Our Technology

What Is Mercury?

It’s the name of our core messaging engine. Although we offer many different services based around messaging, they all run on one unified platform, called Mercury.

Divide and Conquer

Mercury isn’t one server, it’s lots of them. Our servers are organised into task groups. Each task group has a specific job to do, and no other. Every server has a number of counterparts all doing the same job. In the event of server failure, the counterparts take up the slack providing uninterrupted service.

To put this into context, lets have a look at the lifecycle of an email:

• Email Arrives: The email is delivered to our MX server group, who receive emails and plan their route through the system. This way, emails are not sent to server groups that don’t have a part in that email’s lifecycle, improving load management and efficiency. At this point, a few basic security checks are performed to ensure the integrity of the sender server and message content. The obvious fakes are removed at this point.
• Filtering: Based on the destination account’s settings, the message is now scanned for spam and viruses. Again, task groups are dedicated to this role.
• Delivery: ‘Good’ emails are now handed to the mailbox or relay task groups for storage or onward delivery to customer SMTP servers.
• Collection: Again, a separate bank of servers handle access via POP3 or IMAP for collection/viewing of messages.

Load Control: We Love Surprises

Anyone who has spent time looking after email services will be able to tell you some story or other about the day 25,000 emails arrived to one account in an hour as some ‘zombie’ PC somewhere got a bit over-enthusiastic.

When we designed Mercury, our team had seen this sort of thing before so protection measures against sudden mailbombing and mail-loops (when two mail servers start sending the same message back and forward, thousands of times an hour) were built directly into the system.

Mercury’s monitoring systems can not only detect sudden increases in mail activity but also takes measures to prevent it spiralling out of control. These ‘booby traps’ also prevent massive influxes of spam.

Long-term trend monitoring ensures our servers never get to the point of overloading. Our modular system allows any task group to have additional servers introduced at any time, increasing capacity.

Here’s Looking At You

All this architecture is monitored 24/7 by our load analysis systems. Many recovery procedures are automated to get a fix in place in the quickest time possible. Regardless of the situation, our support team is automatically alerted, day or night.

Can you say all that about your email service? If not, maybe we should talk.