Intelligent message splintering is the mechanism that allows for differing recipient-based content security rules to be applied independently to message with multiple recipients.
Each recipient is evaluated for each policy in the appropriate mail policy table (Incoming or Outgoing) in a top-down fashion.
Each policy that matches a message creates a new message with those recipients. This process is defined as message splintering:
If some recipients match different policies, the recipients are grouped according to the policies they matched, the message is split into a number of messages equal to the number of policies that matched, and the recipients are set to each appropriate "splinter."
If all recipients match the same policy, the message is not splintered. Conversely, a maximum splintering scenario would be one in which a single message is splintered for each message recipient.
Each message splinter is then processed by anti-spam, anti-virus, Advanced Malware Protection (incoming messages only), DLP scanning (outgoing messages only), Outbreak Filters, and content filters independently in the email pipeline.