Why MySQL might not benefit from having a mothership

The bit that caught my eye, shortly into the post itself, was:

The conversation went something like “I was talking to so-and-so, and he said, you know, you guys really need Sun/MySQL, because without the mother ship, things will fall apart and your own business will fail.”

While at the User Conference I had a conversation with a colleague o’ his on this very subject, causing me to wonder if the aforementioned “he” is in fact “me”.  Assuming so, that is not at all what I said, so I’ll try again.

My employer has nothing to do with anything I write here, etc.

Everyone Needs the MySQL Mothership

What I actually said was: I believe that a successful commercial enterprise called “MySQL” is necessary in order to create the types of opportunities that exist in the “MySQL Ecosystem” today.

I’m not biased because I’m responsible for running MySQL Support.  I’m “biased” because I was the DBA who made the business decisions to deploy products against MySQL for three different companies, placing not only our businesses but also our customers in the hands of MySQL.

Some of the reasons I made those decisions are no longer true of MySQL (currently, at any rate), but one of the primary reasons I selected MySQL over, say, PostgreSQL was because “MySQL” owned its product.

It was a commercial company producing an open source product.  It was not an open source project.  It was not a company with a South Park-esque business plan (”1. Monetize other people’s work  2. Who cares?   3. Profit”).

The database is not a web server, performing generally simple tasks.  The database is not the kernel, vital but invisible.  The database is not a commodity.

The database is my business and my customers’ business!

I can’t trust that to an open source project, because I need a commercial relationship.  I need accountability, responsibility, reliability.  I expect to be here tomorrow, and I need you to be there with me.

I can’t trust that to an IRC channel, a forum and a mailing list, populated by people who may not know what they are talking about, who may not care about my problems and who just may be elitist jerks preferring to insult me than to help me.

I can’t trust third party generalists.  No matter how much Linux distribution or “open source stack” vendors know about supporting their “thing”, they cannot properly support my database (”my business and my customers’ business”) because they simply cannot know it well enough.

I can’t completely trust third party specialists, however trustworthy they may be, even though I may choose to use their services.  It’s the “Tomorrow” question all over again, because specialists (by definition) don’t do everything.

If MySQL, the commercial company producing an open source product, had not existed then every one of those companies would have been forced to continue deploying against MSSQL and Oracle.

There are certainly businesses without such concerns, or who are unable or unwilling to make those particular business decisions (the .com bubble certainly did not teach anyone anything), but the particular needs I had are very much prevalent among companies actually willing to enter into business relationships with (”pay money to”) a particular vendor.

It’s rather like a developing star system (I’ll refrain from naming the star “Sun”).  If the star has sufficient mass, it will draw more things to it, perhaps at an accelerating pace.

Yes, most things will be drawn into orbit around the star itself.  We may even find that this power it wields is confining or restricting by its very nature.  But, if the star lacks sufficient mass to hold the system together, the entire thing will eventually fly apart.

In such a case the best result may be that some of the pieces form smaller, local systems and drift off to wherever, but those smaller systems will never possess the same “drawing power” as the larger system.

Or: Opportunities will always exist, but the number and nature of those opportunities depend heavily upon the “gravity of the star” organizing the system.

We all need a successful commercial venture named “MySQL” at the heart of the “MySQL solar system” in order to continue enjoying the same types of opportunities, to continue moving into new regions with ever-increasing momentum.

Some folks are happy, even happier, within the context of a “small, local system”.  I just hope that they remember to look beyond themselves and recognize that even if their own lives would be improved by destroying the star, there are millions more that would not.

Perhaps I ask too much.

I doubt that planetmysql.org includes me in our feed anymore, and generally that works out fine for all concerned.  My rantings while quitting smoking were probably at least “a minor contributing factor” in the filtering we now employ.

It’s not for lack of things to write (trust me), so, why the reticence?

I suppose mostly it’s that I don’t like to read blog entries about MySQL.  Every time I dive into Ye Olde Blogosphere I see us revising the 80/20 rule into more of a 90/10 rule, or maybe 95/5, or 97/3.  Or: 3% of everything is Scottish; the remaining 97% is crap.

But the real reason is drama and my fondness for avoiding same.

Being online lost its newness and novelty for me some 26 years ago, so perhaps that causes me to be less inclined to take my 15 nanoseconds of soapboxery seriously.  I am certainly less inclined to take that belonging to others very seriously.

MySQL was blasted for being a “toy database” because it lacked features.  It continued to be blasted about some of them long after their lack was rectified.  All along, an ever-increasing number of users and customers happily went about their lives and their businesses with it.

Now it’s en vogue to bash MySQL for making business decisions, or for producing software that contains bugs, or whatever.  And again: all along, an ever-increasing number of users and customers happily go about their lives and their businesses with it.

It’s so easy to lose any sense of perpective.

What’s the goal?  What’s the mission?  What’s the point?

I don’t define the answers to those questions, except perhaps for the Support team, but from my perspective the MySQL “mission” is about enabling data management:  Everyone has data; everyone needs to manage their data.  Data management has become a utility even more vital than electricity, and we’re there helping to make it happen all over the globe.

It matters much more to me that we still try to accomplish our mission than that the road sometimes gets a little rough.  Nothing is perfect, therefore imperfection does not surprise me.  Everyone makes mistakes, therefore blunders do not offend me.  Some things must be seen through to the end to assess their outcome, therefore experimentation does not worry me.  All questions are worth asking.

I care too much, I’m told, but really I only care about the one thing: the mission.

