Configuring an SMTP Connector

 

The primary uses of an SMTP connector are to connect to the Internet or to other mail systems and to define additional options on an SMTP Internet gateway. Because an SMTP connector creates an isolated route for Internet mail, it eases administration and troubleshooting if mail flow problems occur.

This section focuses on the connector's use as a connection method to deliver Internet mail. To configure an SMTP connector to deliver Internet mail, you first must consider the following configuration requirements.

How to Route Mail for Outbound Delivery?

When you configure a connector, you can either use DNS to route all outgoing mail through the connector, or you can specify a smart host to which the connector routes all mail.

Using DNS to route all outgoing mail through the connector

If you use DNS to route outgoing mail, the SMTP connector uses DNS to resolve the IP address of the remote SMTP server, and then it delivers the mail.

If you select this routing method, verify the following information:

  • Verify that your DNS server can successfully resolve names on the Internet.

  • If you use an external DNS server to resolve names, and this server is configured at the SMTP virtual server level (that is, using a different DNS server than the one specified on your network connection), make sure that this external DNS server can resolve names on the Internet.

Specifying a smart host

The smart host handles DNS resolution and delivers the mail. Although you can specify a smart host on an SMTP virtual server, it is a good idea to set the smart host on the connector itself. The smart host setting on the SMTP connector overrides any smart hosts configured on the SMTP virtual server.

If you select this routing method, you specify an IP address or name for the smart host. The IP address and name for the smart host must meet the following requirements:

  • If you specify an IP address for the smart host   Enclose the IP address in brackets (for example, [10.0.0.1]), and make sure that the IP address is not the IP address of the Exchange server.

  • If you specify a name for the smart host   Ensure that the name is a fully qualified domain name (FQDN). (For example, "Server Name" is not an FQDN. However, servername.contoso.com is an FQDN.) Also, make sure that the name is not the FQDN of the Exchange server.

If you do not have a smart host in your network, contact your Internet service provider (ISP) to determine what IP address or FQDN to use for the smart host. After you have the IP address or FQDN, make sure that the IP address or FQDN meets the previous requirements.

Which Servers to Use as Local Bridgehead Servers?

An SMTP virtual server hosts a connector. When you create a connector, you designate at least one Exchange server and one SMTP virtual server as bridgehead servers. The connector inherits size restrictions and other settings from the SMTP virtual server. However, you can override these settings on the connector. You can also designate multiple bridgehead servers for load balancing, performance, and redundancy.

To send outbound mail, the connector uses the outbound port configured on the SMTP virtual server. If your organization sends lots of mail externally, it is a good idea to designate dedicated Exchange servers and SMTP virtual servers as gateway servers or bridgehead servers receiving Internet mail. Using dedicated servers as gateway servers means that other mailbox servers do not have to assume the additional overhead of a gateway server.

Which Domains Should Be Included in the Address Space?

The address space defines the mail addresses or domains for the e-mail messages that you want routed through a connector. For example, an address space of * (asterisk) encompasses all external domains. A connector with this address space is can route all external e-mail messages.

Exchange routes messages through a connector based on the closest match to an address space. If you had a connector with the * address space and then created a second connector with an address space of *.net, Exchange would route all mail sent to a domain with a .net extension through the second connector. This routing difference occurs because Exchange selects the connector that has the most similar address space to the outbound mail.

On connectors with an identical address space, costs work the same way as they do on routing group connectors. For example, you create two SMTP connectors to the Internet, Connector1 and Connector2, and each has the address space of *. Because Connector1 has better network connectivity, you always want to use this connector (unless it becomes unavailable) to send mail to the Internet, and you give Connector1 a cost of 1. Then, you give Connector2 a cost of 2. As long as Connector1 is operating correctly, Exchange always sends messages through that connector because it has the lowest cost. If Connector1 becomes unavailable, Exchange uses the connector with the next lowest cost, Connector2.

Important

Do not list your inbound domains on an SMTP address space for a connector. Your inbound domains are listed in your recipient policies. (For more information, see "Configuring Recipient Policies.") If you list some or all your inbound domains in the SMTP address space, you may receive non-delivery reports (NDRs) that indicate a mail loop. (These NDRs may have the diagnostic code 5.3.5.) By specifying domains on the Address Space tab in the connector's Properties dialog box, you can configure these domains as routable domains.

What Is Appropriate Scope for the Connector?

You can select either a whole organization or a routing group for the connector's scope. For example, you have two routing groups and each routing group has a server that has an SMTP connector to send mail to the Internet. For this configuration, you may choose to specify a routing group scope for each of the connectors. Specifying a routing group scope forces the servers in each routing group to use the connector in that routing group. However, a routing group scope also means that, if the group's SMTP connector becomes unavailable, messages queue in the routing group until the connector becomes available again. Because of the restrictions imposed by a routing group scope, you would most likely set an SMTP connector to this scope if it is acceptable to have messages queuing when a connector becomes unavailable, or if the network cannot handle the extra traffic from one routing group sending Internet mail through an SMTP connector of another routing group. Otherwise, you must assign the connector an organization-wide scope and permit users in your whole organization to use any acceptable SMTP connector.

For detailed instructions, see How to Create an SMTP Connector for Internet Mail Delivery.

For detailed instructions, see How to Enable Anonymous Access for an SMTP Connector.