Oracle has repeatedly declared its intent to invest heavily in MySQL technology in its effort to up-end Microsoft's SQL Server business. What it didn't say, but which should have been clear, given Oracle's treatment of its own database customers, is that MySQL customers were going to have to pay for those investments. Through …
could, and hopefully will, indeed get a whole bunch of new users coming from the mysql base.
However applications should soon follow up too. Some, like Wordpress e.g., only work (out of the box) with mysql.
Hopefully they'll adjust soon enough :)
I already did that switch ... back in 2006, that is. I was being increasingly annoyed by MySQL's insistance on saying that transactions and referential integrity was superfluous, and treating the *real* RDBMS engines like InnoDB and BDB as second-class engines. The BACKUP command never worked with anything other than MyISAM, and most of their improvements were made on MyISAM. It also allowed atrocities like '2006-02-31' in its date fields ... so eventually, I switched all my dev stuff to Postgres.
Everyone's up in arms now, but those in arms fail to recognize that the InnoDB and BDB engines have been under Oracle's umbrella for quite some time now. If Oracle had wanted to kill MySQL, they could've just killed those engines.
Postgres needs to think about version migration before it will get the adoption of MySQL.
I understand the reason for being cautious with opening older DBs, but for those of us that use postgres packaged by our distros, this can lead to the situation where a database is completely unavailable because the package has been up-issued. Fixing that is a headache we could all do without.
Migrate new developments.. Already doing one..
But when will I will migrate my system with 100,000 lines of code and 103 tables, which costs the companys 100+ users $0 and runs sweet.. err never!
Postgres would be a much more accepted if every fan-boy post wasnt preceeded by how much they hate MySQL
What difference between free and non free versions
That will be interesting -- if oracle release a whole load of functionality that is available in the paid for version only -- it will prob cause many to jump ship.
Mind you: for a lot of people mysql is good enough just as it is; they don't really need new features.
One angle the article didn't cover: as MySQL is open-source, what are the odds of customers deciding to ditch Oracle's support and go solo with it – or perhaps turning for independent, more reasonably-priced support providers?
Moon Macrosystems' market keeps growing!
Is't it dual license?
I'm sure that I remember that MySQL is under a dual license where the applicable license depends on what you want to do with it.
...and I have never understood how that is even possible,
The GPL side of the release specifically nixes such shenanigans.
> I'm sure that I remember that MySQL is under a dual license
It is. MySQL can be had under GPL for no money, or under a commercial licence for lots.
> where the applicable license depends on what you want to do with it.
What you want to do with it makes no difference - what matters is whether or not you want the GPL version.
If you do, your derviative work (i.e. what you build with it) must also be GPL. Some people don't want that - so they have to pay Oracle a few first-born children instead.
I still don't get it
"If you do, your derviative work (i.e. what you build with it) must also be GPL"
But only if you redistribute it. Otherwise you are quite free to keep any changes you make close to your chest*
So purchasing the commercial licence gets you what exactly? Or are you saying that if you purchase a commercial licence you can modify the code and redistribute it as a closed source product?
I don't see how that can work.
* Assuming GPLv2 of course
@Vic, re: derivative work
>> What you want to do with it makes no difference - what matters is whether or not you want the GPL version. If you do, your derviative work (i.e. what you build with it) must also be GPL. Some people don't want that - so they have to pay Oracle a few first-born children instead. <<
"Derivative work" does not mean "what you build with it".
"Derivative work" means something more like "what you build from it".
This is an absolutely massive difference for a database product.
> But only if you redistribute it. Otherwise you are quite free to keep
> any changes you make close to your chest*
That's a somewhat moot point; you only need to distribute source for any program you might distribute (N.B.: all the source, not just changes). Nevertheless, the code is still covered by GPL if you don't redistribute - it just doesn't mean you have to do anything to comply.
> So purchasing the commercial licence gets you what exactly?
The right to redistribute without complying with the GPL.
> Or are you saying that if you purchase a commercial licence you can
> modify the code and redistribute it as a closed source product?
> I don't see how that can work.
You don't need to.
But if you were a closed-source vendor that wanted database functionality in your product, you might have bought a commercial licence from MySQL. They have customers...
> This is an absolutely massive difference for a database product.
