Polymorph: Zak Greant's Blog

Oracle buys Sleepycat

April 1st is still more than a month away and at least one rumour about Oracle's upcoming purchases is true: today the software giant annnounced their acquisition of Sleepycat Software, the makers of Berkeley DB (and various other products).

One interesting point is that Berkeley DB was already seeing competition from SQLite (which is an excellent, fast and free (as in beer and freedom) RDBMS). I wonder how much the acquisition is going to drive adoption of SQLite?

Additionally, Oracle now owns both half of MySQL's transactional storage engines, which perhaps gains them another measure of control over the Swedish upstart. (The other engines are NDB and Gemini)

It looks like MySQL is digging in though - they have just procured another 18.5 million dollars of investment (primarily from larger companies with a vested interest in MySQL's success). I bet that the money goes towards building a transactional database engine that they hold complete copyright on.

See also:

Tags: , , , , , , ,

Related posts


Posted on Tuesday, February 14th, 2006 at 18:01

You can follow any responses to this entry through the RSS 2.0 feed.

You can leave a response, or trackback from your own site.

7 Responses to “Oracle buys Sleepycat”

  1. Anonymous (5 comments) Says:
    February 14th, 2006 at 19:07

    NDB :)

    The clustered database has transactions, and BDB is virtually unused.

  2. Zak (71 comments) Says:
    February 14th, 2006 at 19:46

    D'oh! Very true. I bet that we see rapid progress turning NDB into a disk-based storage engine, as (last time I looked) it was still a memory-only engine.

  3. Anonymous (5 comments) Says:
    February 14th, 2006 at 23:36

    Currently, in the latest 5.1 (I think), unindexed data was saved on disk in NDB clusters; only indexed data was "RAM only". And even that will change soon.

    So, yeah…I think MySQL Cluster could be a good replacement for Oracle-haters. :-P

  4. Lukas (5 comments) Says:
    February 14th, 2006 at 23:39

    Doesn't 5.1 reduce the memory requirement to only cover the indexes already?

  5. jim (1 comments) Says:
    February 15th, 2006 at 0:57

    disk-based data for cluster is already in 5.1. (indexes still need to find into memory, though.)

    brian keeps threatening to add transactions to the archive storage engine. i've thought it would be fun to make the memory storage engine transactional, although obviously you'd be missing that d in acid.

  6. Mark (1 comments) Says:
    February 15th, 2006 at 2:13

    Zak, see http://www.flamingspork.com/blog/2006/02/15/information_schemafiles-querying-disk-usage-from-sql/

    (disk based cluster is pretty much already there in 5.1.6, just hammering out the remaining issues)

    The feature has been "trickling" out during the 5.1.x alpha/beta releases ;)

  7. Anonymous (5 comments) Says:
    February 15th, 2006 at 2:52

    See the release notes for 5.1. Disk storage is already there. :)

    James Day

Leave a Reply

Polymorph is powered by Wordpress running on Apache, Ubuntu Linux, MySQL and PHP.

The marvelous illustration of the Mad Hatter is by the late, great John Tenniel.
Like many great parts of our culture, it is in the public domain.

Contact: zak@greant.com | Gnu Privacy Guard Key

Entries (RSS) and Comments (RSS)