We have two Cisco IronPort C360 appliances doing antispam and antivirus filtering before mail makes its way into Exchange. These have not been in a DMZ, and we're now making the effort to move these into a DMZ.
We use our corporate AD servers for LDAP recipient validation (so that invalid recipients are dropped). We were planning to move an LDAP server into the DMZ with the appliances, with email addresses pushed to it from production, for recipient validation. *BUT* I came across SMTP call-ahead, which seems like it would suit our needs better since we wouldn't need a dedicated LDAP server and our architecture would be more secure...
IronPort would open an SMTP session to Exchange to verify the recipient. Exchange hub servers don't do this out of the box, antispam agents need to be installed. This fellow has a good guide: This fellow has a good guide: http://www.jjclements.co.uk/2010/09/23/exchange-2010-recipient-filtering-on-a-hub-transport-server/
I'm wondering if what we're doing is a terrible idea. Does anyone have any experience with SMTP call-ahead? I would appreciate any thoughts, validation, or feedback.