I think you're drawing a distinction that was neither stated nor implied...
The definition of "derivative work" is laid down in copyright legislation. If you think my description does not mesh with that definition, then feel free to substitute the one from CDPA88, since that was my meaning.
Nevertheless, the point I was making is that the GPL does *not* prevent you making commercial applications. It does not care what you want to do with the code - indeed, distributors are explicitly prohibited from making any prohibitions according to intent. GPL code is Free to use - even for stuff I don't want you to use it for...
What? Am I missing something? Oh noes, my company's running illegal unlicensed software! Watch out for the FAST force!
Oh, you're talking about support from Oracle? Who the hell pays Oracle for support?
Re: "MySQL price"?
> Oh, you're talking about support from Oracle?
No. MySQL is dual-licenced: you either accept the GPL, or you have to pay money to use it in a non-GPL context.
freedom loving and tight fisted..
Yep that is me in a sentence, definitely not your typical Oracle DB customer then... i'm sure i heard that knowing your customer is key to doing business....
Is it me
or does the Oracle drone's letter sound more like the language of a phishing e-mail ?
It's not just you...
I was expecting to see a reference to the Oracle Bank and an offer to "Transfer The Left Over Funds ($ 15.5 Million) Of One Of My Banks Clients Who Died In A Plane Crash".
re: Is it me
Not just that - the author seems to be trying to distance himself from the nasty megacorp people who are putting the prices up.
My thoughts exactly.
The overuse of the word "soon" to terminate sentences.
The almost childish phraseology.
Sounds like a scam to me....
Oracle are even removing the names of their employees, now... does their arrogance know no bounds?!
What about Sun's other free offerings?
OpenOffice looks like another potential cash cow. Microsoft charges >£100 for its office suite; if Oracle positioned OpenOffice as a reliable, professional product rather than as hippy freeware and charged maybe £35 for it, they could gain a lot of business customers.
Beer, because they'll be making a lot of beer-tokens.
Don't look now, but.....
Oracle is already selling "Oracle Open Office" - $50 for the Standard edition, $90 for the Enterprise edition.
Support is extra - $40 per user.
Wait and see?
Seeing as the pricing increases haven't been announced yet I'd say wait and see rather than panicing and proclaiming doom and gloom. After all, anyone with half an ounce of sense expected Larry to bump the price up on MySQL, it's just how big the bumping is which will lead to desertions or customers simply shrugging and paying the extra. If they think Oracle's development of MySQL is worth it then even a largish price increase will probably be swallowed without too much grumbling.
How much of my code is in their code base? I think I still have the right to revoke that. Anyone got a contact to send Oracle's ISP a DMCA shutdown notice?
> How much of my code is in their code base?
That depends. On how much of your code have you given them the copyright? That will set the maximum limit.
MySQL has long had a requirement on contributions for copyright to be handed to them. If they don't get those rights, the code doesn't go in. This has allowed them to continue the dual-licencing model.
This is a double-edged sword; should one of the GPL forks become dominant, it could have features not in the "official" Oracle release - and Oracle cannot just import that code into their own codebase without entirely deswtroying their ability to sell commercial licences...
> I think I still have the right to revoke that.
No, you don't.
> Anyone got a contact to send Oracle's ISP a DMCA shutdown notice?
That would be unwise.
Perhaps no reason to forget MariaDB.
"Perhaps no reason to forget MariaDB."
Had never heard of it; I'll check it out. Thanks.
Postgres - the hurdle
The one hurdle that has been there for years is replication. There are several ways of doing it, they are all ugly and not as trivial as MySQL.
"We're being told"
Who by, God?
What sort of crap "nothing to do with me, even though its the same company" bullshit is this?
Postgres? Still awaiting for proper LOB support
Postgres is a strange database, it implements fancy fields but lacks a modern LOB support. That's usually symptoms of not very well managed open source projects where drivers are not users' needs, but what developer would like to code tonight.
Re: Postgres? Still awaiting for proper LOB support
"Postgres is a strange database, it implements fancy fields but lacks a modern LOB support."
References, please. Maybe your need for large objects is actually driven by the needs of other database systems whose text support is an unsatisfactory choice between relatively narrow varchar types and inconvenient "treat it like a file" (B)LOBs.
