User Guide for AsyncOS 14.0.1 for Cisco Secure Email Gateway - MD (Maintenance Deployment)
Bias-Free Language
The documentation set for this product strives to use bias-free language. For the purposes of this documentation set, bias-free is defined as language that does not imply discrimination based on age, disability, gender, racial identity, ethnic identity, sexual orientation, socioeconomic status, and intersectionality. Exceptions may be present in the documentation due to language that is hardcoded in the user interfaces of the product software, language used based on RFP documentation, or language that is used by a referenced third-party product. Learn more about how Cisco is using Inclusive Language.
Anti-spam processes
scan email for incoming (and outgoing) mail based on the mail policies that you
configure.
One or more scanning engines
scan messages through their filtering modules.
Scanning engines assign a
score to each message. The higher the score, the greater the likelihood that
the message is spam.
Based on the score, each
message is categorized as one of the following:
Not spam
Suspected spam
Positively-identified spam
An action is taken based on
the result.
Actions taken on
messages positively identified as spam, suspected to be spam, or identified as
unwanted marketing messages are not mutually exclusive; you can combine some or
all of them differently within different incoming or outgoing policies for
different processing needs for groups of users. You can also treat positively
identified spam differently from suspected spam in the same policy. For
example, you may want to drop messages positively identified as spam, but
quarantine suspected spam messages.
For each mail policy,
you can specify thresholds for some of the categories, and determine the action
to take for each category. You can assign different users to different mail
policies and define different scanning engines, spam-definition thresholds, and
spam-handling actions for each policy.
You can license and enable both these solutions on your
email gateway, but you can only use one in a particular mail policy. You can specify a different anti-spam solution for different groups
of users.
How to Configure the
Email Gateway to Scan Messages for Spam
Procedure
Command or Action
Purpose
Step 1
Enable anti-spam scanning on the
email gateway.
Note
Remaining steps in this table apply to both scanning engine options.
If you have feature keys for both Cisco IronPort Anti-Spam and Intelligent Multi-Scan, you can enable both solutions on the
email gateway.
(Recommended) Enable IP Reputation Service scoring for each inbound mail flow policy, even if you are not rejecting connections based on IP Reputation Scores.
For each inbound mail flow policy, ensure that “Use SenderBase for Flow Control” is On.
If your
email gateway does not connect directly to external senders to receive incoming mail, but instead receives messages relayed through a mail
exchange, mail transfer agent, or other machine on your network, ensure that relayed incoming messages include the original
sender IP address.
Your
email gateway ships with a 30-day evaluation key for the Cisco Anti-Spam software. This key is not enabled until you accept the license
agreement in the system setup wizard or Security Services > IronPort Anti-Spam pages (in the GUI) or the systemsetup or antispamconfig
commands (in the CLI). Once you have accepted the agreement, Cisco Anti-Spam will be enabled, by default, for the default
incoming Mail Policy. An alert is also sent to the administrator address you configured (see the System Setup Wizard, Step 2: System) noting that the Cisco Anti-Spam license will expire in 30 days. Alerts are sent 30, 15, 5, and 0 days prior to expiration.
For information on enabling the feature beyond the 30-day evaluation period, contact your Cisco sales representative. You
can see how much time remains on the evaluation via the System Administration > Feature Keys page or by issuing the featurekey
command. (For more information, see Feature Keys.)
Cisco Anti-Spam: an
Overview
IronPort Anti-Spam
addresses a full range of known threats including spam, phishing and zombie
attacks, as well as hard-to-detect low volume, short-lived email threats such
as “419” scams. In addition, IronPort Anti-Spam identifies new and evolving
blended threats such as spam attacks distributing malicious content through a
download URL or an executable.
To identify these
threats, IronPort Anti-Spam examines the full context of a message-its content,
methods of message construction, the reputation of the sender, the reputation
of web sites advertised in the message, and more. IronPort Anti-Spam combines
the power of email and web reputation data, leveraging the full power of the
world's largest email and web traffic monitoring network — SenderBase — to
detect new attacks as soon as they begin.
IronPort Anti-Spam
analyzes over 100,000 message attributes across the following dimensions:
Email reputation —
who is sending
you this message?
Message content —
what content
is included in this message?
Message structure —
how was this
message constructed?
Web reputation —
where does the
call to action take you?
Analyzing
multi-dimensional relationships allows the system to catch a broad range of
threats while maintaining accuracy. For example, a message that has content
claiming to be from a legitimate financial institution but that is sent from an
IP address on a consumer broadband network or that contains a URL hosted on a
“zombie” PC will be viewed as suspicious. In contrast, a message coming from a
pharmaceutical company with a positive reputation will not be tagged as spam
even if the message contains words closely correlated with spam.
If you have not
enabled IronPort Anti-Spam in the system setup wizard:
Click
Enable.
Scroll to
the bottom of the license agreement page and click
Accept to accept the agreement.
Step 3
Click
Edit
Global Settings.
Step 4
Select the
check box for
Enable
IronPort Anti-Spam Scanning.
Checking this box enables the feature globally for the
email gateway.
Step 5
To optimize the throughput of your
email gateway while still being able to scan increasingly larger messages sent by spammers, configure the thresholds for message scanning
by Cisco Anti-Spam.
Option
Description
Message Scanning Thresholds
Enter a value for Always scan messages smaller than —The recommended value is 1 MB or less. Messages smaller than the always scan size will be fully scanned, except in cases of “early exit.” Messages larger than this size are partially scanned if they
are smaller than the never scan size.
Cisco advises not to exceed 3 MB for the always scan message size. A larger value may result in decreased performance.
Enter a value for Never scan messages larger than —The recommended value is 2 MB or less. Messages larger than this size will not be scanned by Cisco Anti-Spam and the X-IronPort-Anti-Spam-Filtered:
true header will not be added to the message.
Cisco advises not to exceed 10 MB for the never scan message size. A larger value may result in decreased performance.
For messages larger than the always scan size or smaller than the never scan size, a limited and faster scan is performed.
Note
If the Outbreak Filters maximum message size is greater than Cisco Anti-Spam’s always scan message, messages smaller than the Outbreak Filters maximum size are fully scanned.
Timeout for Scanning Single Message
Enter the number of seconds to wait for timeout when scanning a message.
Enter an integer from 1 to 120. The default value is 60 seconds.
Scanning Profile
Choose from any of the following scanning profiles to catch spam messages:
Normal - Enable this option for a balanced approach to block spam.
