In-Depth

Creating a Cloud Storage Strategy

When it comes to storing data, there are numerous choices IT can make, all with varying levels of complexity.

• By Bruce Hoard
• 08/01/2012

Cloud storage may well be the new frontier in protecting data: not fully tested, but full of promise for pioneering types.

Gene Ruth, a Gartner Inc. research director, says there's a wide variation when it comes to what users are willing to do with the cloud. Depending on how risk-averse large organizations are, some are using cloud compute in a serious and broadly applied fashion. Others are advancing more slowly and building private cloud environments that include the cloud compute -- which, as Ruth points out, is very different from cloud storage. These slow movers might start with private cloud capabilities, but they're gravitating toward hybrid environments in which they can push some of the workloads out into the cloud. Doing that is "relatively complex" operationally, because it requires management tools that can survey entire internal and external environments, which Ruth says are "two separate gardens." Unfortunately, these tools either don't exist or are only now surfacing.

Cutting-Edge Tools
Not every company or organization stays on top of the evolving enterprise storage/virtualization/cloud landscape. Ruth says some Gartner clients are very advanced and know all the right questions to ask, while others are less involved with IT because it's not a strategic pillar, so they tend to have smaller IT staffs. Overall, he believes that most companies replacing or acquiring new storage equipment understand there are a multitude of advanced technologies -- such as thin provisioning, solid-state drives (SSDs), auto-tiering and de-duplication -- to evaluate.

The issue in the minds of many potential buyers is the credibility of these cutting-edge tools. Generally speaking, the more visionary the company, the more likely it is to be looking further forward and wondering how to best integrate cloud capabilities into its operations in an effort to achieve a lower total cost of ownership, all while becoming more agile and focusing on providing its businesses with a competitive edge.

Ruth's advice -- aimed at organizations starting from the ground up on implementing enterprise storage in virtualized datacenters with advanced cloud environments -- changes with the degree of risk such organizations are willing to take. For the risk-averse, he suggests a more traditional approach that involves building out the lower portion of the storage stack by using established products from large vendors such as IBM Corp., EMC Corp., NetApp, Hitachi Data Systems Corp. and Hewlett-Packard Co.

"When you go up the next step, you're typically in a cloud-like environment, where you want to build a common provisioning point -- a consolidation point," he says. "That's where it gets a little trickier."

Smart Software
Having reached this juncture in the road, some users look toward storage virtualization appliances, while others look to use storage resource managers to provide a common management umbrella over all of the heterogeneous storage residing lower in the stack. Both of these tactics have their advantages and disadvantages.

The third choice is another category of products created to logically provide common provisioning service levels while provisioning heterogeneous storage environments, and ultimately giving IT administrators views into storage that are more service-oriented and less technologically defined. Ruth says the companies in this space are developing software that automates workflows handled by skilled operators.

Ruth says the emerging family of cloud gateways is essential to connecting server virtualization to the public cloud. These gateways look and feel like storage arrays, but they connect on the back-end to public services providers such as Microsoft Windows Azure and others to provide WAN optimization, de-duplication and encryption -- all the integral technologies that make it appear as though public cloud storage is local. This configuration enables big organizations to take initial steps forward and use public storage for unstructured data files, and as a support mechanism for relatively small server virtualization environments.