Earlier, I wrote about Microsoft's announcement which gives support for ConfigMgr 2012's management of Windows Azure Virtual Machines. You can read that here: On-premise ConfigMgr 2012 Comes to the Cloud.

That brought up an interesting discussion and thought process that I think is worth highlighting.

As Microsoft moves steadily into positioning Windows Azure as the cure-all for IT services, there's one single thing that stands as a serious roadblock to their ultimate success. I know you've heard me talk about IT Pros being a sort of roadblock that Microsoft has to compete with for business computing dollars, and there are many that hold security and privacy, in relation to the NSA overreach, as a prominent factor for anti-Cloud adoption, but there's also one more component that has yet to be fully addressed: Availability.

Vendors promise such severely high SLAs for their Cloud products that they can never hope to achieve them. SLAs are really just marketing fodder. Whoever puts the biggest SLA in print, seems to get some sort of market value from it and a lot of Cloud vendors have gone ahead and attached financial promises to failed SLAs. But, in the end, as we've seen reported over the last many months, those SLAs simply cannot be achieved. Amazon, Google, Microsoft and others have had multiple reported outages over the last year.

As Microsoft and other vendors position the Cloud as a business computing solution, downtime is not an option and they don't seem to get that. Today, with apps, services, and data residing on-premise, business users understand who to contact when there are issues, and the IT group is responsible for SLAs they are actually experienced and successful at attaining.

During today's discussions it became clear that availability and redundancy is still a serious issue, and even if a few companies do attempt a Cloud move, they'll quickly pull back when outages become an issue. The thought is that Cloud vendors are approaching services development all wrong.

The industry should be modeling the Cloud after email spammers and data thieves. You never hear of a Zombie server breaking or going offline for hours or days. It actually takes a massive effort to stop them. The redundancy built into the processes for infection proliferation is legendary. The bad guys are utilizing Cloud-based services so well that it takes multiple vendors and sometimes federal agencies to shut them down. On the reverse, the good guys fall on their faces when they stub a toe.

Until Cloud vendors can change their perspective and build their services with a keen eye to availability and redundancy, the Cloud is a neat thing for consumers who can just go watch TV if the Internet is down, but it's nothing more than that.