Data filtering profiles prevent sensitive information such as credit card or social security numbers from leaving a protected network. The data filtering profile also allows you to filter on key words, such as a sensitive project name or the word confidential. It is important to focus your profile on the desired file types to reduce false positives. For example, you may only want to search Word documents or Excel spreadsheets. You may also only want to scan web-browsing traffic, or FTP.
You can use default profiles, or create custom data patterns. There are two default profiles:
CC# (Credit Card)—Identifies credit card numbers using a hash algorithm. The content must match the hash algorithm in order for data to be detected as a credit card number. This method will reduce false positives. SSN# (Social Security Number)—Uses an algorithm to detect nine digit numbers, regardless of format. There are two fields: SSN# and SSN# (no dash).
Weight and Threshold Values
It is important to understand how the weight of an object (SSN, CC#, pattern) is calculated in order to set the appropriate threshold for a condition you are trying to filter. Each occurrence multiplied by the weight value will be added together in order to reach an action threshold (alert or block).
Example: Filter for Social Security Numbers Only
For simplicity, if you only want to filter files with Social Security Numbers (SSN) and you define a weight of 3 for SSN#, you would use the following formula: each instance of a SSN x weight = threshold increment. In this case, if a Word document has 10 social security numbers you multiply that by the weight of 3, so 10 x 3 = 30. In order to take action for a file that contains 10 social security numbers you would set the threshold to 30. You may want to set an alert at 30 and then block at 60. You may also want to set a weight in the field SSN# (no dash) for Social Security Numbers that do not contain dashes. If multiple settings are used, they will accumulate to reach a given threshold.
Example: Filter for Social Security Numbers and a Custom Pattern
In this example, we will filter on files that contain Social Security Numbers and the custom pattern confidential. In other words, if a file has Social Security Numbers in addition to the word confidential and the combined instances of those items hit the threshold, the file will trigger an alert or block, depending on the action setting.
SSN# weight = 3
Custom Pattern confidential weight = 20
The custom pattern is case sensitive.
If the file contains 20 Social Security Numbers and a weight of 3 is configured, that is 20 x 3 = 60. If the file also contains one instance of the term confidential and a weight of 20 is configured, that is 1 x 20 = 20 for a total of 80. If your threshold for block is set to 80, this scenario would block the file. The alert or block action will be triggered as soon as the threshold is hit.

Related Documentation