Aggressive - Enable this option to provide stronger emphasis to block spam. When enabled, tuning the Anti-Spam policy thresholds have
more impact on spam detection than the Normal profile with a larger potential for false positives.
Note
When using the new aggressive scanning profile mail policy adjustments to Anti-Spam thresholds have a larger impact than before.
Therefore when enabling the aggressive profile, any Anti-Spam policy thresholds previously adjusted should be reset to default
settings and then reevaluated for the best balance of spam catch rate vs. false positive potential.
Step 6
Submit and
commit your changes.
Configuring Intelligent Multi-Scan and Graymail Detection
This section describes how to configure Cisco Intelligent Multi-Scan and Graymail Detection and Safe Unscubscribing:
Cisco Intelligent Multi-Scan incorporates multiple anti-spam scanning engines, including Cisco Anti-Spam, to provide a multi-layer
anti-spam solution.
When processed by Cisco Intelligent Multi-Scan:
A message is first scanned by third-party anti-spam engines.
Cisco Intelligent Multi-Scan then passes the message and the verdicts of the third-party engines to Cisco Anti-Spam, which
assumes responsibility for the final verdict.
After Cisco Anti-Spam performs its scan, it returns a combined multi-scan score to AsyncOS.
Combining the benefits of the third-party scanning engines and Cisco Anti-Spam results in more caught spam while maintaining
Cisco Anti-Spam’s low false positive rate.
You cannot configure the order of the scanning engines used in Cisco Intelligent Multi-Scan; Cisco Anti-Spam will always be
the last to scan a message and Cisco Intelligent Multi-Scan will not skip it if a third-party engine determines that a message
is spam.
Using Cisco Intelligent Multi-Scan can lead to reduced system throughput. Please contact your Cisco support representative
for more information.
Note
The Cisco Intelligent Multi-Scan feature key also enables Cisco Anti-Spam on the
email gateway, giving you the option of enabling either Cisco Intelligent MultiScan or Cisco Anti-Spam for a mail policy.
Important
When Cisco Intelligent Multi-Scan is enabled during system setup, it is enabled for the default incoming mail policy with
the default values for the global settings.
Before you begin
Activate the feature key for this feature. See Feature Keys. You will see the IronPort Intelligent Multi-Scan option only if you have done so.
Procedure
Step 1
Select Security Services > IMS and Graymail.
Step 2
If you have not enabled Cisco Intelligent Multi-Scan in the system setup wizard:
Click Enable.
Scroll to the bottom of the license agreement page and click Accept to accept the agreement.
Step 3
Click Edit IMS Settings.
Step 4
Select the check box for Enable Intelligent Multi-Scan enable the feature globally for the
email gateway. However, you must still enable per-recipient settings in Mail Policies.
Graymail messages are messages that do not fit the definition of spam, for example, newsletters, mailing list subscriptions,
social media notifications, and so on. These messages were of use at some point in time, but have subsequently diminished
in value to the point where the end user no longer wants to receive them.
The difference between graymail and spam is that the end user intentionally provided an email address at some point (for example,
the end user subscribed to a newsletter on an e-commerce website or provided contact details to an organization during a conference)
as opposed to spam, messages that the end user did not sign up for.
Graymail Management Solution in
Email Gateway
The graymail management solution in the
email gateway comprises of two components: an integrated graymail scanning engine and a cloud-based Unsubscribe Service.
The graymail
management solution allows organizations to:
Identify graymail using the
integrated graymail engine and apply appropriate policy controls.
Provide an easy mechanism
for end users to unsubscribe from unwanted messages using Unsubscribe Service.
In addition to these,
the graymail management solution also help organizations to provide:
Secure unsubscribe option for
end users.
Mimicking an unsubscribe option is a popular phishing
technique. For this reason, the end users are generally wary of clicking
unknown unsubscribe links. For such scenarios, the cloud-based Unsubscribe
Service extracts the original unsubscribe URI, checks the reputation of the
URI, and then performs the unsubscribe process on behalf of the end user. This
protects end users from malicious threats masquerading as unsubscribe links.
Uniform subscription
management interface for end users. Different graymail senders use
different layouts for displaying unsubscribe links to the users. The users must
search for the unsubscribe link in the message body and perform the
unsubscribing. Irrespective of the graymail senders, the graymail management
solution provides a common layout for displaying unsubscribe links to the
users.
Better visibility for
administrators into various graymail categories. The graymail
engine classifies each graymail into three categories (see
Graymail Classification)
and the administrators can set policy controls based on these categories.
The graymail engine classifies each graymail into one of the following
categories:
Marketing
Email. Advertising messages sent by professional marketing groups,
for example, bulletins from Amazon.com with details about their newly launched
products.
Social Network
Email. Notification messages from social networks, dating websites,
forums, and so on. Examples include alerts from:
LinkedIn, for jobs that
you may be interested in
CNET forums, when a user responds to your post.
Bulk Email. Advertising messages sent by
unrecognized marketing groups, for example, newsletters from TechTarget, a
technology media company.
How Graymail
Management Solution Works
The following steps
illustrates the workflow of graymail management solution:
Figure 1. Graymail Management Solution Workflow
Workflow
Procedure
Step 1
The
email gateway receives an incoming message.
Step 2
The
email gateway checks if graymail detection is enabled. If graymail detection is enabled, go to Step 3. Else, go to Step 8
Step 3
The
email gateway checks if the message is spam, virus, or malware positive. If positive, go to Step 8. Else, go to Step 4
Step 4
The
email gateway checks if the message is graymail. If the message is graymail, go to Step 5. Else, go to Step 8
Step 5
The
email gateway applies the configured policy actions such as, drop, deliver, bounce, or quarantine to the spam quarantine.
Step 6
The
email gateway checks if safe unsubscribing enabled. If safe unsubscribing is enabled, go to Step 7. Else, go to Step 8.
Step 7
The
email gateway adds a banner with unsubscribe button to the message. Also, the
email gateway rewrites the existing unsubscribe links in the message body.
Step 8
The
email gateway processes the message through the next stages of its email work queue.
What to do next
For an overview of how email is processed through the system, from reception to routing to delivery, see Understanding the Email Pipeline
The following flow
diagram shows how safe unsubscribing works.
Figure 2. Safe
Unsubscribing Workflow
Workflow
Procedure
Step 1
End user
receives a message with the graymail banner.
Step 2
End user clicks
on the Unsubscribe link.
