Blur (Incubating) Wiki
Migrating from Github, currently incomplete.
Blur is an open source search platform capable of querying massive amounts of data at incredible speeds. Blur is built on top of Lucene, Hadoop, Thrift, and ZooKeeper. Tables consist of a series of shards (Lucene indexes) that are distributed across a cluster of commodity servers. All index information is stored in HDFS, cluster coordination is handled by Apache ZooKeeper, Apache Thrift is used for the RPC, and Lucene handles all of the indexing and data storage.