From: Shakespeare on
Op 28-5-2010 0:20, Tim X schreef:
> <knip>
> Many web hosting companies provide MySQL. Very few provide Oracle. I
> think this is mainly because a lot of web applications really only use
> databases in a very limited way - essentially a repository for key-value
> pairs. MySQL is pretty good for this type of use, while Oracle is
> overkill.
> Tim

The Oracle license model hardly allows for using an Oracle database in
the way most MySQL databases are used for websites:

1) XE (free) allows only small servers, and only one installation per
server (even on vmware, citrix or whatever) and has some serious
limitations in commercial use.
2) SE One : only 2 sockets allowed, has to be licensed for processor
license, so will start at about 5000 US $, 1000 $ Yearly support and has
to be licensed by the client, not the hosting company. So the license
can not be shared among clients, even if they would share the same database.
3) SE (15000 US $), EE (40000 US $) are just too expensive to use for
ordinary websites.

So from a licensing point of view, it could be smart for Oracle to brand
a MySQL fork as OMySQL of MyORCL to be used in the same way MySQL is
used now.

* Prices are rounded estimates of list prices.

From: Robert Klemme on
On 28.05.2010 00:25, Tim X wrote:
> galen_boyer(a) writes:
>> Mladen Gogala<gogala.mladen(a)> writes:
>>> On Wed, 26 May 2010 10:14:40 -0700, Steve Howard wrote:
>>>> It's replacement in the world? I would guess MySQL, based on Oracle
>>>> Corp.'s ownership if nothing else.
>>> Hmmm, that doesn't compute, from my point of view. That is precisely why
>>> I opened this topic. Why would a huge corporation suddenly abandon its
>>> main cash cow in favor of a freebie? Slowly killing the freebie seems
>>> more likely to me, now that Oracle Corp. owns it. I am not sure as to why
>>> people are not trying to run away from MySQL, but they obviously are not.
>> I don't think Oracle owns it. Nobody can "own" open source. They
>> bought off the main braintrust and set up agreements so those guys
>> wouldn't start up again.
>> But, the way I understand it, anybody could continue development on MySql.
> Yep, they could take a fork from it and do what they want. I suspect
> Oracle won't even try to kill it off. Mor likely, they just won't do
> anyting with it and it willl slowly wither from bitrot (or be replaced
> by a forked version). I'm not sure Oracle even really cares. It was
> something they just got with the Sun bundle. It probably wasn't even
> included as a line item in the purchase process.

If I was Oracle this is not what I'd do. If I was Oracle I'd invest as
much effort as needed to keep MySQL alive (because it has such a large
user base) - meaning: fixing bugs, adding minor features but not in a
way to make it competition for Oracle DB - plus invest in making the
migration from MySQL to Oracle as smooth as possible (at least on sales
slides). It's a fact that the volume of data to be processed is ever
increasing. With that trend many MySQL installations will hit a
scalability wall sooner or later (see Mladen's example with Cacti) and
this is exactly the point where the nice sales guy from Oracle corp. can
offer a "smooth migration" to the real thing - maybe even with a
"initial migration bonus".

Mladen, I believe people are not running from MySQL because the user
base is so large that Oracle cannot simply kill it off. Many MySQL
users will also work with Oracle DB and if they offend those on the
MySQL side this can have negative effects on the Oracle side as well.

If I personally would be looking for a replacement for an Oracle
instance I would not turn to MySQL. Instead, I would go to PostgreSQL
which - from what I am reading - has a better track record of
scalability. It also has a similar locking model as Oracle where reads
are not blocked by updates. And while the plug in architecture of MySQL
is cool and allows for some nice things to be done (MySQL basically is
then a SQL parsing and processing engine) it does not feel right for a
heavy duty database. I know, this is purely subjective.

Kind regards


remember.guy do |as, often| as.you_can - without end
From: Robert Klemme on
On 30.05.2010 12:01, Robert Klemme wrote:
> And while the plug in architecture of MySQL
> is cool and allows for some nice things to be done (MySQL basically is
> then a SQL parsing and processing engine) it does not feel right for a
> heavy duty database. I know, this is purely subjective.

If you need a more factual argument here: I believe the separation
between SQL engine and storage engine automatically limits what an
optimizer can do to improve query plans since the deails of storage are
unknown. I'm not into the details of the MySQL programming model so I
may be wrong (i.e. if the storage interface also contains functionality
to obtain information about access costs).



remember.guy do |as, often| as.you_can - without end