Red Hat: Cloud will anoint next Microsoft

Openness is a funny kind of thing

As you might expect, Red Hat is waving the banner for openness in clouds. And the three principles that Whitehurst wants everyone to agree one are openness, transparency, and choice.

Openness is a funny kind of thing, as he explained, such as when Facebook and Google work together (either directly or through proxies) on various projects and standards. Red Hat wants cloud computing standards to evolve out in the open, and for customers to have the choice of running their applications on any cloud, written in any language, and have them move around at will.

"We are at a very significant inflection point that will determine if clouds will develop with users determining the direction or some big mega-vendors are going to co-opt the terminology and deliver the same old stuff under new names that they have been delivering for years."

Ahem. How many hypervisor and management tool stacks are now being called clouds? All of them, that's right. How many servers with some switches and storage running this software hodge-podge are called clouds in a box? Same answer.

What makes CloudForms, Red Hat's latest incarnation of its cloudy server and management stack, any different? The fact that it will eventually be available as open source code? Ditto for OpenShift, Red Hat's platform cloud, also announced this week and based on CloudForms.

The real question, and Whitehurst hit it precisely, is that we are at a point in the history of computing where we might be choosing the next Microsoft. (That next Microsoft could be VMware, or it could be Microsoft, or it could be Red Hat, Google, Apple, or some other company.)

Microsoft is coming at ya

"If we opt for the old business models, then what we are deciding right now is who is going to be the next Microsoft." In a typical paradigm shift, as we are undergoing now according to Whitehurst, from distributed, static computing to cloudy, flexible computing, you only get one or two winners. I would argue sometimes three. You get a winner and two contenders, typically. But the point is still the same. "And if we are just choosing the next Microsoft, then there is only going to be a certain amount of value that can be created, and I think we know where the vast majority of that value will be extracted."

The dominant vendors get to roll in mountains of dough. Red Hat may be a lot of things, but even at the $1bn run rate it currently has for annual revenues, it is not rolling in it compared to proprietary system vendors. Look at what Microsoft has been able to do by eliminating the competition on the desktop for more than a decade and then transferring that to dominant share in servers. Microsoft wants your phone now, too.

"In the next generation, we have a hypervisor, and in choosing it, you are determining who the next Microsoft will be – whether there will be an open source alternative and choice," Whitehurst said. "The only rational choice when you have the opportunity is to be open."

What this argument about open versus closed misses is that might still makes right. (Well, it doesn't actually do that, but it gives that effect. Right makes right, after all.) Red Hat itself is bragging this week how the CloudForm stack for provisioning bare metal and virtual servers and packing up applications into templates so they can be fluffed out onto many different types of clouds, is based on integrating 65 different open source projects. Only very large companies have the resources to buy control of these projects, as Red Hat has done, or participate in the ones that it doesn't control and then integrate it into a cohesive package.

Who else in the open source community but Red Hat and possibly Citrix Systems has the resources to do this in 2011? (Red Hat didn't mention OpenStack, the alternative open source cloud fluffer championed by NASA and Rackspace, in any of the keynotes at Red Hat Summit. Funny that.) VMware probably has more people working on its ESX hypervisor and vSphere and vCloud extensions than Microsoft has working on Windows Server, Hyper-V, and Systems Center.

The real question you have to ask yourself is this: do you want the architecture and the actual code of the clouds of the near future and far future to be determined by the meritocracy of the open source community, which has its own prejudices and politics, or proprietary software vendors like Amazon, Google, Microsoft, and VMware, which are in it for the profits as much as the prestige.

In the end, there will be many clouds, they will be annoyingly close to compatible in some cases, but not quite. And they will all give us something to complain about and a way to make a living for years to come. Just like mainframes, minicomputers, Unix, and client/server before clouds. ®