Last week we told you about how Twitter is migrating its search stack from Ruby to Java. But Twitter is also known for being an early adopter of Scala. This presentation by Marius Eriksen at the Commercial Users of Funtional Programming 2010 conference explains how Twitter uses Scala to scale.
Scaling Scala at Twitter by Marius Eriksen (Twitter) from Scott Smith on Vimeo.
A few of the things Twitter uses Scala for:
- Social adjacency store (FlockDB)
- Namesearch
- “Who to Follow”
- Kestrel/queueing
- Streaming API
- Storage systems
- Geo
Eriksen mostly focuses on the geospatial applications powered by Scala at Twitter. Eriksen and another engineer built Rockdev, Twitter’s geospatial backend, without any previous experience in Scala or the Java Virtual Machine.
Eriksen also talks about the role distributed databases, Apache Cassandra in particular, fit into the team’s scalability projects.
See also: How Twitter Uses NoSQL.
