Red Hat's Deltacloud at Tip of Epic Change

by Paul Rubens

OS Roundup: With arguments against cloud computing becoming irrelevant, Red Hat and other OS vendors are scrambling for slots as key players in the 21st Century data center -- wherever it may reside.

Earlier this month Red Hat announced its Deltacloud initiative to "enable an ecosystem of developers, tools, scripts, and applications which can interoperate across the public and private clouds."

The potential impact on the ability of users, developers and IT departments to consume cloud services via a common set of tools is, according to Brian Stevens, Red Hat's CTO, "epic."

These days, you must be careful when using the "E" word, because often as not, it's accompanied by its nemesis: the "F " word — as in "epic fail." Especially when talking about cloud computing. Plenty of arguments are made against the concept, like "It was done in the '60s with a different name and we've moved on," or "It's just the ASP model all over again, " or "It's not secure." And there's always the venerable old favorite: "It's not reliable."

It's going to take more than a few ropey arguments to stop operating systems companies like Red Hat from moving toward a cloudy future. The first two arguments — about the '60s and ASPs — are pretty much true, but they aren't arguments against cloud computing at all. All they are really saying is that the concept was tried, but the technology and infrastructure required weren't up to it at the time. Big deal: Leonardo Da Vinci tried to make a helicopter back in the 15th century and failed, as he didn't have a suitable engine. It doesn't mean helicopters were destined never to work — just ask Igor Sikorsky.

Security is another red herring: Of course it's of paramount importance to many organizations, but third-party cloud providers may well have the resources and skills to be more secure than your average corporation. Besides, if you run your cloud internally, the issue of handing data and security responsibilities to a third party pretty much goes away.

What about reliability then? Let's look at Google, and its recent "Gfail" caper during which Gmail was unavailable for about 100 minutes at the start of September. Anyone relying on Google's cloud-based email solution was left stranded and unable to do anything about the lack of service, totally reliant on Google engineers to identify the problem and get it sorted sharpish.

Which, of course, they did. Now can you tell me honestly, hand on your heart, that your organization has never had an unplanned email outage? Of course it has. Corporate email systems have an average of 53 minutes of downtime each month, according to Wash.-based Osterman Research. Can you imagine the outcry if Gmail was down for 53 minutes a month?

Like it or not, cloud computing — whatever that means — is not going away. And that's going to present a wagon load of opportunities and threats for anyone who makes enterprise operating systems, of both the server and desktop variety. It's impossible to predict who will end up top dog once cloud computing really makes an impact. One thing is crystal clear to Microsoft, Red Hat, Canonical and many other operating system companies: Doing nothing is not an option. You have to be in to win.

Paul Rubens is a journalist based in Marlow on Thames, England. He has been programming, tinkering and generally sitting in front of computer screens since his first encounter with a DEC PDP-11 in 1979.

Follow ServerWatch on Twitter

This article was originally published on Tuesday Sep 22nd 2009
Mobile Site | Full Site