People keep asking me, Hey Steve, do you think Microsoft is behind this government inquiry into IBM over antitrust issues? The answer is, of course they are. For 20 years Microsoft has been trying -- and failing -- to pry customers away from those fugly old mainframes. So now they're turning to the government to fight this battle for them.
Thing is, the government may have a case.
Don't get me wrong. I'm no fan of Microsoft. In fact I hate both of these companies, Microsoft and IBM, about the same amount. And for pretty much the same reasons. A love of mediocrity, treating customers like dirt, stamping out competition. If there were some way that the government's case could end up with both Microsoft and IBM being nuked off the planet, I'd call that a step in the right direction. Alas, probably not.
But yes, IBM is incredibly abusive to its customers. They drove everyone else out of the mainframe market, and as soon as they were the only game in town they began to squeeze their customers on pricing, not just for the hardware itself, but for software and maintenance too. And they've made sure that it is so painful to migrate off a mainframe that it's practically impossible. Basically, they've got you by the you-know-whats.
That's why all of our big banks in the U.S. are still running mainframes. They can't get off. It's ridiculous. It's as if all the major airlines were still using planes from the 1960s. Meanwhile new banks all around the world are all on newer, less expensive platforms, which also happen to be more flexible and adaptable to change. Same is true for airlines -- the old ones still rely on big expensive mainframes, while the new ones, like Jet Blue, run new-age systems and save loads of money.
Once IBM drove all the other hardware companies out of the mainframe business, they set about consolidating the software business too. Consider the case of Compuware, a big mainframe software company. IBM created a knock-off of Compuware's program and started selling it cheap, practically giving it away. Compuware sued, alleging IBM stole its code. IBM dragged the case out for years, during which time Compuware almost went out of business, because how could they sell their product when their only customers were IBM mainframe users and those guys were getting a clone of their product for nothing, or almost nothing? Finally, with Compuware on its deathbed, IBM settled the lawsuit for pennies.
Oh, these are not nice people, my droogies.
Then there are the clone upstarts -- like Platform Systems, which came up with an emulator that would run apps that were created for the IBM mainframe OS. In other words, they gave you a migration path off the mainframe. IBM sued for patent infringement (just to mess them up) and then eventually bought them -- and they were never heard from again.
Another cloner is T3, which is also in a legal hassle with IBM, and is partly funded by Microsoft. The Borg won't admit this, but they put money into T3 so they could pick this fight with IBM and draw them into an antitrust trap. Which now they've done.
The real story here is that this is about Microsoft trying to crack the glass house. They covet the billions that IBM makes with mainframes and have believed since the 1980s that they would one day take that business away from IBM.
You have to hand it to IBM -- they've done a brilliant job of defending the fortress. They have to do this, because without mainframes, IBM is done. A huge part of their services business revolves around servicing these big expensive mainframes. The mainframe is to IBM what Windows is to Microsoft -- it's the cornerstone of their business. Which is why Microsoft wants to destroy it.
So Ballmer is going nuts. He's looking at all that money and he believes it should be his. He should be running the data centers for banks and airlines. It's a fat market and Microsoft wants it. People think Ballmer is obsessed with Google. But this business, the glass house, is the real white whale.
IBM will defend itself by showing how much the cost of MIPS (millions of instructions per second, a measure of compute power) has come down over the past N years. But the cost of MIPS statistic is misleading, because while the cost comes down, IBM keeps shoving more and more MIPS down the throats of its customers.
How do they do it? Well, they roll out a new "update" of the OS, and it's super MIPS hungry. They discontinue support for the old OS, so you have to buy the new one, but the new OS forces you to buy 2x or 3x the MIPS you had before, even though you're just running the same workload. In other words, this whole "update" is just a tax from IBM.
To rub a little salt in the wound, IBM will put out a press release saying how they shipped a record number of MIPS this quarter, as if the world is all eagerly embracing the mainframe again, when really they just jammed those MIPS into customer sites.
This is what mainframe customers have been screaming about for years.
So why don't these aggrieved customers just switch if they're so unhappy? Well, that's kind of tricky when IBM either sues or buys up anyone that comes along with a migration path.
Also, you can't just up and move the whole data center. You need to do it in pieces. Guess what IBM does to you if you tell IBM you're going to start moving parts of your data center off the mainframe? What do you think happens to pricing on the stuff that's still on the IBM mainframe?
Also, thanks to the magic of the big outsourcing craze from a few years ago, in many cases IBM Global Services is running the mainframes and data centers for its customers. In a lot of cases IBM even owns those data centers.
And if you do decide to migrate, well, they'll make sure that migration is painful as hell. Like, gee whiz, wouldn't it be awful if during that migration your data center went down for 24 hours? Or 48? Maybe you ought to just stick with that good old reliable mainframe. Sure it's expensive, but it's a small price to pay for peace of mind.
This is how IBM operates -- just like Tony Soprano. And that is how the mainframe still exists in 2009.
True confession: Way down deep, I secretly admire them. We all do. Ballmer, Gates, McNealy, all of us who came after IBM -- we've studied them, and learned from them, and copied them.
They're pioneers, really. And in a twisted way, they're still the best at what they do. Not the tech. They suck at tech. But at the other stuff? No one can touch them.