Can you imagine a database that stands as much traffic as you want, adding and removing nodes automatically, working interruption-free even if a volcano starts spitting lava over your data centres? That's the promise of distributed systems like Cassandra, Voldemort or Riak. But, how do they do it? In this talk I will tell you the basics of distributed databases. Come to my talk if you want to learn about CAP, BASE, incremental scalability, replication or consensus. I will also explain geeky things like gossip, vector clocks, hinted handoff, or entropy. By the end of the session you should understand which is the architecture that allows Amazon to have their shopping cart available 24 hours a day, 365.25 days a year, or Linkedin to be able to spam you non-stop. When was the last time you saw Amazon or Linkedin websites down?