The first release of Oracle Berkeley DB with support for SQL came out roughly six months ago and was on par or better in terms of performance with SQLite while keeping 100% drop-in API compatible with the SQLite3 APIs. We tested it in Firefox, Chrome and other popular and demanding SQLite-based open source applications to make sure it worked out of the box and it did! The advantages of BDB's storage engine were clear to everyone who tried it for themselves; BDB is faster, more scalable, more concurrent and offers greater operational efficiency for the same workload. This release delivers even more functionality, more speed, and much more BDB goodness to users of the BDB SQL storage API.
The Berkeley DB SQL API provides a very easy to use, well understood interface for Berkeley DB users -- SQL, JDBC, ODBC and ADO.NET. Building applications with Berkeley DB just got a whole lot simpler.
With the addition of the SQL API, Berkeley DB clearly has pulled into the lead in the embedded/mobile/edge SQL database race. Our plan is simple:
- Ubiquity by supporting SQL, JDBC, ODBC and ADO.NET,
- Flexibility by allowing the application developer to chose the API that works best for them,
- Provide best-in-class storage engine features, performance, concurrency and scalability,
- Stay fully compatible with the SQLite API,
With this release, we continue to widen the gap between Berkeley DB and the other embedded storage solutions that are available. We are committed to sustaining a competitive advantage for customers who choose Berkeley DB's SQL storage APIs. Our value is in our ability to put your products ahead of your competition without adding complexity or forcing unnecessary compromises. We want you to have an obvious advantage over your competition simply by choosing Berkeley DB. The fact is that Berkeley DB users now have two streams of technical innovation working for them.
This release:
- delivers a complete, embeddable SQL92 database
- as a library under 1MB size
- drop-in API compatible with SQLite version 3
- no-oversight, zero-touch database administration
- >15 year battle tested Berkeley DB BTREE for concurrent transactional data storage
- support for encrypted SQL database files
- support for Apple's iOS platform on iPhone/iPad/iPod mobile devices
- real binary file management with access using SQL (BFILEs)
- online database file space reclamation (VACUUM) that can be used while the database is operational without stopping or disrupting active transactions
- faster on many common operations than SQLite when single threaded, much faster under concurrent load
- we've reduced database file size, a more optimized index layout
- added an option to split databases into multiple files, one per-table, rather than all in one file
- we've added an option to share cache, locks, and transaction logs across multiple separate databases to reduce overhead when dealing with dozens (or more) databases in use at once
- easier performance monitoring/measurement using DTrace or Systemtap probes on Solaris
- we've added a new command line tool called "db_replicate" which turns any SQL or non-SQL database environment into a replicated HA cluster in one simple step, no code required (use it to test HA, keep a hot/live backup, or even in production!)
- and a lot more!
Berkeley DB is hands-down the best edge, mobile, and embedded database available to developers. We're moving more and more of the core BDB features into the SQL API with every release. We hope you're excited because we certainly are!
No comments:
Post a Comment