Руководство Пользователя для Cisco Cisco Email Security Appliance C170

The appliance accepts email using multiple mail protocols: SMTP and QMQP. However, all outgoing 
email is delivered using SMTP, which is why the 

deliveryconfig

 command does not require that the 

protocol be specified.

Several of the features or commands described in this section will affect, or be affected by routing 
precedence. Please see the “Assigning Network and IP Addresses” appendix for more information.

By default, the system uses an IP interface or IP interface group for email delivery. Any currently 
configured IP interface or IP interface group can be set. If no specific interface is identified, AsyncOS 
will use the hostname associated with the default delivery interface in the 

SMTP HELO

 command when 

communicating with recipient hosts. To configure IP interfaces, use the 

interfaceconfig

 command. 

These are the rules for using Auto selection of email delivery interfaces:

If the remote email server is on the same subnet as one of the configured interfaces, then traffic will 
go out on the matching interface.

When set to auto-select, static routes you have configured using 

routeconfig

 take effect. 

Otherwise, the interface that is on the same subnet as the default gateway will be used. If all of the 
IP addresses have an equivalent route to the destination, then the system uses the most efficient 
interface available. 

When the Possible Delivery feature is enabled, AsyncOS treats any message that times-out after the body 
of the message is delivered, but before recipient host acknowledges receipt of the message, as a “possible 
delivery.” This functionality prevents recipients from receiving multiple copies of a message if 
continuous errors at their recipient host prevent acknowledgement of receipt. AsyncOS logs this 
recipient as a possible delivery in the mail logs and counts the message as completed. It is recommended 
that the Possible Delivery feature remains enabled.

You also specify the default maximum number of concurrent connections the appliance makes for 
outbound message delivery. (The system-wide default is 10,000 connections to separate domains.) The 
limit is monitored in conjunction with the per-listener maximum outbound message delivery 
concurrency (the default per listener is 600 connections for private listeners and 1000 connections for 
public listeners). Setting the value lower than the default prevents the gateway from dominating weaker 
networks. For example, certain firewalls do not support large numbers of connections, and this could 
induce Denial of Service (DoS) warnings in these environments.