Four Ways to Solve Cloud Integration, for Better or Worse
The cloud is cheaper than standard IT infrastructure. This has been a given for so long that hardly anyone questions its veracity. And after all, who would argue the cost advantages of leasing resources from an outside provider vs. building and maintaining complex internal infrastructure?
Well, some very bright minds in the IT industry are starting to do just that.
Rob Enderle, writing for CIO.com, for one, notes that speed and flexibility, while important, do not necessarily translate into lower costs. The hard truth, of course, is that spending on IT is dropping while spending on cloud services is increasing, but this has more to do with timing and availability rather than simple economics. Indeed, recent analyses show that once internal infrastructure begins deploying cloud services of its own, it can meet enterprise needs for about $100 per user while Amazon and other providers come in at around $200 per user when purchased individually or in small groups, as is the practice with many business units. And a proprietary platform like Oracle can run as much as $500 per user.
It would seem, then, that when enterprise executives rush to implement cloud services, they turn to the most convenient solution, rather than the cheapest. According to 2ND Watch, more than 90 percent of business units have adopted cloud-based services while 60 percent have gone around IT to implement them. At the same time, about 74 percent of IT shops are delivering internal cloud services, which is impressive at this early stage of the game, but they are responsible only for about 37 percent of the cloud activity in the enterprise industry. To me, that indicates that many users would gladly accept cloud services from in-house sources but look to the outside nonetheless because internal options are either not available or cannot be delivered in a timely fashion.
But this state of affairs might not go on much longer as demand for cloud services increases. According to new research from Frost & Sullivan, cloud providers have been so focused on building capacity and otherwise improving infrastructure over the past few years that little attention has gone to things like contract automation, invoice tracking and client compensation. In many cases, these functions are handled manually, which gets to be problematic when the number of contracts increases and expectations for customized service exist across an increasingly diverse customer base. If this continues, cloud providers could find themselves in the same position as telecommunications carriers who are still struggling with bundled service fulfillment and the management of multiple business-to-business contracts.
What’s needed, then, is a way to predict how the pay-as-you-go pricing structure of the cloud will affect user costs in the face of changing usage patterns and other variables, says RightScale’s Hassan Hosseini. The cloud, after all, is designed to provide elastic services to meet the needs of dynamic data environments, but if the pricing structure is fixed, it could leave some organizations under water. Key to implementing a proper cost model, then, is to determine the kind of growth that is expected—steady, continual growth, seasonal fluctuations or intermittent spikes following, say, a new product launch—and then build those expectations into service agreements. Forecasting is not an exact science, but at least a set of reasonable expectations can be leveraged to smooth out the cost of long-term cloud dependencies.
Ultimately, knowledge workers are likely to employ a mix of public, private and hybrid cloud services as they seek the most efficient and effective means to accomplish their goals. Cost is but one factor in the equation to determine optimal utility, although it is an important one.
But don’t get caught in the trap of thinking the cloud is automatically cheaper than internal IT. As enterprise infrastructure becomes more cloud-like and third-party providers struggle to meet the needs of an increasingly diverse data community, local resources will likely prove their worth again very soon.