It helps keep everything else in perspective.

Virtual Box 2.0 dropped today.  Now I have my 64 bit guests AND apparently the lockup under OS X was a VBox defect of some sort or other, because it does not occur under 2.0.

Now if only the Windows “AMD64″ installer (ie: for the 64 bit host) would execute I might actually be able to use this thing for something serious.  Either it dislikes Intel CPUs or dislikes 32 bit Vista; VMWare doesn’t care, and neither do I, so VBox loses again.

I finally managed to get VBox functional on OS X.  Installed Ubuntu and discovered that I can lock up the VM at will by “ls -lR /”.  This only happens under OS X and actually reinforces my suspicion that my Macbook Pro has a hardware problem.  But maybe VBox is buggy on that platform.  I switched to Windows as the host OS for reliability (…)

I decided to try OpenSolaris (for the first time) under VBox.  I have a long-standing hatred of Solaris’ installation procedure and “user experience”, and have to say that Sun/We/Gaia are going very much in the right direction with this.  Gnu userland utilities would probably cause me to switch to OpenSolaris on my *desktop*, never mind the server.

Ubuntu “server” seemed to have mostly similar performance between VMware Server and VirtualBox.  Ubuntu “desktop” was dramatically better under VirtualBox (and, to continue gushing, Gnome under OpenSolaris/VBox was also dramatically better than Unbuntu/Gnome/VMware).  For whatever reason, Ubuntu/Gnome/VMware is too slow for usability.

So, all in all a mostly positive experience.  I need to investigate whether my Macbook is fubar (very likely), and Sun/We/Gaia need to enable 64 bit guests.  As I’m doing rather important things in my VMs, I confess that I will continue to trust VMware more for at least the near future.

They definitely have some competition, however.  And that’s good for virtualization.

1) Some of my newer computers are dramatically more powerful and energy efficient (all around) than my older systems.  It’s simpler/cheaper/saner to virtualize many/most/all of the older machines and run them on a single physical server.  Consolidating their respective functions onto a single server would take weeks or more of my time, so this was the way to go.

2) Being able to “pack up” an entire server and “take it with me” to different servers (or laptops) makes my life and my work a lot easier.

So I decided to try Sun’s xVM VirtualBox today.  It’s leaner/meaner/maybefaster and “Open Source”, and it supports a better set of host operating systems than VMWare Server itself, so I was actually looking forward to it.

Right up until 30 seconds into it when I encountered an apparent limitation, in the form of not supporting 64 bit guests.  I found this on a forum post, wherein We Freeloading Community are informed that the team is demand driven and focused on the needs of paying customers and that this isn’t one of them.

Sad to think that the paying customers don’t need 64 bit guests.  It made me wonder what they’re virtualizing, really.

Old and/or “small, dedicated, low-load” systems, a’la my #1 above?

Presumably they aren’t pre-provisioning servers for scale-out, because the first thing they’d want then are 64 bit systems, and their Demand would Drive support for this seemingly wild and exotic feature.

Oh well.  Back to VMWare.  A real shame, too, as I was looking forward to moving some of my VMs to the Macbook Pro.

I guess I need to buy VMWare’s Mac product, now.  Their paying customers apparently need the same things I do.

I like some things (a lot); I dislike some other things (a lot).  We’ll see how it goes.

I’m not sure how I will adapt to a large corporation, nor how a large corporation will adapt to me (not at all, I suspect).  I think I have too much entrepreneur in me to fall in love with them as an employer, but who knows.  They seem to be a generous corporation, at least compared to other large USA-based corporations, and regardless of motivation they are exploring some interesting avenues.

Change Is Stability, I’ve heard, so here’s to change.

So if I owe you a beer, or more importantly if you owe me a beer, you know where I’ll be.

Of course and as always, if you are interested in joining the MySQL Support Team you can drop me an email or catch me on Freenode or trip me as I walk by and we can meet up at the conference and discuss the subject!

Well, I don’t know! Those decisions are not mine to make.

I could speculate. (Frankly, I would probably be spot on, too.) But I’d rather not, and that’s not why I’ve resurrected this blog.

Many of the questions amounted to, “Should I stop using MySQL?” or, worse, “Will I have to stop using MySQL?” I decided to answer those here.

The questions were mostly from people who do not know much about Sun and who are afraid that the acquisition will result in mysqld becoming solely closed-source, proprietary, licensed software; afraid that they would have to pay if they do not today, or have to pay more if they do.

I cannot commit Sun any more than I can commit MySQL, but let’s face it: Sun really is the biggest “free software” and “open source” company on the planet. They have released nearly all of their products under the GPL, and I suspect the only holdouts are because of 3rd party patent and licensing complications.

Should you stop using MySQL? Well, that’s up to you, but I think it’s safe to say that such a decision has nothing to do with Sun. In fact, you probably should feel more confident using MySQL because of the acquisition. (If Monty likes it, so should you.)

Will Sun force you to stop using MySQL? I can’t speak for them, but why would they? They presumably want to extend MySQL’s ubiquity. If you aren’t paying for MySQL services or licenses today, they’re far more likely to try creating value that you will want to pay for than resorting to outmoded strong-arm tactics. It’s just not “in character” for where they are, today.

They explain themselves better than I can, so let me just wrap up with one small bit of public speculation: for the many millions of people and companies using MySQL today, I think the unlikely, worst thing that could happen is that nothing directly affecting you will change at all.

That’s more like a Christmas present than a worst-case scenario.

Sunta Claus? Heh. Merry Christmas.