Last week saw MySQL User Conference 2008 in Santa Clara, but I was not able
to make time for it this year either. However, in the wake of Sun's acquisition
of MySQL, it was very interesting to follow what was going on. A few things
that caught my attention:
MySQL 5.1 is nearing
General Availability and an interesting storage engine plugin ecosystem
starts to emerge. It's this latter, but related event that I see as the first
real sign of validation for MySQL's long-ago chosen path of pluggable storage
systems instead of focused effort on making one good general-use engine.
Oracle/Innobase announced InnoDB Plugin
for MySQL 5.1, which much-awaited features which promise a great deal of
help for daily management headaches. More than that, InnoDB Plugin's release
under GPL lifts quite a lot of the concern I'm sure many users like us have had
about the future viability of InnoDB as MySQL storage engine.
A couple of data warehousing solutions are launched, also based on MySQL 5.1
-- Infobright
is one I've already researched somewhat (looks very interesting, as soon as a
few current limitations are lifted), Kickfire
I know nothing about right now but would love to learn more of.
There's a
huge amount of coverage graciously provided by Baron Schwartz that I have
yet to fully browse through.
A few remarks by Mårten Mickos regarding MySQL's business model seem to have
kicked up a bit of a sandstorm. I don't really understand why; I read these to
just verify that the direction MySQL took last year is to continue this year as
well. I don't see any major changes here regarding the licensing structure,
software availability, or support models. Frankly, it seems like yet another
case of Slashdot readers not reading, let alone understanding, what they're
protesting against, and press following up on the noise.
I do understand the critique made against MySQL's chosen model, though. In
fact,
I went on record last September to say that I understand that critique. I
still see the same issues here. I believe we represent a fairly common profile
of a MySQL Enterprise customer in that what we want from it is not the
bleeding-edge functionality but a stable, well-tested product that we can
expect to get help for if something does go wrong. We don't see great value in
having access to a version of software that isn't generally available to "less
advanced" or more adventurous users for free in a community version. In fact,
we see it as a negative that such functionality exists, because it hasn't
received the community testing, feedback and improvements that makes great open
source software as good as it is. While new functionality is interesting, and
we're trying to spend time getting familiar with new stuff in order to use it
in production later, it simply isn't prudent to put business-critical data in a
system that hasn't received real-world testing by as large a community as
possible (unless you have no other alternative, and then you takes your
chances).
Yet it seems to me that this is essentially what Sun/MySQL continue to
propose for the Enterprise customers by delivering "value add" functionality in
a special version of the server or plugins to it, possibly in a closed-source
form that further reduces transparency and introduces risk. Mårten, I'd prefer
it to be otherwise. How can I help you change your mind about this?