Today we had our quarterly board meeting. One of the items discussed was our accomplishments with the recent upgrades to our email hosting system, as well as our plans for what’s to come. Here are some details…
We started the design for these upgrades towards the end of last year. The entire process took about 12 months, and it was completed on November 11. The major objectives of the upgrades were:
- to provide complete redundancy for every front-end and back-end application;
- to split all applications into independent server clusters;
- to reduce mail delivery times down to the sub ten second range, during all hours of the day;
- to simplify how we scale each application;
- to build everything using commodity hardware;
- and most importantly, to migrate customers to the new architecture with minimal impact.
We hit all of these objectives dead on, and we are very proud of it. We rolled out the upgrades much slower than originally planned, however that gave us time to measure system performance and make adjustments prior to bringing a large number of customers onto the new system.
And today [insert this post’s title here].
We have already begun our next two infrastructure initiatives:
- Measure Everything. This project will enable us to track historical metrics for over 170 data points within our system. We have codenamed this project "Rockefeller". Have you ever heard this quote?
- Extremely fast, virtually unlimited storage. We were one of the first business email companies to offer gigabyte mailboxes, but we know you want more. The demands placed on the inbox are growing exponentially, and email wasn’t design to be used in the manner in which people use it today. A lot of companies try to offer bigger and bigger mailboxes, but the performance of those mailboxes suffer. This is great news for companies who can solve this problem. We feel we have an awesome approach, and it is codenamed "Mercury".
"Measure everything of significance. I swear this is true. Anything that is measured and watched, improves."