Storage costs have reached a point where they attract significant attention from senior management. As a result, storage budgets are being carefully scrutinized, and managers are facing unprecedented pressure to reduce costs. While it's possible to cut budget line items here and there, addressing the systemic problem of rising storage costs requires a more strategic solution. One effective approach is to adopt an internal service provider...
or utility model.
Conceptually, this shift to a consolidated model based on standardized tiers of services makes sense--it drives efficiency through standardization and improved utilization, reduces risk by providing predictability and consistency, and ensures higher levels of service with offerings aligned with business needs.
This final element, alignment with business needs, is the key to a successful transition to an internal service provider model, but it can also be one of the biggest challenges. It requires teamwork and strong cooperation across a number of functions within the organization. This means not only working with other IT groups, but garnering the support and participation of the business units targeted by these services.
Storage teams working to gain the support and involvement of lines of business frequently express frustration over users' seeming lack of interest or unwillingness to improve efficiency and consistent levels of service. But old habits die hard. Fortunately, while transforming storage may be a challenge, we're not talking about a Herculean effort to bridge the gap between IT and business units.
The fundamental problem is a lack of alignment. The business is focused on revenue generation and profits, while the storage group is concerned with functional support and cost containment.
We see many organizations with most of their data residing on expensive Tier 1 storage. When budget pressure comes from top management, storage managers typically think about moving some of that data to lower tiers. But while a storage manager may be enthusiastic about implementing multiple tiers of storage to save money, alarms go off for the business manager responsible for revenue.
Clearly, the two points of view are out of sync. What's needed is an alignment of policy, process and incentives.
Policies are the internal laws that govern an organization. Strategic policies, in particular, set the tone for the entire organization and are the first step in ensuring that all groups are marching in the same direction. By "strategic," I'm referring to organization-wide guidelines from senior management that set the parameters for how storage is to be used and managed. These policies are applicable to business units and IT alike.
Unfortunately, many organizations lack a formal set of strategic storage policies. This doesn't mean policies don't exist--it simply means policy decisions about storage are made at a lower, tactical level by people who may not understand the full impact of those decisions. This results in the first source of misalignment between business and IT.
A good example of this misalignment concerns data retention policies. Who should decide how long an organization retains data? The answer forms the basis for a strategic policy that designates responsibility for data retention. Without a clear owner, the decision-maker often ends up being a storage or backup administrator who simply applies a default retention policy that may be wholly inappropriate.
Planning process alignment
The planning process is often another area of misalignment. Within the system development lifecycle, storage requirements aren't often considered until the project's final stages. At that point, the storage team has little opportunity to provide any input that could have a material impact on the deployment of storage for the project.
The storage team lacks a seat at the planning table, and planners may not be concerned with--or even aware of--the issues that impact storage. The challenge here involves process and organization. To affect systemic change, the planning process must be modified to ensure that storage is "designed" into the project.
From an organizational perspective, this requires a business analyst who's a member of the storage team. It's critical to present options and tradeoffs in business terms, not technical ones. Discussing the required attributes related to data protection, performance and availability in terms of the application and related business issues establishes credibility and demonstrates the benefit of involving storage in the planning process.
Goal and incentive alignment
Aligning policies and processes are significant strides toward building the cooperative relationship required to proceed with a services provider model, but an additional step is needed. The divergence of goals and incentives that may exist between the storage group and business unit must be closed.
As noted earlier, business is focused on revenue, while storage is focused on support and cost. What if it were possible for the groups to share each other's goals? The business unit needs to be assured that meeting revenue goals won't be compromised by changes within the storage infrastructure. Likewise, for the storage team to successfully contain costs, the business unit must share the cost-constraint burden. Some available tools, if applied properly, can help with these endeavors.
Service level agreements (SLAs). An SLA is a contract that commits (in writing) to a specified level of service with clearly stated metrics and remedies for non-compliance. SLAs are used on a regular basis in networking, and are an essential part of any outsourcing agreement, but they haven't been applied regularly to storage infrastructures.
It should be noted that an SLA is different than a goal. A storage group may have an implicit goal to provide 24x7 services to all users and respond immediately to an outage. However, it's unlikely they would contractually commit to this level of service for any but the most critical applications.
Cost modeling. If you had the choice between a Chevy and a BMW for the same price, which one would you pick? A fundamental reason for the disconnect between storage and business units is the lack of a shared burden with regard to cost. If business users feel no direct cost impact, storage is, in effect, free and they'll opt for Tier 1 every time.
A comprehensive cost model details the capital and operational cost of providing storage services. Combined with regular reporting on how storage assets are being utilized and by whom, it can go a long way to raise awareness of the impact of disk and tape storage consumption.
It should be noted that a cost model shouldn't simply be an allocation mechanism. Instead, it should be a tool that raises awareness and drives behaviors, such as efficiency, in accordance with corporate cost management initiatives.
In last month's Best Practices column (see The metric system), Stephen Foskett wrote about metrics. Successful implementation of SLAs and cost modeling is dependent on accurate metrics and reporting. It's tough to argue with hard facts, and a storage manager armed with the data to demonstrate adherence to SLAs or to show that utilization is growing at a triple-digit rate will make a far greater impact.
Ultimately, alignment is about building trust. A business unit or application owner needs to believe they can get the storage and services they need when they need them and in a non-disruptive manner. Overallocated, underutilized storage is often the result of previous experiences where obtaining storage was too difficult and provisioning took too long. As a result, users learn to inflate requests.
The goals of business and IT need not be at odds. When senior management is prepared to step up and provide the aligning force of policies as the first step, dramatic gains can be made in the practical alignment of IT with business goals by following the simple guidelines discussed here. Whether we're talking about revenue generation or operational costs, both are essential to the health of the organization. After all, isn't everyone concerned about the bottom line?
Note: In the January 2005 Best Practices column on data growth, I referred to a book by Chris Stakutis and John Webster prior to its publication. In the period between my article and the book's publication, the title changed to Inescapable Data: Harnessing the Power of Convergence.