A Log Structured Store for Streaming Data
Click here to download now
Rating:
(0)
(0)Overview: Event stream processing systems have focused on in-memory databases to achieve high performance, but there are several classes of application that require the persistence of some subset of the events processed and/or results produced at very high data rates. This paper describes a facility for streaming data to disk using a versioned AVL-tree index layered over an append-only log-structured store that operates at between 45-60% the throughput of an in-memory red-black tree or hash table.
