OpenAcademy at Columbia University
Redis provides a base feature set from which a graph database engine can be built efficiently. Redis is a data-structure server, providing blazingly fast read/write speeds. It has mature support for asynchronous replication that forms the backbone of highly available and reliable systems. Benchmarks and some initial tests demonstrate that Redis is a good choice for such a graph search and analysis engine.
Chae Jubb (@cjubb39) and Sireesh Gururaja (@gsireesh) considered the existing Apache Solr. On top of this Apache Solr search engine, we added features for a graph database. That graph database is backed by Redis, a data-structure server. By integrating with Solr, we were able to add graph-search functionality to an already-existing search engine by simply adding a few search query modifiers. With this new search feature, we were able to search a traditional databse alongside a new graph database. Redis is used as the underlying storage of this database. It is mature enough that it can be used as the backbone of an available and reliable system. Ultimately, the graph is stored in this high-speed, graph database.
You can also check out the Columbia Redis team's final presentation here.
Go back