Step 3
Unsubscribe
Service extracts the original unsubscribe URI.
Step 4
Unsubscribe
Service checks the reputation of the URI.
Step 5
Depending on
the reputation of the URI, the Unsubscribe Service performs one of the
following actions:
If the URI is malicious, the
Unsubscribe Service will not perform the unsubscribe process and displays a
block page to the end user.
If the URI is not malicious,
depending on the URI type ( http or mailto ), the Unsubscribe Service sends an
unsubscribe request to the graymail sender.
If the request is
successful, the Unsubscribe Service displays the “Successfully unsubscribed”
status to the end user.
If the
first unsubscribe request fails, the Unsubscribe Service displays the
“Unsubscribe process in progress” status and provides a URL that can be used to
track the status of the unsubscribing.
End
users can use this URL to track the status at a later point. After the first
failed attempt, the Unsubscribe Service sends periodic unsubscribe requests for
a duration of four hours.
If an
end user checks the status of the unsubscribe process at a later point,
If one
of the requests within the four hour duration (from the first failed attempt)
is successful, the Unsubscribe Service displays the “Successfully unsubscribed”
status to the end user.
If none of the requests
within the four hour duration (from the first failed attempt) are successful,
the Unsubscribe Service displays the “Unable to subscribe” status to the end
user and provides a URL that can be used to unsubscribe from the graymail
manually.
Configuring Graymail
Detection and Safe Unsubscribing
Requirements for
Graymail Detection and Safe Unsubscribing
For graymail detection, anti-spam scanning must be enabled globally. This can be either the IronPort Anti-Spam, the Intelligent
Multi-Scan feature, or Outbreak Filters. See Managing Spam and Graymail.
For safe unsubscribing,
Add the safe unsubscribing feature key.
The end user machines must be able to connect to the cloud-based Unsubscribe Service directly over the Internet.
Graymail Detection and Safe Unsubscribing in Cluster Configurations
You can enable Graymail Detection and Safe Unsubscribing at the machine, group or cluster level.
Enable Graymail Detection and Safe Unsubscribing
Procedure
Step 1
Select Security Services > IMS and Graymail.
Step 2
Click Edit Graymail Settings.
Step 3
Check Enable Graymail Detection.
Step 4
Check Enable Safe Unsubscribe.
Step 5
(Optional) Check Enable Automatic Updates to enable automatic update of the engine.
The
email gateway fetches the required updates for the particular engine from the update server.
To configure Graymail Detection and Safe Unsubscribing global settings in CLI, use the imsandgraymailconfig CLI command. For more information, see CLI Reference Guide for AsyncOS for
Cisco Secure Email Gateway.
Configuring the
Incoming Mail Policy for Graymail Detection and Safe Unsubscribing
Click the link
in the
Graymail column of the mail policy to modify.
Step 3
Depending on
your requirements, choose the following options:
Enable graymail detection
Enable safe unsubscribing
Choose
whether to apply the above actions on all messages or only on unsigned
messages.
Note
The
email gateway considers a message signed if it is encrypted using S/MIME or it contains an S/MIME signature.
Actions to be taken on
various graymail categories (Marketing Email, Social Network Email, and Bulk
Email):
Drop,
deliver, bounce, or quarantine (to the spam quarantine) the message
Note
If
you plan to use safe unsubscribing option, you must set the action to deliver
or quarantine.
Send the message to an
alternate host
Modify subject of the
message
Add custom headers
Send
the message to an alternate envelope recipient
Note
If
you are sending a graymail positive message to an alternate envelope recipient,
banner will not be added.
Archive
the message
Note
If you are planning only to monitor the detected graymail, you can enable graymail detection per policy without having to
configure actions for various graymail categories. In this scenario, the
email gateway takes no action on the detected graymail.
Step 4
Submit and
commit your changes.
What to do next
Note
You can also
configure outgoing mail policies for graymail detection. Keep in mind that, in
this scenario, you cannot configure safe unsubscribing.
To configure policy settings for Graymail Detection and Safe Unsubscribing in CLI, use the policyconfig command. For more information, see CLI Reference Guide for AsyncOS for
Cisco Secure Email Gateway .
IronPort-PHdr Header Added During Graymail Scanning
The IronPort-PHdr header is added to all messages that are processed by the Graymail engine when:
Graymail engine is enabled globally on the
email gateway.
Graymail scanning
is enabled for a specific mail policy.
Note
If Graymail scanning is not enabled for a specific mail policy, the IronPort-PHdr header is still added to all messages, if
the Graymail engine is enabled globally on the
email gateway.
The IronPort-PHdr header contains encoded proprietary information and is not customer-decodable. This header provides additional
information about debugging issues with your Graymail configuration.
Note
If Anti-Spam engine or Outbreak Filter is enabled for a specific mail policy, the IronPort-PHdr header is added to all messages
that pass through the specific mail policy.
Bypassing Graymail Actions using Message Filters
If you do not want to apply graymail actions on certain messages, you can use the following message filters to bypass graymail
actions:
Message Filter Action
Description
skip-marketingcheck
Bypass actions on marketing emails
skip-socialcheck
Bypass actions on social network emails
skip-bulkcheck
Bypass actions on bulk emails
The following example specifies that messages received on the listener “private_listener” must bypass graymail actions on
social network emails.
internal_mail_is_safe:
if (recv-listener == 'private_listener')
{
skip-socialcheck
();
}
Monitoring
Graymail
You can view data
about detected graymail using the following reports.
Report
Contains
the Following Graymail Data
More Info
Overview
page > Incoming Mail Summary
The number
of incoming graymail messages under each graymail category (Marketing, Social,
and Bulk) and the total number of graymail messages.
The number
of incoming graymail messages under each graymail category (Marketing, Social,
and Bulk) and the total number of graymail messages for all the IP addresses,
domain names, or network owners.
Incoming
Mail page > Incoming Mail Details > Sender Profile (drill down view)
The number
of incoming graymail messages under each graymail category (Marketing, Social,
and Bulk) and the total number of graymail messages for a given IP address,
domain name, or network owner.
The number
of incoming graymail messages under each graymail category (Marketing, Social,
and Bulk) and the total number of graymail messages for all the users.
Internal
Users page > User Mail Flow Details > Internal User (drill down view)
The number
of incoming graymail messages under each graymail category (Marketing, Social,
and Bulk) and the total number of graymail messages for a given user.
