Open Clouds and Google’s Future

The idea that clouds should or could be closed is a straw man.

Even Cisco agrees they should be open. So Red Hat CEO Jim Whitehurst repeating this truth is no big whoop. 

But lock-in isn't all about software. Just because Cloud A and Cloud B run open source software does not eliminate the lock-in problem. They could be running different open source systems.  They could be incompatible in some other way.

More important is the fact that clouds are big, and systems built in clouds are very big. You ever try to move between two Windows machines? It can take up a full day. Now you're talking about moving between two mainframe equivalents. It's going to take time, it's going to take money, and it's going to be a considered decision.

So HP executives gave away nothing in the Cloud Memogate that has tongues wagging in the Valley.

The hope of all the computing giants (HP, Microsoft, IBM), Web hosts (Rackspace, etc.) and telco companies (AT&T, CenturyLink) crowding into the cloud computing space is that their reputations, their brand names, and their current customer lock-in, will be enough to see them through into the new era.

They couldn't be more wrong.

In the long run, I think, the cheapest infrastructure wins. So cloud computing plays into Google's wheelhouse. That company is tip-toeing into the space for the same reason it doesn't provide ISP or wireless phone service. Which is that attracting government interest is bad for the business model.

What killed Microsoft, what killed IBM before it, and what continues to kill the telcos, isn't what government does but the mere threat of government action, the necessity of responding to such a threat, of even getting involved with government in any way. Government action can take all sorts of forms – patent suits by competitors, private antitrust complaints, class action suits by customers – and they all have the same result. They create corporate arteriosclerosis.

The more government action is real or imagined, the more lawyers you hire, the more bureaucrats you put between innovators and delivery of product, the slower you can move.

And the most pernicious form of government interference with business is (surprise) the contract.

In the end this will prove the greatest and most-lasting damage of the Bush Administration, I feel. They created a vast ring of government contractors around Washington, companies devoted exclusively to getting big-money deals to run government functions. Contracting involves taking on a lot of costs you don't charge for up-front, before you get the contract, which you have to get out (plus a profit) after the contract is signed. Funny how conservatives who sought to privatize government in order to make it “more efficient” never figured this out. They thought they were making government as lean-and-mean as corporate America. They were, in fact, making corporate America as fat and bloated as they accused government of being.

So expect Google to tiptoe slowly into clouds and other new markets, and to avoid government contracting, to resist anything that might threaten its reputation and place it in government cross-hairs. Don't be evil isn't the plan. Don't give the appearance of evil to some politician is the real plan.

As with all such plans, expect this one to fail. When it does, the best plan for Google becomes the road Microsoft refused to take.

Split up. Clouds do it all the time, in the face of mere wind.