Introduction to Druid, fast distributed data store

I would like to introduce a very nice time-series database product, that allows Plumbr to receive and process billions of events per day without breaking a sweat. After several disappointing evaluations of competing products Druid’s ease of use and performance characteristics made it our lucky ticket. This is the first time in many years that I am impressed with a product so much I would like to talk to everybody about it 🙂

In this talk I will present several architectural decisions made by Druid authors that allows it to receive a constant stream of real-time events, make them available for complex analytical queries almost instantly and answer those queries in a few seconds. They are very good examples of good solutions to complex problems that seem obvious in the hindsight.