MoneyWatch Tech

Cloud Outages: Are Vendors Selling Before They're Ready?

By Erik Sherman

Updated on: August 18, 2008 / 11:07 AM EDT / MoneyWatch

It's official: the media declares that cloud computing is changing the world. For anyone who's made more than one circuit around the high tech block, this sounds like many other bits of jargon -- SaaS, MSP, ASP, web services, thin client, client server, even mainframe timesharing -- and for good reason. The concept of running, from a single location, major computing services for thousands of users is actually an old one.

But in their rush to jump aboard the latest fad gravy train, many companies are stumbling and taking their customers with them. As a result, they have tossed the mud that will sully the cloud name and require yet another new term in the future to entice customers now jaded and distrustful.

There seems to be no end to the corporations that are declaring themselves "cloud computing" providers, at least in intent:

HP, Intel, and Yahoo working together
IBM
Dell
AT&T

Clearly the smell of freshly cut budgets has hit the water. Dell even tried, unsuccessfully, to trademark the term "cloud computing." Many assume that clouds are lined with gold, not silver. There is also a ready supply of thunderstorms hitting the unwary:

Apple's MobileMe is a disaster.
Google has seenmajor outages.
Amazon's storage offering took a tumble.
Citrix -- long a purveyor of "thin client" technology -- found its GoToMeeting and GoToWebinar services fail as well.

Providing high availability, high redundancy, low downtime services is far more difficult than many think. As happens virtually every time some new fad, or just some new name for an old fad, hits the streets, companies leap in without understanding exactly what they're getting themselves and their customers into. The result is one pile-up after another until a valid concept gets tarnished and many who could have benefited write the whole thing off.

In the case of cloud computing, I think there are some distinct factors driving the outages we've seen:

Hosting services that must be "always on" is a lot different from providing technology to data centers regain control of far-flung desktop applications.
Even if a company provides services, there's a big difference between offering intermittent action, such has taking an order, running a search, or handling an email, and continuous use of applications that puts far more strain on a computing architecture.
A company needs experience dealing with both the desktop and the server or data center, because the interaction between these two is far more complex than hanging a terminal, or terminal equivalent, off a central machine.
If you've never had to absolutely, positively, cross-your-heart-and-hope-to-die promise services to customers with savvy, money, and a willingness to stake you out over an ant hill if you mess up, then you're jumping into deep, shark-infested waters armed only with a rubber ducky.
One size doesn't fit all. Serving individuals with low-intensity needs is a completely different ballgame than taking on a medium to large organization. The same approach, infrastructure, and procedures won't satisfy both.
Offering computing services can't be an afterthought; it must be a core business. If you aren't ready for the operational, legal, and ethical issues that come up, you simply shouldn't be involved. Running a big data center, no matter how large, doesn't translate into hosting competence.

Some later comers to the cloud party, such as Microsoft and IBM have received some flack over not providing "real" cloud computing services. Big deal. Here are two companies that actually have the experience to make something work and the track record with corporate IT departments. If they're tentative or measured in how they approach a market opportunity, maybe they know something that many of these other companies don't know but shoudl learn in a hurry.

For once I wish that more vendors, instead of being driven by the need for immediately financial gratification, would consider what they were doing. It's generally far smarter to stay out of a market unless or until you can develop the chops to make customers happy. Otherwise, reality tends to rain on your business plan parade and leave everyone around you drenched.

Storm clouds image via stock.xchng user cempey, standard site license.