Every successful technology innovation follows a similar path from development, deployment and finally, acceptance.
The cloud is probably somewhere in the development/deployment phase, considering the continued introduction of new services and new architectures coupled with the rapid uptake of cloud computing models on the part of leading enterprises. Characteristically, then, we've moved past the heady days of the early roll-out period, when the possibilities seemed limitless, and are now steeling ourselves for the challenges ahead to make it all work.
However, few people are giving much thought to the end game. When all is said and done, what will enterprise architectures look like once the transition to the cloud has run its course?
According to Gartner, more than half of all enterprises will be on the cloud in one form or another by the end of the decade. The chief driver, not surprisingly, is the need to foster improved ROI for enterprise infrastructure. Most data centers have grown so unwieldy over the years that few organizations can calculate the value of what they have or whether it produces a net gain or loss to the organization. Cloud architectures, on the other hand, are purpose-built for ROI, given that users will know exactly what data services they are paying for and how well, or how poorly, that expense is contributing to the bottom line.
Still, the fact remains that cloud architectures and traditional data center designs represent radically different approaches to data infrastructure and operations, according to Gartner's Ray Paquet. This is true even when the enterprise data center adopts more cloud-like capabilities. First off, there is the fact that most public clouds scale horizontally, which means greater reliance on low-cost commodity hardware, with a software abstraction thrown in to protect applications from inevitable failures. As well, clouds tend to shun complicated SAN and NAS environments in favor of direct attached storage (DAS), with advanced replication services taking the place of expensive RAID configurations. In a way, then, the cloud harkens back to a more basic infrastructure, at least on the hardware level.
Of course, it's not like enterprises have much choice when it comes to embracing the cloud. As JOI Media's Brock Willis notes, the cloud is much more adept at meeting the mobile and collaborative needs of the modern work force, which demands that data be accessible to multiple users simultaneously, using a variety of devices. Gone are the days when sharing files via email and FTP servers was good enough, or that software had to reside on the end points with file access via a virtual private network. Centralized data repositories providing anywhere/anytime/by-any-means access is the name of the game now.
Still another reason why the cloud is inevitable is that is where enterprise software development is headed. Planview's Patrick Tickle notes that most startups these days are focusing exclusively on cloud-based solutions, forcing even long-time developers like him to follow suit in order to deliver the support, implementation and service cost/benefit ratios that users are coming to expect. At the moment, the only choice seems to be what type of cloud-based delivery model to go with: cloud-native, cloud-hosted or cloud-optimized.
Data infrastructure development has long been a matter of keeping up with the Joneses, spurred on by the continued pressure to do more with less. The cloud still generates unease in the front office regarding security and availability, but it won't be long before the operational benefits begin to show up on quarterly reports — and then it will be simply too hard to resist.