If you had enabled
Marketing Email Scanning under anti-spam settings for a mail policy, after
upgrading to AsyncOS 9.5 or later, keep in mind that:
The number of marketing
messages is a sum of marketing messages detected before and after the upgrade.
The total number of graymail
messages does not include the number of marketing messages detected before the
upgrade.
The total number of attempted
messages also includes the number of marketing messages detected before the
upgrade.
Updating Graymail Rules
If you have enabled service updates, scanning rules for the graymail
management solution is retrieved from the Cisco update servers. But in some
scenarios (for example, you have disabled automatic service updates or
automatic service update is not working), you may want to manually update
graymail rules.
To manually update the graymail rules, do one of the following:
In web interface, go to Security Service > IMS and Graymail page, and click Update Now.
In CLI, run the
graymailupdate command.
To know the details of existing graymail rules, see the Rule Updates section of the IMS and Graymail page in web interface or use the graymailstatus command in CLI.
Customizing the
Appearance of Unsubscribe Page for End Users
When an end user
clicks on unsubscribe link, the Unsubscribe Service displays a Cisco branded
Unsubscribe page indicating the status of the unsubscribe process (see
How Safe Unsubscribing Works).
You can customize the appearance of the Unsubscribe page and display your
organization’s branding (such as company logo, contact information, and so on)
using
Security
Services >
Block Page
Customization. For instructions, see
Customizing the Notification That End Users See If a Site Is Malicious.
End-User
Safelist
If the end users in
your organization have configured Safelist for their own email accounts,
graymail messages from a sender in the safelist will not be scanned by the
graymail scanning engine. For more information about Safelists, see
Using Safelists and Blocklists to Control Email Delivery Based on Sender.
Viewing Logs
The graymail detection and safe unsubscribing information is posted to
the following logs:
Graymail Engine
Logs. Contains information about the graymail engine, status,
configuration, and so on. Most information is at Info or Debug level.
Graymail
Archive. Contains archived messages (the messages that are scanned
and associated with the “archive message” action). The format is an mbox-format
log file.
Mail
Logs. Contains information about graymail detection and addition of
banner for safe unsubscribing. Most information is at Info or Debug level.
Troubleshooting Graymail Detection and Safe Unsubscribing
After clicking on the
Unsubscribe link, the end user sees the following message: “Unable to
unsubscribe from...”
Solution
This problem can occur
if the Unsubscribe Service is unable to perform the safe unsubscribe on behalf
of the end user. The following are some of the common scenarios in which the
Unsubscribe Service is unable to perform the safe unsubscribe:
Unsubscribe URI or mailto
address is wrong.
Websites that require the
end users’ credentials to unsubscribe.
Websites that require the
end users to confirm the request of unsubscribing by logging into their email
accounts.
Websites that require
captcha to be solved and the Unsubscribe Service is unable to solve the
captcha.
Websites that require
interactive unsubscribing.
The end users can use
the URL provided at the bottom of the unsubscribe page to unsubscribe manually.
Configuring Global Settings for Intelligent Multi-Scan and Graymail Detection
To optimize the throughput of your
email gateway, you can configure the threshold and the timeout settings for scanning messages by Cisco Intelligent Multi-Scan and Graymail.
These settings are common for both Cisco Intelligent Multi-Scan and Graymail configuration.
Select Security Services > IMS and Graymail
Click Edit Global Settings.
Select the thresholds for scanning with Cisco Intelligent Multi-Scan and Graymail Detection.
The default values are:
Always scan 512K or less.
Note
This setting is not applicable for Graymail Detection and Safe Unsubscribing.
Never scan 1M or more.
Enter the number of seconds to wait for timeout when scanning a message.
When specifying the number of seconds, enter an integer from 1 to 120. The default value is 60 seconds.
Most users do not have to change the maximum message size to be scanned or the timeout value. That said, you may be able to
optimize the throughput of your
email gateway by lowering the maximum message size setting.
Submit and comit your changes.
Defining Anti-Spam
Policies
For each mail
policy, you specify settings that determine which messages are considered spam
and what action to take on those messages. You also specify which engine will
scan messages that the policy applies to.
You can configure
different settings for the default incoming and outgoing mail policies. If you
need different anti-spam policies for different users, use multiple mail
policies with different anti-spam settings. You can enable only one anti-spam
solution per policy; you cannot enable both on the same policy.
Navigate to the
Mail
Policies > Incoming Mail Policies page.
Or
Step 2
Navigate to the
Mail
Policies > Outgoing Mail Policies page.
Step 3
Click the link
under the
Anti-Spam column for any mail policy.
Step 4
In the
Enable
Anti-Spam Scanning for This Policy section, select the anti-spam
solution you want to use for the policy.
Options you see
depend on the anti-spam scanning solution(s) that you have enabled.
For mail
policies other than the default: If you use settings from the default policy,
all other options on the page are disabled.
You can also
disable anti-spam scanning altogether for this mail policy.
Step 5
Configure
settings for positively identified spam, suspected spam, and marketing
messages:
Option
Description
Enable Suspected Spam Scanning
Enable Marketing Email Scanning
Choose an option.
Positively-identified spam scanning is always enabled if anti-spam scanning is
enabled.
Apply This Action to Message
Choose which overall action to take on positively identified spam, suspected
spam, or unwanted marketing messages:
Deliver
Drop
Bounce
Quarantine
(Optional) Send to Alternate Host
You
can send identified messages to an alternate destination mailhost (an email
server other than the ones listed in SMTP Routes or DNS).
Enter an IP address or hostname. If you enter a hostname, its Mail Exchange
(MX) will be queried first. If none exists, the A record on the DNS server will
be used (as with SMTP Routes).
Use
this option if you want to redirect messages, for example to a sandbox mail
server for further examination.
You
can alter text in the Subject of identified messages by prepending or appending
certain text strings to help users more easily identify and sort spam and
unwanted marketing messages.
Note
White space is not ignored in this field. Add spaces after (if
prepending) or before (if appending) the text you enter in this field to
separate your added text from the original subject of the message. For example,
if you are prepending, add the text
[SPAM] with a few trailing spaces.
“Add Text to Subject” field only accepts US-ASCII characters.
Advanced Options (for custom header and message delivery)
(Optional) Add Custom Header
You
can add a custom header to identified messages.
(Optional) Send to an Alternate Envelope Recipient
You
can have identified messages sent to an alternate envelope recipient address.
Click
Advanced
and define an alternate address.
