RethinkDB as a distributed open source database
Summary
After three years of development, the RethinkDB team releases a distributed database with manageable replication and an expressive query language. RethinkDB was released publicly in November 2012. The project is developed as free software.
Ideas
- Replication distributes data across several independent servers.
- Sharding splits tables across machines and disks.
- A web interface shows the cluster state and running queries.
- The query language combines functional operators into understandable data flows.
Insights
- Distributed databases need usable operating tools as well as correct algorithms.
- Replication increases availability and creates new consistency decisions.
- Open development makes architectural decisions and bugs publicly reviewable.
Facts
- The database supports replication and horizontal partitioning.
Recommendations
- Test node loss and recovery under realistic load.
- Monitor replication lag and unevenly distributed shards.
References
Links to the original source and the Web Archive open in a new tab.