There are three features that control which IP interface will be used for message delivery:
The deliveryconfig CLI command controls primary IP interface for all mail delivery. By default, this is set to "Auto," which will deliver mail based on destination IP subnet. If the destination IP is a local subnet, it will use an IP interface that is on that subnet. If the destination IP is not a local subnet, the messages will be delivered from an IP interface on the same subnet as the default gateway.
The altsrchost table overrides the deliveryconfig setting for specific messages. In this table, you must specify the source IP address or Envelope From to match on and the IP interface to use when the match occurs. This table is configured using the altsrchost CLI command.
The alt-src-host Message Filter action can be used to specify the IP interface for message delivery as an action for your choice of criteria using Message Filter rules. This will override both deliveryconfig and the altsrchost table. Should you choose to set up a filter, reviewing the section titled Message Filters in the AsyncOS Advanced User Guide is highly recommended.
As you can see, there are several options for how to configure the IP address used for message delivery. In order to decide which method to use, you have to determine:
From which IP interface should the majority of your messages be delivered?
Are there any special cases?
For each special case, what is the unique trigger that should be used to set the IP address used for message delivery?
Once you know the answers to these questions, you will know which method or combination of methods is best for your environment. More details on these features can be found in the Content Security User Guides.
For more information about message filters, email delivery parameters and setting up virtual gateway addresses, see the AsyncOS User Guide on the Cisco Support Site.