For
example, you could route messages identified as spam to an administrator’s
mailbox for subsequent examination. In the case of a multi-recipient message,
only a single copy is sent to the alternate recipient.
Archive Message
You
can archive identified messages into the “Anti-Spam Archive” log. The format is
an mbox-format log file.
Spam Thresholds
Use
the default thresholds or enter a threshold value for positively identified
spam and a value for suspected spam.
Understanding
Positive and Suspect Spam Thresholds
When evaluating
messages for spam, both anti-spam scanning solutions apply thousands of rules
in order to arrive at an overall spam score for the message. The score is then
compared to the thresholds specified in the applicable mail policy to determine
whether the message is considered spam.
For highest accuracy,
the threshold for positive identification as spam is quite high by default:
Messages scoring between 90 and 100 are considered to be positively identified
as spam. The default threshold for suspected spam is 50.
Messages with scores below
the suspected spam threshold will be considered legitimate.
Messages above the suspected
threshold but below the positive-identification threshold will be considered to
be suspected spam.
You can configure
your anti-spam solution to reflect the spam tolerance levels of your
organization by customizing the Positive and Suspected spam thresholds in each
mail policy.
You can change the
positively identified spam threshold to a value between 50 and 99. You can
change the threshold for suspected spam to any value between 25 and the value
you specified for positively-identified spam.
When you change the
thresholds:
Specifying a lower number (a
more aggressive configuration) identifies more messages as spam and may produce
more false positives. This provides a lower risk that users will see spam but a
higher risk of having legitimate mail marked as spam.
Specifying a higher number (a more conservative configuration) identifies fewer messages as spam and may deliver more spam.
This provides a higher risk of users seeing spam but less risk that legitimate mail will be withheld as spam. Ideally, if
set up correctly, the message subject will identify the message as likely spam and message will be delivered.
You can define
separate actions to take on positively-identified and suspected spam. For
example, you may want to drop “positively identified” spam but quarantine
“suspected” spam.
Configuration
Examples: Actions for Positively Identified versus Suspected Spam
Spam
Sample
Actions
(Aggressive)
Sample
Actions
(Conservative)
Positively Identified
Drop
Deliver with “
[Positive Spam] ” added to the subject of
messages, or
Quarantine
Suspected
Deliver
with “
[Suspected Spam] ” added to the subject of
messages
Deliver
with “
[Suspected Spam] ” added to the subject of
messages
The aggressive
example tags only suspected spam messages, while dropping those messages that
are positively identified. Administrators and end-users can check the subject
line of incoming message for false positives, and an administrator can adjust,
if necessary, the suspected spam threshold.
In the conservative
example, positively identified and suspected spam is delivered with an altered
subject. Users can delete suspected and positively identified spam. This method
is more conservative than the first.
For a further
discussion of aggressive and conservative policies in mail policies, see
Managed Exceptions.
Unwanted Marketing
Messages From Legitimate Sources
If you had configured Marketing Email Settings under anti-spam settings for a mail policy, after upgrading to AsyncOS 9.5
for Email, Marketing Email Settings under anti-spam settings will be moved under graymail settings of the same policy. See
Managing Spam and Graymail.
Using Custom Headers
to Redirect URLs in Suspected Spam to the Cisco Web Security Proxy:
Configuration Example
You can rewrite
URLs in suspected spam so that when a recipient clicks a link in the message,
the request is routed through the Cisco Web Security proxy service, which
evaluates the safety of the site at click time and blocks access to known
malicious sites.
Enabling Different
Anti-Spam Scanning Engines in Different Mail Policies: Configuration Example
When using the System
Setup Wizard (or systemsetup command in the CLI), you are presented with option
to enable either Cisco Intelligent Multi-Scan or the Cisco Anti-Spam engine.
You cannot enable both during system setup, but after system setup is complete
you can enable the anti-spam solution that you didn’t choose, by using the
Security Services menu.
After the system is
set up, you can configure the anti-spam scanning solution for incoming mail
policies via the
Mail
Policies > Incoming Mail Policies page. (Anti-spam scanning is
typically disabled for outgoing mail policies.) You can even disable anti-spam
scanning for a policy.
In this example, the
default mail policy and the “Partners” policy are using the Cisco Anti-Spam
scanning engine to quarantine positive and suspected spam.
Figure 3. Mail Policies -
Anti -spam Engine Per Recipient
To change the
Partners policy to use Cisco Intelligent Multi-Scan and scan for unwanted
marketing messages, click on the entry in the Anti-Spam column corresponding
with the Partners row (“use default”).
Select Cisco
Intelligent Multi-Scan for the scanning engine, and select Yes to enable
unwanted marketing message detection. Use the default settings for unwanted
marketing message detection.
The following figure
shows Cisco Intelligent Multi-Scan and unwanted marketing message detection
enabled in a policy.
Figure 4. Mail Policies -
Enabling Cisco Intelligent Multi-scan
After submitting and
committing the changes, the mail policy looks like this:
Figure 5. Mail Policies -
Intelligent Multi-Scan Enabled in Policy
Protecting
Email Gateway-Generated Messages From the Spam Filter
Because automated email messages that are sent from the
email gateway (such as email alerts and scheduled reports) may contain URLs or other information that may cause them to be incorrectly
identified as spam, you should do the following to ensure their delivery:
If either
anti-spam scanning engine is enabled for a mail policy, each message that
passes through that policy will have the following headers added to the
message:
X-IronPort-Anti-Spam-Filtered: true
X-IronPort-Anti-Spam-Result
The second header
contains information that allows Cisco Support to identify the rules and engine
version used to scan the message. Result information is encoded proprietary
information and is not customer-decodable.
Cisco Intelligent Multi-Scan
also adds headers from the third-party anti-spam scanning engines.
You can define additional
custom headers to be added to all messages for a given mail policy that are
positively identified as spam, suspected to be spam, or identified as unwanted
marketing mail. See
Defining Anti-Spam Policies.
Reporting
Incorrectly Classified Messages to Cisco
Messages that appear
to be incorrectly classified may be reported to Cisco for analysis. The
reported messages are used to enhance the accuracy and effectiveness of the
product.
You can report
incorrectly classified messages that belong to the following categories:
Missed spam
Message marked as a spam,
but is not a spam
Missed marketing message
Message marked as a
marketing message, but is not a marketing message
Enter your organization’s domain name in the Domain field to register your domain with the portal.
