Security, availability and performance — you can almost recite the enterprise's list of concerns over cloud computing by memory these days. That doesn't mean these issues have been overblown or overhyped, but it does mean organizations should pay attention to the potential consequences of placing ever-greater reliance on infrastructure that is not their own.
In fact, more than a few voices are calling attention to what they say are key blind spots in the drive to reap the cost-savings and flexibility gains that the cloud promises. These could certainly come back to haunt the industry if the cloud ultimately proves too unwieldy for enterprise-class data requirements, particularly if organizations aren't able duplicate popular services on their internal infrastructure.
Nearly all of the top concerns stem from one of the fundamental aspects of the cloud: loss of control. As Mezeo Software found in its latest survey, more than 80 percent of CIOs rank things like data leakage as a top worry. And the uncomfortable truth is that cloud computing will likely assume a role in the enterprise whether it is actively sought or not. With the BYOD and mobile communications trends taking hold, a fair amount of organizational data is bound for cloud storage unless strict governance and policy guidelines are put in place. Unfortunately, less than half of survey respondents said they are actively working on the problem.
New technologies like the cloud are usually adopted as a means to enhance performance. But is it possible that the cloud can actually degrade data operations? According to tech consultant Rick Blaisdell, it can if key pieces of the cloud infrastructure, primarily those governing network latency and bandwidth, memory, storage and CPU performance, are not carefully managed. As things are shaping up now, enterprises are already laying the groundwork for a fully federated data infrastructure, with various resources and applications deployed across traditional, cloud and hybrid tiers. Expect performance issues to persist until IT can figure out what goes where.
But if the cloud poses a risk to data environments as they are now, what about all the new developments that are supposedly optimized for broad-based, distributed architectures? Collaborative environments, for example, are supposed to be tailor-made for the kind of scalability and universality that the cloud entails. That may be true, according to Forrester Research, but until collaboration vendors can demonstrate their ability to handle enterprise-class services and a commitment to ensuring their availability over the long run, adoption of these new environments will be tepid at best.
There is also the danger that successful cloud deployments can be laid low by unforeseen circumstances. According to readwriteweb.com's Brian Proffitt, up to half of all email traversing the Web these days is generated automatically as devices and applications exchange information in line with their duties. What happens when an app sends a message to another app that has been moved to the cloud? In many cases, it will send another message, and then another, until a message is received. Multiply this by hundreds or thousands of apps, and enterprises moving to the cloud could be setting themselves up for their own denial-of-service attack.
Despite all of this, and a number of high-profile cloud failures in recent months, enthusiasm for cloud services appears to be on the rise. This makes sense because the cloud is still new and all its promises are front and center at most organizations.
The warts won't become known until enterprises start to place critical reliance on the cloud. By then, of course, it will be too late to turn back. And in the end, IT will have to do what it always does when technology presents an intractable problem: Develop new technology to solve it.