the preferred location for anti-spam filtering. The technology includes a series of agents that are registered with Exchange and are invoked at specific points in the SMTP pipeline. FPE can also be integrated with Forefront Online Protection for Exchange (FOPE) to provide an additional layer of filtering for your messaging environment.

When you deploy FPE, the anti-spam features that are built in to Exchange are disabled. To learn more about how the FPE anti-spam solution works, see Using Antispam Filtering.