Note
Make sure that you enter a valid domain name, for example, example.com is the domain name in the following email address: user@example.com. If you have multiple domains in your organization, make sure that you add all the domains.
Check the I own this domain check box if you are the owner of the domain entered in step 'd.'
Note
If you do not check the 'I own this domain' check box, then you will only have domain view access rights. For more information,
see the Cisco Talos Email Status Portal Help page at https://talosintelligence.com/tickets/email_submissions/help
Click Submit.
After you click Submit, an email with a 6-digit character verification code is automatically sent to postmaster@domain.com (where domain.com is the domain you entered in step 'd') to confirm the domain ownership.
If your organization is not using postmaster@domain.com or your administrator does not have access to the postmaster mailbox, create a message filter (on all your
email gateways) to redirect messages from SubmissionPortal@cisco.com sent to postmaster@domain.com to a different email address. The following is a sample message filter:
redirect_postmaster: if (rcpt-to == "postmaster@domain.com") AND (mail-from == "^SubmissionPortal@cisco.com$") { alt-rcpt-to
("admin@domain.com"); }
Enter the 6-digit character verification code in the Domain Ownership Verification Code dialog box to confirm the domain ownership.
Click Submit Verification Code.
After you click the Submit Verification Code button, you are automatically provided with admin access rights. A registration
ID is auto generated that can be viewed in the Manage Accounts section of the portal. You can use the registration ID for
all the
email gateways in your organization.
Note
A registration ID is a unique identifier to identify submissions made from the Cisco Email Security Gateways that belong to
a particular organization.
Registering when an administrator in your organization is already registered on the portal:
Enter your organization’s domain name in the Domain field to register your domain with the portal.
Note
Make sure that you enter a valid domain name, for example, example.com is the domain name in the following email address: user@example.com. If you have multiple domains in your organization, make sure that you add all the domains.
Click Submit.
After you click Submit, an email notification is sent to the administrator who is already registered on the portal. This administrator
must log in to the portal, and click Approve in the Permission Requests section of Manage Accounts to approve your registration request.
After your registration request is approved, a registration ID is auto generated that can be viewed in the Manage Accounts
section of the portal. You can use the registration ID for all the
email gateways in your organization.
Note
A registration ID is a unique identifier to identify submissions made from the Cisco Email Security Gateways that belong to
a particular organization.
Step 2
Add the registration ID generated from Cisco Talos Email Status Portal for all the
email gateways in your organization.
Log in to your
email gateways using the web interface.
Go to System Administration > Cisco Talos Email Status Portal Registration.
If your
email gateway is part of a cluster, set the mode to cluster level.
Click Set Registration ID.
Enter the registration ID obtained from the Cisco Talos Email Status Portal in the Registration ID field.
Submit and commit your changes.
If your
email gateway is not part of a cluster, you must repeat steps 1 through 6 on all the
email gateways in your organization.
You can also use the portalregistrationconfig command in CLI to set the registration ID.
How to Report
Incorrectly Classified Messages to Cisco
After you report an incorrectly classified message to Cisco, you will receive an email notification based on the option you
select under the Email Notification and Reports button in the Manage Account section of the portal.
Note
The 'My Submission Notifications' and 'My Submission Reports' options under the 'Email Notification and Reports' button are
set to off by default. For more information, see the Cisco Talos Email Status Portal Help page at https://talosintelligence.com/tickets/email_submissions/help
Cisco Email Security
Plug-In is a tool that allows users (email administrators and end users) to
report incorrectly classified messages to Cisco using Microsoft Outlook. When
you deploy this plug-in as part of Microsoft Outlook, a reporting menu is added
to the Microsoft Outlook web interface. You can use the plug-in menu to report
incorrectly classified messages.
Forwarding Incorrectly Classified Message as an Attachment
Depending on the category of the message, you can forward each incorrectly classified message as an RFC 822 attachment to
the following addresses as shown in the table below:
Email Submission
Definition
Submission Method
User Consideration for Submitting
Spam / Phish
Unsolicited and undesired. Spam/Phish is never legitimate and may also be malicious (phish, virus, malware, scams, etc.)
spam@access.
ironport.com
phish@access.
ironport.com
virus@access.
ironport.com
Outlook Plugin ‘Spam’, ‘Phish’, or ‘Virus’ button
Delivered to the user’s inbox, but user considers message to be spam or Phish.
Detected as spam, but user considers message legitimate.
Legiti
mate
Legitimate (good) email, not spam. Also known as ‘Ham.’
ham@access.
ironport.com
Outlook Plugin ‘Not Spam’ button
Marketing/graymail messages not detected as marketing/graymail.
Market
ing / Graymail
Marketing is legitimate (not Spam) email that is commercial bulk email. Usually subscription based, sometimes unwanted.
Users may have knowingly or unknowingly solicited mail from the sender. For example swiping a badge at a conference or making
an online purchase, etc. Legitimate subscription based marketing email will have a working unsubscribe mechanism.
Graymail is a broader category that Includes Marketing as well as other legitimate bulk mail.
ads@access.
ironport.com
Outlook Plugin ‘Marketing’ button
Detected as spam, but user considers message legitimate
Not Market
ing / Graymail
Legitimate email (not Spam) that is not bulk and not subscription based. Usually person-to-person and/or transactional.
not_ads@access.
ironport.com
Detected as Marketing/Graymail, but user considers the message to be transactional or otherwise not Marketing/Graymail.
You can achieve best results if you use one of the following email programs to forward the message:
Apple Mail
Microsoft Outlook for Mac
Microsoft Outlook Web App
Mozilla Thunderbird
Caution
If you are using Microsoft Outlook 2010, 2013, or 2016 for Microsoft Windows, you must use the Cisco Email Security Plug-In
or the Microsoft Outlook Web App to report incorrectly classified messages. This is because Outlook for Windows may not forward
the message with the required headers intact. Also, use the mobile platforms only if you can forward the original message
as an attachment.
How to Track Your Submissions
After your receive an email notification with the submission details, you can view and track your submission on Cisco Talos
Email Status Portal.
Determining Sender
IP Address In Deployments with Incoming Relays
If one or more mail exchange/transfer agents (MX or MTA), filtering servers, etc. stand at the edge of your network, between
your
email gateway and the external machines that are sending incoming mail, then your
email gateway cannot determine the IP addresses of the sending machines. Instead, mail appears to originate from the local MX/MTA. However,
IronPort Anti-Spam and Cisco Intelligent Multi-Scan (using the IP Reputation Service) depend on accurate IP addresses for external senders.
