MySQL and the freedom to fork

Patrick Galbraith has initiated a fascinating discussion with his post, "What is the official branch of MySQL?" I did a double-take when I first saw it, and I can't quite shake the question from my mind. It implies much of the power, and peril, of open source.

The question is critical because it implies that open source can become much bigger than the developer--whether an individual or a company--that created it. While Linus Torvalds, for example, remains central to Linux kernel development, Linux has become much, much bigger than Torvalds. Companies and foundations have been set up to guide and monetize it. Billions of dollars are earned and lost each year because of Linux.

In the case of MySQL, it has sprouted forks and iterations/distributions on the popular open-source database. OurDelta (a superset of MySQL started by ex-MySQL employee Arjen Lentz), Drizzle (belatedly recognized as an "official" Sun/MySQL fork), and MariaDB (created by MySQL co-founder Monty Widenius) are just a few of MySQL's off-shoots, but a few is enough to prompt Galbraith's legitimate question, particularly if you're an enterprise looking to buy into the "true" MySQL code branch.

For Sun, the forks arguably both enhance and diminish the company's ability to recoup its $1 billion investment in MySQL. Forks siphon off development that could be focused on the main code branch, and could also redirect dollars to these branches.

On the positive side, however, the greater the proliferation of MySQL forks in the marketplace, the more salient and powerful the MySQL brand becomes, and hence the better able to command support subscription revenue. Red Hat Enterprise Linux, for example, becomes increasingly valuable as Linux variants multiply: RHEL becomes the safe, grounded choice for ISVs and enterprises.

Identifying the "official" branch of MySQL depends largely upon what you want. If you're an enterprise looking for the safe, standard build, Sun/MySQL is what you want. But if you're looking to build a Web-enabled business, Drizzle may be the right choice. Or if you're on the cutting edge and feel that Sun's support is too slow, OurDelta may offer the best sanctuary.

In short, figure out what you really want from MySQL before deciding what "official" means.

Follow me on Twitter at mjasay.

[pendragn]

For most enterprises, the choice of "what software to use" is based on whether or not the software is supported (enterprise level) by the developing company/dev house. Many forks of many different software branches are out there, but some are clearly more suited for enterprise deployments than others, for the mere fact that support and consulting and patching services are there for businesses to rely on.

I think a business would not think of running software that they could not get adequate levels of support for, plus the guarantee that enterprise long-term needs are being thought about and taken seriously (patching schedules, end-of-life concerns).

This is an example of why I would rather pay for a commercial database rather than using an open source alternative. There is no roadmap. There is no official standard. If there is a bug, since I am NOT going to fix and I am NOT going to apply a patch applied by a random company, I'd rather pay and use a real product. In cases like these, support from a company like Sun who doesn't control its own software is throwing good money after bad.

[another_cissp]

I guess it depends on the support you require. Open source is well supported in the community and answers can normally be found in minutes rather than calling up some tech and spending hours going from level 1 to level 3 support.

[Dalkorian]

He does have a point though. If you're unable or at least not interested in understanding the product in question, paying someone else to get it for you might be the best answer. If you're willing to get your hands dirty and learn something, open source will be more prominent to you.

As snooty as that remark sounds, it's true. Case in point, the owner of a retail business might not have the background, time, desire or inclination to learn an open source product. They just want to call someone when it doesn't work as expected and be hand held through the process so they can get what they are trying to do done.

Notice we're missing the market where you pay someone to support open source products for you.

[odubtaig]

No, the comment with no name is both ignorant and ill-informed. It's the sort of comment where I usually suspect wilful ignorance of the kind that doesn't want to hear anything contradictory even if God came down and said it.

Sun not only controls the core MySQL branch as available from them in its entirety, it also controls the development direction, what contributions are included and has a sizeable internal development team. It is written to a widely documented and unchanging (added to, not modified) standard set by MySQL AB, then Sun, which any 3rd party developer can rely upon.

Any halfwit with ability to type 'mysql roadmap' into a search engine (Google, Live or Yahoo, pick one) will find http://dev.mysql.com/doc/refman/5.1/en/roadmap.html right at the top.

Yes, that's the official, publically available roadmap for MySQL 6.0 in the freely accessible manual.

Paying customers of Sun are paying maintenance and therefore when they submit a bug report, it will be fixed by Sun if it hasn't already been fixed by the community.

Need I go on?

[Sourdust]

@odubtaig, you could have gotten your point across without insulting anyone. Posts like yours hurt open source by making it seem like there is no friendly help out there. Who's going to want to seek out the community for help if they think they'll be talked down to and mocked? They'll assume the reply will start with "any halfwit with the ability to type..."

[odubtaig]

You're right. My replies are sometimes coloured by who it is I'm replying to and this one has a history of being deliberately ignorant to fit an agenda.

If I think someone honestly doesn't know then I'm much more polite, but I guess anyone reading casually doesn't know that.

[mylinuxsupport2009]

Open source IS the community and not the software brand. If the community forks the code and the core development effort shifts from MySQL to" OurSQL" the world continues and SUN is out a billion dollars. This is the true power of open source and is the most important reason to abandon proprietary software and embrace open source. The community not any individual or corporation is really in charge. As to support, with open access to source code 3rd parties can provide world class software support including bug patches and customization. In the future owning "software" brands will be seen as a liability and not an asset.

MyLinuxSupport.com

[rapier1]

Nothing you say here does a lot to reassure the enterprise consumer or developer. No company wants to invest several million dollars in an enterprise app that depends on people thinking "its cool" for it's support structure.

[odubtaig]

'Cool' has nothing to do with it, usefulness and whether the community on which these companies depend is treated with respect is everything. There have been few major disruptions but they have all come down to that.

The Community may not have invested a lot money in a package but they will have invested time and effort (after all, isn't that what your company's money spent really represents?) which still makes it difficult to move away so if they're moving away to some other package, you have to start asking why that is.

As it is, say it comes to your next support renewal time and you don't like the way things are, you're looking at an alternative package which offers better terms and a development roadmap which suits you better while still maintaining compatibility with what you're using now. Could this be done with Oracle? MS SQL?

I'd say the option of being able to move to another supplier almost pain-free should you wish is a good one. You don't get that with proprietary packages. Buy from a proprietary vendor and you could be subject to lock-in with a treadmill upgrade cycle, they could go bust meaning no updates or bugfixes ever and once they have you locked in there's no compelling reason for them to make their software as good as it could be because they no longer have to compete for your money.

Paying for support from an open source vendor, however, means that if they go bust you can pay someone else to fix problems, your data can be directly accessed without having to pay the toll of extra licenses or paying extra just to run their software on a newer system and because other companies can fork, alter, tailor and support the same software they are _always_ competing for your money so they have to work hard for it.

[Nitin_sawant89]

Open source will bring boom in commercial software industry

[mylinuxsupport2009]

Rapier1,

It's not a bad thing when the industry, developers, and consumers pay attention to vision, talent and contribution. Open source is about transparency and in an open source environment vision, talent and contribution drive things not brands and patent portfolios. When source code is published, all you have to do to find who did what is review the list of contributers and look for the sections of source code their name shows up in.

All parties have equal access to the source code necessary to support products so support organizations have to compete on their talent and business practices not on access to trade secrets and proprietary information.

- mylinuxsupport.com