It’s been way too long since the last post. All I can say is that we’ve been very, very busy. Of course, in this (and any) economy, busy is good. Ok, here’s what we’ve been up to…
We are pleased to announce that we have completed the migration of all customers over to our new distributed cloud based spam filtering environment. We always knew that this was the direction we wanted to go in, even before our service was launched. However, since we are used to dealing with very reliable data centers, the first release of our service wasn’t cloud based and simply used relay servers as an emergency fallback. The original system looked like this:
The location 2 and 3 servers were only engaged when there was a problem with location 1, which ensured that no email was ever lost. However, since these locations only hosted relay servers, in the event of trouble they had to wait for the primary server at location 1 to come back online, which could cause email delays.
Our new distributed cloud filtering system resolves this problem:
This configuration splits up incoming email for each domain between two totally separate data centers, using round robin DNS MX records for the primary MX record (“MX 1″). In the event that one of these locations has trouble, the “MX 2″ and “MX 3″ records point directly to the two locations so that sending mail servers will always be able to find an available server to send to. Since this design eliminates the need of relay servers, this eliminates most any threat of email delays. Yay! 
Oh, and this setup only requires 3 MX records, but, we’ve added support for a fourth. The fourth record will be used when we expand the distributed cloud to include three locations. For now, the “MX 4″ record mirrors “MX 1″ but will later point to the third location. To take advantage of the third location when it comes online, you will need to update your MX records according to the instructions now displayed on the Settings -> mail host page when you login to the SpamHero back office.
If this upgrade doesn’t look like that big of a deal, then perhaps it’s because the sketch above is a little too simplified. This was a big project for us, and it held up a lot of other projects and upgrades that we’ve been anxious to get to. For example, one upgrade it had held up was a speed increase for those users with very large quarantines… if you are one of those users, you’ll notice a huge speed increase there.
With this distributed cloud project out of the way, you should see a steady flow of improvements and new features again. I’ll try to make mention of some of the more notable ones here as they are rolled out.
