In the absence of a filtering service, a receiving mail server will typically receive messages from many thousands of different servers with thousands of different IP addresses. When a domain is protected by an email filter, the receiving mail server should only see inbound mail traffic from that filter - i.e. all message traffic will come from a very small number of IP addresses compared to before.
In some circumstances, anti-spam / anti-virus / intrusion detection software on a firewall and/or on a mail server will interpret the high number of messages from our IP addresses as an indication that we are trying to send spam to that mail server. When this happens, the firewall or anti-spam software at the customer's site may either slow down or block some or all traffic from our servers, thus preventing mail from being delivered.
This can also occur with specific servers, i.e. the receiving mail server may see a high number of incoming connections from just one or two of the filtering servers, and thus may block traffic from those systems but allow traffic from our other servers to be delivered normally.
To solve this problem, the firewall and any other security solution (anti-spam / anti-virus / intrusion detection / etc.) at the customer's site should 'whitelist' our IP addresses, so that messages being delivered from our networks are allowed through without being blocked or delayed.