Like many web outfits, Twitter once relied heavily on Cassandra. Originally developed by Facebook, Cassandra is one of many “NoSQL” databases designed to store data across hundreds or even thousands of machines. This worked well enough, but Twitter soon found it was too difficult to expand the system to new sets of machines. “A lot of systems have this problem,” says Avital, who helped oversee the use of Cassandra at the company. “It’s hard to manage systems when server clusters grow from tens to hundreds to thousands of nodes.” And, perhaps more importantly, there are certain tasks Cassandra just wasn’t suited to, tasks that requires other database tools such as MySQL and something called Gizzard.