Berkeley DB 4.8
By Gregory Burd on Sep 14, 2009
Today we are very happy to announce the availability of an amazing new version of Berkeley DB. This release is full of new features, performance improvements, and much more.
New in Berkeley DB 4.8:
- Performance has been enhanced for multi-threaded/process applications on CMP/SMP systems.
- New C# and .NET API providing more support for Microsoft programmers.
- New support for C++ Standard Template Library (STL) allows developers to use transactions, cache large datasets, and persist data using familiar STL data structures and APIs.
- New API to support referential integrity constraints including abort, cascade, nullify semantics.
- New B-tree compression APIs allowing you to reduce the amount of disk space required to store your data with your custom compression code (or our default implementation). Using compression will reduce disk I/O and allow more data to live in-cache. This space savings in turn improves performance (less I/O and more in-cache data) despite the overhead of compressing/decompressing data. Random reads is one case where performance could suffer when compressing data, be sure to understand your data access patterns and test your performance.
- New support for partitioning database files can improve application throughput and reduce lock contention when data is split across multiple I/O channels and/or disks.
- New bulk load/delete APIs can significantly improve application performance.
- New command line tool for bridging the gap between SQL and Berkeley DB databases called 'db_sql'. This utility will translate the majority of SQL92 DDL into Berkeley DB source code.