Part One | Part Two
SearchWinIT.com: You think that only the largest companies will benefit from cloud computing. Why is that?
SearchWinIT.com: Still, many people are pushing cloud computing as the future.
Minasi: I've been in the computer industry for more than 35 years and there is a cycle. We do outsourcing … we go in-house, we go out-house, we go in-house, we go out-house. Maybe it's micro-computers in the '80s. Maybe it's networks in the '90s. Then maybe client/server. Then, after a while, we say it's not so new or high-tech anymore.
Now that something new is cloud computing and we're hearing it can save us money. [Service providers are saying] 'I've got a building over here where I've got tons of computers, networks and switches and routers … and I'll sell it to you for this rate.' That's where the outsourcing starts.
Where the concept goes wrong is when something goes wrong with the network. When that happens in-house, the boss isn't happy with the IT guy and he doesn't have a job. But with [cloud computing], you get an interface guy that manages you and 300 other businesses. The worst thing you say to him is, 'We're pulling out of your company,' and he thinks, 'Gosh, then we'll only have 20,199 clients.' His bonus may be smaller, but he won't lose his job.
IT isn't like the motor pool. You can't just outsource to a cab company. Imagine running a company without IT, particularly with how much the world has changed in the last 30 years.
The two people who hold the most dangerous positions in a company -- and the ones we have to trust the most -- are the accountants who hold the money and the IT people who hold the data. Without the money or without the data we're up a certain well-known creek without a paddle pretty quickly. That's why I don't think outsourcing is going to work. I remember when time-sharing was a big deal in the late 1960s and early '70s. Years later, along came ASPs … I giggled out loud when I heard about ASPs. That, of course, didn't last.
SearchWinIT.com: Cloud services are still being developed really. How can you be so sure they won't take off?
Minasi: Even if someone builds the Cadillac of cloud services, they'll be out of business within a year because American businesses, publicly-held corporations, are required by law to drive everything to the lowest possible cost. Everything will fall to the same level of mediocrity.
The other reason is … nobody is going to make money running 20 processors and a small SAN. For people to make the most money, they'll have to have warehouses and warehouses of this stuff. Where have we seen this before? Online auction houses. If I'm going to sell the most, I'm going to put it with the one that gets the most eyeballs and there's only one online auction company. The same thing is going to happen with cloud computing: The ones that are successful small boutique shops will get gobbled up by huge ones.
Also, this will be a drive to put the most inexpensive people possible on the help desk, and look what happens with that.
If you're a big company, great. But if you're a big company, you can afford to set up your own cloud.
SearchWInIT.com: So you don't see the cloud as an option for SMBs?
Minasi: The only model that I can think of where there's one person supporting multiple companies and offering great service is [Microsoft] Small Business Server. I have a client who supports about 35 customers. He does it remotely with SBS. He gets a check from each [customer] every month and because it's a small number, they all get the support they need. If there was a way to do that with cloud computing, a way where everybody supports a small number of clients, then that would be optimal. But they're not going to do that. With cloud computing, simple economics says the bigger the better.
SearchWinIT.com: Do you think virtualization will make things easier for IT?
Minasi: It already does. We would not be able to have the level of support we have today if it weren't for technology like that. In 1989, you've got this model of racks and racks of computers doing one thing, running at 5% capacity sucking up power and air conditioning. Server consolidation, if nothing else, makes us love virtualization. Virtualization is the single coolest thing that's happened in the last ten years.
SearchWinIT.com: Will desktop virtualization have the same kind of impact on IT?
Minasi: It's mixed. When I first saw Terminal Services 12 years ago, I thought … the best of the PC and the mainframe. I advised my friend to go into Terminal Services. I said, 'This is going to be big.' I was wrong. [With Terminal Services] the licensing is so byzantine and I think really unfair. Guess what Microsoft is doing with Kidaro? Just as stupid a licensing system with MED-V [Microsoft Enterprise Desktop Virtualization]. So why aren't people using Terminal Services? To read the licensing you need a degree from Harvard Law.
The second problem is if you're going to deliver desktops, what percentage of your day are you really driving your computer hard? Often enough that I have to put a sufficient amount of juice in the back room to deliver your applications. Where are we going to put all those blades? Now that I'm supporting 2,000 users, does that mean I need 2,000 PC blades? Where do they go? Who cools them? Where do we get the power for them?
Have we really saved any money at this point?
SearchWinIT.com: What about a hypervisor on the client approach instead of a VDI approach?
Minasi: It's an application compatibility answer. It's not a desktop support answer. I'm not dismissing it. Application compatibility is a big thing. If I could wave a magic wand, the problem I would solve for the world is to make all the bad apps go away and have them come back as correctly written 32-bit and 64-bit apps.