Published: 14 Apr 2003
According to the Meta Group, Stamford, CT, storage capacity will grow at a compound annual rate of 92% through 2005. That's quite a lot of new spinning disk year after year. CIOs will have to find or buy excess capacity to accommodate this new requirement.
The good news is that disk drive prices continue to spiral downward to the tune of 40% to 50% per year. But the real cost isn't hardware--it's storage operations. Peripheral Research, Santa Barbara, CA, claims that $1 of storage hardware carries a cost of $7 for operations. With budgets flat and headcounts frozen, how can IT departments meet this operational challenge?
The obvious answer is through process automation. Throw some software from Computer Associates (CA), EMC, Tivoli, or Veritas at the problem, configure these tools to meet your storage policies and you're done, right? Wrong.
Today's storage management software is incredibly immature, proprietary and limited and likely to remain so for another two to three years. Rudimentary storage management standards are just starting to be ratified and will take time to roll out into products. In the meantime, storage management vendors can swap APIs or write system calls to other vendors' command line interfaces, but this is a painfully slow process that won't come close to solving real enterprise storage problems.
Focus on process
IT managers take heart. The lack of credible storage management software options is actually a blessing in disguise. The dearth of software tools provides an opportunity to address storage growth with business-driven processes and procedures that deliver ROI to the business and IT, while actually solving the problem.
The roadmap for processes and procedures should focus on these three points:
- Accommodating business goals
- Classifying data
- Setting the right SLAs
With these points in mind, IT managers will know where they are and where they must to go. Let's examine each in detail.
IT executives must meet with business managers, CEOs and CFOs to fully understand upcoming business projects and initiatives. What's coming up to drive new revenue streams? Are there any regulatory compliance issues to meet? How will the company touch its customers? Will there be any business process automation? The answer to these questions should help CIOs appreciate the scope of future requirements so they can start to plan for new storage capacity, infrastructure and operations. Once IT managers comprehend the business needs, they should negotiate their staffing and budgetary requirements accordingly.
With storage capacity nearly doubling each year, no company can continue to manage all data the same. To solve this quagmire, the Enterprise Storage Group recommends that enterprise companies classify all enterprise data. One suggestion is to group data into three buckets: critical business data, critical operational data and non-critical operational data. Examples of critical business data are customer databases, transaction systems or product inventory. Critical operational data may be e-mail or user files. Non-critical operational data may be reference data and archives. Once you've established a data classification scheme, you can choose a matching infrastructure, operational policy and staffing model.
Based upon these classifications, choose service levels that meet the business needs for each group. Between data classification and SLA categories, you should be able to better scale the IT department while still delivering proper services.
Now that you know where you want to go, how can you get there? These five key steps are a must:
Establish IT governance. IT departments need standard operating procedures to deliver consistent, repeatable, measurable results. A good place to start is with IT governance models such as the IT infrastructure library (ITIL), IT services management (ITSM) and the control objectives for information and related technology (CobiT). These models provide best practices for an array of IT activities such as configuration management, availability management, service management and security. Operational models shouldn't be approached lightly, as implementing them can be projects unto themselves. Nevertheless, a good operational foundation is imperative to cope with ever-growing storage needs. At the very least, CIOs should look to well-established archetypes such as ITIL, ITSM and CobiT for ideas--especially to help with problem areas.
Set up policies and procedures based upon data classification and SLAs. Now is the time to set up specific operational models. Policies and procedures need to be detailed, specific and a match for each data classification. Set up specific policies and procedures for each and every operational process. For example, IT's response to error messages should be different, based upon whether they come from a storage system housing mission-critical data, critical operational data or non-critical operational data. Mission-critical systems need immediate attention and escalation while the others may best be served by standard escalation procedures by level one support.
Dedicate the right staff to the right projects. Make sure your most skilled personnel are responsible for supporting mission-critical data. Don't forget to give junior staff members the chance for mentoring and career growth. As storage grows and a new critical system comes online, this training experience will pay dividends when you have fully-trained personnel to slot into new roles.
Select the important data to monitor. This is where the lack of mature storage management tools forces you to pay careful attention. All of the storage equipment provides oodles of management data, but capturing and analyzing that data can be a full-time job involving manual processes, spreadsheets and automated scripts. Manage this process by sorting through the mountains of data and defining the two or three metrics you need across the entire storage infrastructure. For example, you may want to capture utilization, performance and error logs from every device to give you a high-level picture of end-to-end performance and behavior. You'll still have to dig into each device, but at least you'll have a pretty good cursory understanding of the entire storage system. Review this data often because it can help you identify trouble spots and plan for future needs.
Define the right metrics and reports. IT managers must monitor business and technical statistics to determine if they are meeting business commitments. They also must check under the hood to monitor the storage and IT infrastructure. To do this accurately, companies should choose the right metrics and then monitor and report on them frequently.
The processes defined in this column will help IT managers cope with storage growth, while storage management vendors evolve and improve their current elementary tools. Eventually, storage management software will actually be ready for prime time. You'll get more from those tools, though, if you've done the hard work of understanding what processes and procedures you need.
- Tiered Storage - Optimizing the Storage Infrastructure –Fujifilm Recording Media USA, Inc.
- Illuminating Insight for Unstructured Data at Scale –IBM