The solution is to configure your
email gateway to work with incoming relays. You specify the names and IP addresses of all of the internal MX/MTAs connecting to the
email gateway, as well as the header used to store the originating IP address.
The following figure shows a very basic example of an incoming relay. Mail from IP address 7.8.9.1 appears to come from IP
address 10.2.3.4 because the local MX/MTA is relaying mail to the
email gateway.
Figure 6. Mail Relayed by MX/MTA — Simple
The following figure shows two other, slightly more complicated examples of how mail may be relayed inside the network and
how mail may be processed by several servers within the network before it is passed to the
email gateway. In example A, mail from 7.8.9.1 passes through the firewall and is processed by an MX and an MTA before being delivered
to the
email gateway. In example B, mail from 7.8.9.1 is sent to a load balancer or other type of traffic shaping appliance and is sent to any
one of a range of MXs prior to being delivered to the
email gateway.
Figure 7. Mail Relayed by MX/MTA — Advanced
Configuring the
Email Gateway to Work with Incoming Relays
Determine whether you will
use custom or received headers to identify the IP address of the original
external sender.
If you will use custom
headers:
Determine the exact header
that will label the originating IP address of relayed messages.
For each MX, MTA, or other
machine that connects to original external senders, set up that machine to add
the header name and the IP address of the original external sender to incoming
messages.
Procedure
Step 1
Select
Network
> Incoming Relays.
Step 2
Click
Add
Relay.
Step 3
Enter a name
for this relay.
Step 4
Enter the IP address of the MTA, MX, or other machine that connects to the
email gateway to relay incoming messages.
You can use IPv4 or IPv6 addresses, standard CIDR format, or an IP address range. For example, if you have several MTAs at
the edge of your network receiving email, you might want to enter a range of IP addresses to include all of your MTAs, such
as 10.2.3.1/8 or 10.2.3.1-10.
For IPv6 addresses, AsyncOS supports the following formats:
2620:101:2004:4202::0-2620:101:2004:4202::ff
2620:101:2004:4202::
2620:101:2004:4202::23
2620:101:2004:4202::/64
Step 5
Specify the
header that will identify the IP address of the original external sender.
When entering a
header, you do not need to enter the trailing colon.
Select the
header type:
Choose
custom headers (recommended) or Received headers.
For custom
headers:
Enter the
header name that you configured the relaying machine to add to relayed
messages.
For
example:
SenderIP
or
X-CustomHeader
For
Received headers:
Enter the
character or string after which the IP address will appear. Enter the number
for the “hop” to check for the IP address.
Using custom headers
is the recommended method of identifying original senders. The machine
connecting to the original sender needs to add this custom header. The value of
the header is expected to be the IP address of the external sending machine.
For example:
SenderIP: 7.8.9.1
X-CustomHeader: 7.8.9.1
If your local MX/MTA
can receive mail from a variable number of hops, inserting a custom header is
the only way to enable the Incoming Relays feature. For example, in the
following figure, both path C and D lead to IP address 10.2.3.5; however, path
C has two hops and path D has one. Because the number of hops can vary in this
situation, you must use a custom header in order to have Incoming Relays
configured correctly.
Figure 8. Mail Relayed by
MX/MTA — Variable Number of Hops
If configuring the MX/MTAs to include a custom header containing the sending IP address is not an option, you can configure
the incoming relays feature to attempt to determine the sending IP address by examining the “Received:” headers in the message.
Using the “Received:” header will only work if the number of network “hops” will always be constant for an IP address. In
other words, the machine at the first hop (10.2.3.5 in Figure - Mail Relayed by MX/MTA — Advanced) should always be the same number of hops away from the edge of your network. If incoming mail can take different paths (resulting
in a different number of hops, as described in Figure - Mail Relayed by MX/MTA — Variable Number of Hops) to the machine connecting to your
email gateway, you must use a custom header (see Custom Header).
Specify a parsing character or string and the number of network hops (or Received: headers) back to look. A hop is basically
the message traveling from one machine to another (being received by the
email gateway does not count as a hop. See Configuring Logs to Specify Which Headers Are Usedfor more information). AsyncOS looks for the first IP address following the first occurrence of the parsing character or string
in the Received: header corresponding to the number of specified hops. For example, if you specify two hops, the second Received:
header, working backward from the
email gateway is parsed. If neither the parsing character nor a valid IP address is found, the
email gateway uses the real IP address of the connecting machine.
For the following
example mail headers, if you specify an opening square bracket ( [ ) and two
hops, the IP address of the external machine is 7.8.9.1. However, if you
specify an closing parenthesis ( ) ) as the parsing character, a valid IP
address will not be found. In this case, the Incoming Relays feature is treated
as disabled, and the IP of the connecting machine is used (10.2.3.5).
In the example in
Figure - Mail
Relayed by MX/MTA — Advanced the incoming relays are:
Path A — 10.2.3.5 (with 2
hops when using received headers) and
Path B — 10.2.6.1 (with 2
hops when using received headers)
The following table shows example email headers for a message as it moves through several hops on its way to the
email gateway as in Figure - Mail Relayed by MX/MTA — Advanced. This example shows extraneous headers (ignored by your
email gateway) which are present once the message has arrived in the recipient’s inbox. The number of hops to specify would be two.
Table 1. A Series of
Received: Headers (Path A Example 1)
1
Microsoft Mail Internet Headers Version 2.0
Received: from smemail.rand.org ([10.2.2.7]) by smmail5.customerdoamin.org with
Microsoft SMTPSVC(5.0.2195.6713);
Received: from ironport.customerdomain.org ([10.2.3.6]) by
smemail.customerdoamin.org with Microsoft SMTPSVC(5.0.2195.6713);
2
Received: from mta.customerdomain.org ([10.2.3.5]) by ironport.customerdomain.org
with ESMTP; 21 Sep 2005 13:46:07 -0700
3
Received: from mx.customerdomain.org (mx.customerdomain.org) [10.2.3.4]) by
mta.customerdomain.org (8.12.11/8.12.11) with ESMTP id j8LKkWu1008155 for
<joefoo@customerdomain.org>
4
Received: from sending-machine.spamham.com (sending-machine.spamham.com [7.8.9.1])
by mx.customerdomain.org (Postfix) with ESMTP id 4F3DA15AC22 for
<joefoo@customerdomain.org>
5
Received: from linux1.thespammer.com (HELO linux1.thespammer.com) ([10.1.1.89])
by sending-machine.spamham.com with ESMTP;
Received: from exchange1.thespammer.com ([10.1.1.111]) by linux1.thespammer.com
with Microsoft SMTPSVC(6.0.3790.1830);
Subject: Would like a bigger paycheck?
