Kerio Mailserver supports both Caller-ID and SPF. Not many
other servers support these, but a few of the big people do, so it
could be worthwhile to turn these on. There's usually no downside:
if a domain doesn't use these, setting these checks won't hurt anything.
It's only when a domain does implement these but a "wrong" server connects to you that the filters are triggered.
There are two similar technologies available for performing
email policy records check in Kerio MailServer. The first one
is Caller-ID created by Microsoft, the other one is a project
named SPF (Sender Policy Framework). Both technologies provide
explicit verification of message senders. During this verification
process, the IP addresses of SMTP servers that send mail from
the specific domain are published. For each domain that supports
at least one of the above technologies, a TXT record is stored
in DNS with a list of IP addresses that send email from the
specific domain. Kerio MailServer then compares the IP address
of the SMTP server with IP addresses contained in this DNS
record. This method guarantee verification of sender's
trustworthiness for each message. If the DNS record does not
contain the IP address the message was sent from, such message
has a falsified address and it is considered as spam. This
way, it is quite easy to distinguish, whether the message is
spam or not.
Messages received from a server that has no IP address list in the DNS record will be always delivered. For the email policy purposes, these emails will not be considered.
As Kerio recommends, you probably shouldn't block outright if a message
violates these policies. Instead, just increase the spam score. That
helps prevent accidents - a missing but legitimate mail server sending
mail that has no other spam characteristics won't be affected. It also lets you set custom rules that override spam setting for specific needs -
you can't override a block.
Sounds like there's nothing to lose and everything to gain. If someone pretends to be sending from Microsoft.com but isn't using a legitimate Microsoft mail server, the Caller-ID will catch them because Microsoft implements this. If someone spoofs mail from a domain that doesn't use
either of these methods, the checking has no affect, positive or negative. As more domains start using either of these, the filters become more
However, I recently heard of a situation where this setting caused
a problem. In this configuration, a user had accounts that he would forward to his main account. These were different domains hosted on the same server. If a company like Microsoft tried to send to one of those forwarded accounts, the Kerio server would (incorrectly) think that it should find itself (the domain he forwarded from) in Microsoft Caller-ID records when the forwarded mail was checked.
Of course that is wrong - that check should bypass on forwards within the server itself - it should only be done on first arrival - but for at least right now, it will fail under
these conditions. The solution is simple enough - include the forwarding domain in the list of addresses that aren't checked for Caller-ID or SPF.