Using Storage to Resolve System Dilemmas

By Jesse Freund

From the "damned if you do, damned if you don't" file comes data storage. Any CIO can tell you it's expensive to archive data, but the cost of not storing business information is almost certainly higher. Thus, storage capacities are growing more than 50 percent annually at most companies, and closer to 100 percent at many others, according to the Enterprise Storage Group, an industry research organization. In order to understand how businesses can reduce the cost of storage, and even turn storage systems into an engine for ROI, it's important to examine the source of all the data.

Back it up

Historically, the need to recover data in the event of a system failure has been the major contributor to ever increasing storage capacities. While most businesses complain about the high cost and laborious process of backing up data, these same enterprises realize that the results of not doing so represents a greater danger. And, yet, according to Enterprise Research analyst Peter Gerr, one-fifth of all backups fail. "Ask a company which day of the week they don't want to have backed up, and they'll look at you like you're crazy," he explains. "But that's what's happening." In other words, despite being faulty, expensive, and time-consuming, most businesses stomach backups because a total system loss is even more unpalatable.

Compliance confusion

In many industries, the need for backup capability has been sent into exponential overdrive by the explosion of compliance issues. Today, more than 10,000 local, state, and federal laws and regulations govern how data must be protected, monitored, maintained, and secured. These regulations represent a double-edged sword: More data must be retained, and this data must be stored longer. For example, in the healthcare industry, HIPAA regulations mandate that patient data must be preserved for the life of a patient plus two years, which is no small feat when humans typically live 70-plus years and a single CAT scan alone can create gigabytes of data. According to the Enterprise Research Group, compliance records will grow from 376 petabytes of data in 2003 to 1,622 petabytes in 2006.

Reference this

While data has traditionally been stored for emergency recovery purposes, and increasingly for compliance, the digitization of corporate assets has breathed new life into old data. Today's businesses need to retain copies of digital assets not just for recovery in case of emergency but also for active reference as a strategic asset. Items like contracts, e-mails, check images, transaction records, and medical information need to be stored and retrieved on a regular basis. The importance of retrieval has changed the nature of recovery and, as a result, led the charge away from tape-based storage technologies towards disk-based solutions that offer faster and more reliable access.

Identifying pain points

The rapidly changing data storage environment and the exponential growth of data itself represent a major problem at many organizations. According to the Storage Networking Industry Association, the top five pain points associated with storage are:

1. Cost
2. The challenge of managing growth and meeting capacity
3. The inability to manage storage assets and infrastructure
4. The lack of integrated and/or interoperable solutions
5. Increasing complexity of storage information

Because storage is typically manually intensive and time consuming, it's not surprising that cost is the greatest concern associated with storage. In fact, pain points 1 and 2 directly relate to cost. But the high cost of storage also creates other indirect problems. Because of the expense, businesses tend to start small and grow storage solutions in a piecemeal fashion. This typically results in storage islands with multiple management domains and numerous tool sets. As pain points 3 through 5 note, storage infrastructures tend to become overly complex, heterogeneous, and unwieldy as a result of the high cost of storage technologies.

Assign data a business value

Any storage strategy must begin with an audit of storage needs and practices. What kind of data needs to be highly available? What kind of connectivity and reliability does this demand? How long should backup and recovery take? How will future growth impact data storage requirements? The goal behind asking the tough questions is to align storage solutions with business needs. Companies must define data to determine what's public and what's not. Businesses must also identify what's in data repositories, and determine the applications used to create the data. And, finally, enterprises must figure out acceptable service levels. Do employees need immediate access to stored emails? Should recovery take four hours or four days?

Centralize, centralize, centralize

Once the data has been assigned a business value, companies need to develop a centralized approach to process, decision-making, and infrastructure. Storage must become an organization-wide priority, and departments across an organization must buy in to the storage strategy. The goal of a unified storage strategy and environment is to develop a universal administrative tool and a common file system for all storage elements. By consolidating data and removing technological redundancies, as well as streamlining purchases and approval, a centralized storage strategy can pay big dividends.

Toward tiered storage

In the end, some data needs to be stashed away for a rainy day: Some data needs to be stored and protected to meet compliance requirements. And some data needs to be made highly available.

In the dynamic world of data storage, where most companies already possess a disjointed patchwork of technologies and systems, the notion of tying these resources together to create a tiered storage strategy represents a great hope and an even bigger challenge. "Most organizations already have a patchwork of technologies, so tiered storage makes sense. You don't want to put your least valuable data on your highest cost and best performing technology. You want to put the right data on the right device," explains Gerr. "We're still in the dark ages of understanding storage as a corporate asset. We've been storing with blinders on for a long time."

Jesse Freund is a Contributing Writer at Business 2.0 and a frequent contributor to Wired.