Date: Wed, 21 Sep 2005 13:46:07 -0700
From: "A. Sender" <asend@otherdomain.com>
To: <joefoo@customerdomain.org>
Notes for the above
table:
The
email gateway ignores these headers.
The
email gateway receives the message (not counted as a hop).
First hop (and incoming
relay).
Second hop. This is the
sending MTA. The IP address is 7.8.9.1.
The
email gateway ignores these Microsoft Exchange headers.
The following table
shows the headers for the same email message, without the extraneous headers
Table 2. A Series of
Received: Headers (Path A Example 2)
1
Received: from mta.customerdomain.org ([10.2.3.5]) by ironport.customerdomain.org
with ESMTP; 21 Sep 2005 13:46:07 -0700
2
Received: from mx.customerdomain.org (mx.customerdomain.org) [10.2.3.4]) by
mta.customerdomain.org (8.12.11/8.12.11) with ESMTP id j8LKkWu1008155 for
<joefoo@customerdomain.org>;
3
Received: from sending-machine.spamham.com (sending-machine.spamham.com [7.8.9.1])
by mx.customerdomain.org (Postfix) with ESMTP id 4F3DA15AC22 for
<joefoo@customerdomain.org>;
The following figure
shows the incoming relay for path A (above) as configured in the Add Relay page
in the GUI:
Figure 9. A Configured
Incoming Relay with Received Header
The Incoming Relays feature provides the various IP Reputation Service related filter rules (reputation, no-reputation) with
the correct IP Reputation score.
Incoming Relays, HAT, IP Reputation Score, and Sender Groups
HAT policy groups do not currently use information from Incoming Relays. However, because the Incoming Relays feature does
supply the
Reputation score, you can simulate HAT policy group functionality via message filters and the $reputation variable.
Incoming Relays and Directory Harvest Attack Prevention
If a remote host attempts a directory harvest attack by sending messages to the MX or MTA serving as an incoming relay on
your network, the
email gateway drops the connection from the incoming relay if the relay is assigned to a sender group with a mail flow policy with Directory
Harvest Attack Prevention (DHAP) enabled. This prevents all messages from the relay, including legitimate messages, from reaching
the
email gateway. The
email gateway does not have the opportunity to recognize the remote host as the attacker and the MX or MTA that’s acting as the incoming
relay continues to receive mail from the attacking host. To work around this issue and continue receiving messages from the
incoming relay, add the relay to a sender group with a mail flow policy that has unlimited messages for DHAP.
Incoming Relays and Trace
Trace returns the Incoming Relay’s IP Reputation Score in its results instead of the reputation score for the source IP address.
Incoming Relays and Email Security Monitor (Reporting)
When using Incoming Relays:
Email Security Monitor
reports include data for both the external IP and the MX/MTA. For example, if
an external machine (IP 7.8.9.1) sent 5 emails through the internal MX/MTA (IP
10.2.3.4), Mail Flow Summary will show 5 messages coming from IP 7.8.9.1 and 5
more coming from the internal relay MX/MTA (IP 10.2.3.5).
The IP Reputation score is not reported correctly in the Email Security Monitor reports. Also, sender groups may not be resolved
correctly.
Incoming Relays and Message Tracking
When using Incoming Relays, the Message Tracking Details page displays the relay’s IP address and the relay’s IP Reputation Score for a message instead of the IP address and reputation score of the original external sender.
Incoming Relays and
Logging
In the following log example, the IP Reputation score for the sender is reported initially on line 1. Later, once the Incoming Relay is processed, the correct
IP Reputation score is reported on line 5.
The following example shows a typical log entry containing Incoming Relay information:
Wed Aug 17 11:20:41 2005 Info: MID 58298 IncomingRelay(myrelay): Header Received found, IP 192.168.230.120 being used
Configuring Logs to
Specify Which Headers Are Used
Your
email gateway only examines the headers that were present when the message was received. So, additional headers added locally (such as
Microsoft Exchange headers, etc.) or when the message is received by the
email gatewayare not processed. One way to help determine which headers are used is to configure AsyncOS logging to include the headers
you use.
Test your
configuration using the
X-advertisement: spam
header.
For testing
purposes, Cisco Anti-Spam considers any message with an X-header formatted as
X-Advertisement: spam to be spam.
The test
message you send with this header is flagged by Cisco Anti-Spam, and you can
confirm that the actions you configured for the mail policy (Defining Anti-Spam Policies)
are performed.
Send a test
email that includes the following header to a user in that mail policy:
X-Advertisement: spam
Use SMTP
commands with Telnet to send this message to an address to which you have
access.
Step 3
Check the
mailbox of the test account and confirm that the test message was correctly
delivered based upon the actions you configured for the mail policy.
For example:
Was the
subject line altered?
Was your
additional custom header added?
Was the
message delivered to an alternate address?
Testing Anti-Spam
Configuration: Example Using SMTP
For this example, the
mail policy must be configured to receive messages for the test address and the
HAT must accept the test connection.
# telnet IP_address_of_IronPort_Appliance_with_IronPort_Anti-Spam port
220 hostname ESMTP
helo example.com
250 hostname
mail from: <test@example.com>
250 sender <test@example.com> ok
rcpt to: <test@address>
250 recipient <test@address>
ok
data
354 go ahead
Subject: Spam Message Test
X-Advertisement: spam
spam test
.
250 Message MID accepted
221 hostname
quit
Ways Not to Test
Anti-Spam Efficacy
Because IronPort
AntiSpam and Cisco Intelligent Multi-Scan rules are added quickly to prevent
active spam attacks and quickly expire once attacks have passed, you should not
test efficacy using any of the following methods:
Evaluating using resent or forwarded mail or cut-and-pasted spam
messages.
Mail lacking the proper headers, connecting IP, signatures, etc.
will result in inaccurate scores.
Testing “hard spam” only.
Removing the “easy spam” using IP Reputation Service, blocked lists, message filters, etc. will result in a lower overall catch rate percentage.
Resending spam caught by another anti-spam vendor.
Testing older messages.
The scanning engine adds and removes rules rapidly based on current
threats. Testing using old messages will therefore lead to inaccurate test
results.