Without question, the notion of the data center as a collection of hardware and software resources residing safely in a secure, on-site facility is coming to an end.
The cloud is only the latest facet of a long process of resource distribution that began even before the first megabit broadband connection was made. These days, it is the rare enterprise that doesn’t have some portion of its institutional knowledge stored on third-party infrastructure.
Symantec, in fact, estimates that nearly half of all data is kept outside the enterprise data center, with about 23 percent now residing on cloud services and much of the rest, somewhat disturbingly, sitting on laptops, smartphones and other mobile devices. The trend is accelerated among small and medium-sized businesses, which have long struggled with the cost of building internal infrastructure adequate to their needs. Increasingly, though, enterprises are faced with availability and reliability challenges from their outside resources, which is why Symantec and others are arguing that protection measures should shift from devices and infrastructure to the data itself.
Off-site services will only become more popular as providers work quickly to outdo their competitors in terms of cost and services. Already, new platforms are targeting the cloud market with capabilities like multitenancy and high-throughput operations. A case in point is SolidFire’s new all-SSD storage systems, the SF3010 and SF6010, designed to improve VM density to up to 216 per rack and push resource utilization up to 85 percent. That gives cloud providers the ability to meet QoS requirements for more users even as it lowers operating costs across the board.
By this point, however, much of the industry has moved beyond the public vs. private debate, with most organizations determined to create a hybrid environment. But exactly how will these be configured, and where will their respective responsibilities lie? According to Adept Cloud’s Frank-Robert Kline, a reasonable scenario would have applications and services on the public component, while data is kept in-house. This provides a great deal of flexibility when it comes to deploying new collaborative and mobile access services, while preserving data safely behind the corporate firewall. The one flaw in such a plan is that, as mentioned above, a good portion of enterprise data is already on the cloud, meaning enterprises will face both a significant capital expense and a whale of a migration problem getting it back.
Enterprises should beware not to become too diverse in their selection of cloud storage infrastructure, however. As InforWorld’s Galen Grumen points out, some applications work better with select cloud providers than others, so as the mobile environment becomes more diverse, file sharing and collaboration could prove complicated. This could be alleviated through some kind of integration layer or device through which all cloud access must pass. Companies like AppSense are moving in this direction with applications like DataNow, although he says the user interface needs a little work.
The unknown future is always more intimidating than the safe, comfortable past. In the case of cloud storage, however, enough experience has been gained so that most enterprises are starting to see more benefits and fewer liabilities.
It still means cloud environments need to be designed and implemented properly, but at least the pitfalls of distributed architectures are better understood.