We often talk about our unique outbound email infrastructure when talking to customers, but we’ve failed to make it a well known part of our product. We simply gloss over it with a mention about taking IP reputation off the table on our front page. You might wonder what exactly we do to achieve this, and we’d like to share. First, let’s start with a map of our outbound flows:
That might be a bit hard to see, but here’s what happens:
Email goes out to one of two redundant outbound filter servers. From there, it goes out to one of three random relay servers. These relay servers house roughly 250 IP addresses each (Chicago, as an exception, has over 500 IPs from two different ranges), and they rotate outbound mail randomly across them. Each relay has different IP ranges, they are not just the same ranges broken out and split up. If an email gets rejected from the first try for a reason that relates to IP reputation, it is retried again 2 times from within that same relay using a different IP each time. If it fails those two extra times, the email is sent to our Charlotte relay where it again tries 3 more times. If it fails those 3 times, we send it to a third party mailing service (SpamWall) for final delivery attempts. Depending on the reason for the failure, the email can sit in that queue for up to 3 days before bouncing back to the sender.
So you see what really makes MXroute special on the product side is this outbound infrastructure that takes every measure necessary to get your email to your recipient, no matter the cost.
Note: As of May 22, 2020 there has been a change in the information provided here. It isn’t a significant change, but it does alter some of the details. The Chicago relay server has been removed, and the Dallas relay expanded by roughly 500 more IPs (from a brand new /23). This is only an improvement on the system, and represents the continual work being done behind the scenes to build up the infrastructure.