Scalable Mail System

How Rackspace Now Uses MapReduce and Hadoop to Query Terabytes of Data
This is a very good post about the battle of building a scalable system. This shows you how much time and effort you can spend to test the best architecture that satisfies your needs. If you read this article, if you don't care about the technical solution, if you are convinced that a solution based on relational DB would still be satisfying..No problem. Just watch the life time they took to grow their system and the huge development time they did.

Was it possible to get the logging v3+ with 2,3 weeks of work?

Anyway, the article revealed when i was working as tech and engineering on zimbra :). whenever a problem raised, i grabbed the logs stored on the local disk, and try to analyze data to check the problem..OH! those were the days my zimbra friend...