Cloud Computing`s Sky-High HypeBy Tony Kontzer | Posted 11-11-2009
Cloud Computing`s Sky-High Hype
Hype and IT go hand in hand. From local area networking to the Internet, new computing paradigms regularly come on the scene with much fanfare, only to have reality set in as the kinds of functionality and business models IT shops need take five, 10, even 20 years to evolve.
At this point, it's safe to group cloud computing in that evolutionary pantheon.
Over the last few years, technologies carrying the "cloud" label have poured into the market, promising huge reductions in investments, newfound flexibility in meeting processing demands and welcome advances in user friendliness. For many small and medium-sized companies, sales organizations and entrepreneurial IT teams, those benefits have been real. But as just about any big-company IT executive will attest, the forecast for cloud computing is murky, at best.
Sure, IT leaders will entrust their expense management and human-resources applications to software-as-a-service (SaaS) providers; turn to platform-as-a-service (PaaS) offerings to build niche cloud applications; or dial up infrastructure-as-a-service when they need temporary computing resources. But a host of factors is preventing the enterprise from placing mission-critical systems in the cloud.
In addition to perceived shortcomings in areas such as security, compliance and monitoring, the cloud has been undermined by the absence of a business model that fully reflects the technology's flexibility. And the growing desperation among vendors to slap the word "cloud" on any product or service that has even a single cloudlike feature has raised CIOs' skepticism.
"At this point, anyone who is offering anything Web-based has found that using the term 'cloud' in the product name will get more coverage in the media than if they described it another way," says Dan Kusnetzky, vice president of research operations for IT research firm The 451 Group.
Plus, often lost amid all of the hand-wringing over definitions, pricing models and technology shortcomings is the well-known secret about the cloud: The individual components aren't all that new. SaaS? Not all that different from the application service providers of the late 1990s. Accessing pooled computing resources via the cloud? Sounds suspiciously like grid computing.
If it sounds like a market rife with confusion, well, it is. But a lot of it could be cut through if only everyone would agree on exactly what cloud computing is.
The way Kusnetzky sees it, a product or service must exhibit a number of characteristics to be considered a true cloud offering. It must be publicly accessible through any Web-connected device. It must feature programmatic interfaces and reside on a multi-tenant architecture with separation between clients. It must offer customers granular cost visibility, elasticity and rich management capabilities. And it must allow customers to rapidly provision resources through a self-service interface. "Not everything people are presenting as a cloud has these characteristics," Kusnetzky says.
Debunking the buzz
Debunking the buzz
Yuvi Kochar, CTO for The Washington Post Co., says that the lack of an agreed-upon definition of the cloud forces IT executives to spend too much time determining whether products fit into the category. "There are many existing services that everyone has tried to morph into cloud computing, and some of those are relevant and rightly fit into that definition," Kochar says. But, he notes, "a souped-up version of virtualization doesn't fit into a definition of cloud computing."
And even when vendors' offerings can legitimately be called cloud computing, many are being tripped up by their own underdeveloped approaches to pricing.
"There aren't enough business models emerging out of the cloud beyond the raw cost proposition," says K. Ananth Krishnan, CTO for Tata Consultancy Services. "I think the 'aha' moment is yet to come."
Actually, some CIOs have had an "aha" moment, as in: "Aha! I should have known the big vendors are still doing things the old way." Jerry Batt, CIO of Pulte Homes, found that when his business went south due to the subprime crisis, he wasn't able to ramp down his investment in a cloud service as much as he wanted to.
Batt acknowledges that he may have jumped on too quickly. His issues with pricing inflexibility came on the heels of numerous technical hiccups that unexpectedly resulted from integrating his internal systems with the cloud service. But as frustrating as those problems were, he says the lack of flexible pricing--which he attributes, in part, to over-consolidation in the vendor market--is responsible for his mistrust of the cloud.
But Krishnan believes that the perceived shortcomings of cloud computing have as much to do with the impatience of the IT world as they do with the immaturity of the market. He points to Nicholas Carr's suggestion, in his most recent book, The Big Switch, that a model for delivering metered IT could evolve much the way it has for electricity. The catch? That electricity-delivery model took shape over 100 years; the cloud computing market has been forming for less than a decade.
In other words, CIOs can learn a valuable lesson about patience from Batt and other who may have over-committed to cloud computing before it could really live up to the hype. That doesn't mean CIOs should crawl inside a hole and wait for a mature cloud computing model to emerge. Rather, they should pick carefully by auditing each of their applications to determine whether the requirements make them good candidates for the cloud--when it finally comes along.
There are plenty of business needs adequately met by today's cloud. On-demand applications are enabling IT shops to outsource non-critical business processes, such as contact management or expense reporting. Platform-as-a-service offerings are letting companies build their own tools to manage similar processes. And renting server time and processing power can be ideal to accommodate entrepreneurial IT projects.
Until the cloud market evolves to where it can deliver the kind of capabilities that will make choosing it a no-brainer for meeting broad enterprise needs, CIOs should adjust their expectations. By doing so, says Kusnetzky, "they may find that some of what they're doing is available today as a product in the cloud."