"That's usually symptoms of not very well managed open source projects where drivers are not users' needs, but what developer would like to code tonight."
Such puerile remarks.
PostgreSQL's developers are way above average when it comes to project management - they have to guarantee backward compatibility and give more than a passing concern to data integrity - and the solution itself is both robust and capable. And many of the principal developers work for companies you have heard of, although any that used to work for Sun may have since quit.
In what way do you think Postgres lacks LOB support? It has a large object datatype (lo) and APIs for chunked/random access to individual objects including JDBC/ODBC support, and it has a byte array (bytea) datatype which which most people generally use for storing images etc. these days.
Don't forget about Ingres
Ingres is also open source (Community Edition) with some enterprise features.
Will Oracle improvements reach the free version?
I guess that's the big question - if Oracle keep their enhancements for the paid version only, then I could see a fork happening for the free version (I believe that's already happened a few times, but none of the forks seem to have gotten any traction). Maybe Red Hat should fork it - people might sit up and notice then!
BTW, one example of MySQL deficiency that Oracle really should look at first is the poor performance of MySQL Cluster with certain types of queries on ndbcluster tables (we're talking 8-10 times slower than standalone MySQL). It was so bad, we had to ditch MySQL Cluster and use a master-master replicated scenario with MyISAM/InnoDB tables instead, which virtually lost no performance compared to standalone MySQL.
It's also about time that MySQL Proxy was worked on and finally brought into the production release family - it's been stuck as a 0.X release (with some nasty flaws) for years now. It's needed if you're trying to use MySQL in a balanced DB server environment and yet little attention seems to be paid to it.
So it will happen soon then
"We're being told that there will be changes to MySQL's pricing and possibly pricing model SOON and wanted to let you know. We have not had a price increase for over 6 years but there will be an increase in the next price list that will be available SOON. We've been expecting the increase for the past couple of months but I'm told it the new price list will be released SOON."
Given that a lot of the core MySQL people (including pretty much the entire optimiser team) have left Snoracle anyway to work on MariaDB (and fix 3+ year old bugs Sun wouldn't fix because the marketing droids instead wanted various gimmicky new features), why throw money at them?
Oracle are running a business not a bloody charity! If you use their product AND want premium support you have to pay whatever they charge. Don't like it? Then tell them you don't like it and either get a better deal or jump to another cheaper product. No it isn't as easy as that and hardly anyone does it because root-canal is more fun, but that's the way the cookie crumbles.
Oracle can nasty, I've been involved in one of their licensing audits and quite frankly 3 weeks as the gimp in Pulp Fiction would have been more pleasant, but if you dance with the devil...
You are one creepy sounding character mate. Are you a lawyer?
> You are one creepy sounding character mate.
Thank you for that. It's always nice to have one's efforts rewarded with such a pleasant response. Twat.
> Are you a lawyer?
That is the wrong question. You ought instead to ask if I am *your* lawyer - because anything else would mean that you should value what you hear as exactly what you paid for it. And the answer to that question, of course, is "no".
Maybe consider Clustrix...
I'd rather avoid the effort of converting to another database like Postgres--or the massive re-architecture required by NoSQL/KVS systems. It might be time to look at MySQL-compatible options that are emerging. The Clustrix guys (clustrix.com) talk about being drop-in compatible from an application/architecture perspective, while also being to provide relational functionality and ACID guarantees across their cluster which "scales on-line from 3 to tens and hundreds of nodes". Managing the cluster as a single, highly-available database would simplify application development. Would have to see this in action though...
PostgreSQL a viable alternativr
The name is actually PostgreSQL, not Postgres. The improvements Oracle is putting into MySQL are basically bringing it up to standards PostgreSQL achieved years ago. PG is a mature, stable, and reliable DBMS and will likely see expanded use as a result of Oracle's decision.
- NASA boffin: RIDDLE of odd BULGE FOUND on MOON is SOLVED
- SOULLESS machine-intelligence ROBOT cars to hit Blighty in 2015
- BuzzGasm! Thirteen Astonishing True Facts You Never Knew About SCREWS
- Worstall on Wednesday YES, iPhones ARE getting slower with each new release of iOS
- Microsoft's Euro cloud darkens: Redmond must let feds into foreign servers