hamsterdb Embedded Storage is an embedded database engine written in ANSI-C. It includes B+Trees with variable length keys and records. It supports in-memory databases and endian-independent files, database cursors, multiple databases in one file, "record number" databases, and duplicate keys. hamsterdb is very fast and highly configurable. It compiles and runs on Unix platforms, Linux, Microsoft Windows, and Windows CE.
hamsterdb Transactional Storage is an ANSI-C implementation of a key/value database engine. It has a lock-free architecture, thus supporting an arbitrary number of long-running transactions and resolves transaction conflicts in memory without accessing the disk. hamsterdb uses logical, idempotent logging for recovery. It is thread-safe and concurrent, moving performance intensive operations to the background. Every hamsterdb handle can be used from arbitrary threads.
This project has been discontinued. The transactional functionality has been merged into the main hamsterdb branch (freecode.com/projects/hamsterdb), and the concurrency/multithreading will be merg...
hamsterdb hamsterdb's first submission was in September 2006. It's a DBM-like library written in ANSI-C concentrating on high performance. It can run as in-memory database, use memory-mappe...