For the most part, IT has treated cloud computing as an extension of traditional enterprise infrastructure. After all, it performs the same function, except there’s more of it and it sits on someone else’s hardware.
But as experience with the technology grows, it is becoming clear that the cloud differs from internal data environments in a number of ways — not all of them positive. So before we congratulate ourselves on expanding the infrastructure footprint at a cost that won’t make the CFO gag, it’s important to know what we’re really getting in the cloud.
The first thing to come to grips with is the distributed nature of the cloud. This may seem like no big deal at first, considering the enterprise has been using distributed architectures for decades. But as tech author Brian Proffitt notes, cloud distribution takes things to an entirely new level because data can now be disseminated across wildly disparate, non-integrated environments. That means institutional knowledge winds up in places that may not be accessible to the organization at large, and when employees leave or cloud providers close up shop, that data is out of reach.
And this “cloud fragmentation” is only likely to get worse, according to Younity CEO Erik Caso. Most cloud providers offer various tiers of service, with many starter programs offering free storage up to a set amount. As enterprise needs increase and more stringent levels of security and availability are required, many users simply jump to a new provider to take advantage of their free or low-cost tiers. It’s the same pattern credit counselors see when clients are over-extended: There’s always another credit card offering low teaser rates.
This may be part of the reason many enterprises are pulling back from the public cloud and are instead turning to the private cloud. In Europe, for example, two-thirds of top enterprises said they relied solely on their own data center infrastructure in 2012, according to a recent QuoCirca report. This was up from about 45 percent in 2011. It seems that many organizations are willing to kick the tires when it comes to third-party infrastructure, but in the end opt to keep data close to the vest.
But even then, the private cloud may not be all it’s cracked up to be, particularly if it is viewed as simply a new way to deploy the same old IT. As Forrester’s James Staten points out, most initial cloud deployments fail to take into consideration the benefits that the public cloud brings to users: namely, cost, convenience and autonomy. If the private cloud does not deliver the same value proposition as the public cloud, guess which one users will choose?
Of course, nearly all cloud services offer visibility, management and other tools to help keep track of data. The problem is, few of these work across multiple clouds, forcing the enterprise to maintain numerous management stacks just to stay on top of things. And again, this does nothing to help integrate data sets and maintain a cohesive environment for institutional knowledge.
Surely, a robust set of cloud policies would go a long way toward maintaining control of data, but these would require adequate means of enforcement. And if not designed and implemented carefully, they could drive more users into the public realm — that value proposition again.
Ultimately, though, it will be up to both the enterprise and cloud communities to work this out. The enterprise will always have a vested interest in maintaining robust infrastructure at the lowest possible cost, and cloud providers would love nothing better than to tap into the enormous revenue streams that enterprise-class services will provide.
Cooperation, then, would benefit us all, while continuation of current service patterns satisfies no one.