Demystifying Enterprise Tiered Storage (Part 1 of 2)
Views of tiered storage differ wildly because there is no clear, concise, universally accepted definition of the technology. We offer a definition that spells out the four attributes of tiered storage.
In recent months, “tiered storage” has entered the IT lexicon as storage vendors hawk their wares amid shrinking IT budgets.
Many IT managers are excited about the cost savings promised by tiered storage yet may be hesitant to adopt it, fearing additional complexity and procedural problems may be introduced into their environment. Storage administrators may voice different concerns: they focus on how to effectively deploy devices that are “so different” into the environment.
These two disparate viewpoints arise because there is no clear, concise, universally accepted definition of tiered storage. Some consider tiered storage to be two or more types of storage based on their access speed, while others cite reduced cost or functionality as the single defining factor. Obviously, a more concise definition is required. We suggest:
Tiered storage is a data-storage environment consisting of two or more kinds of storage delineated by differences in at least one of these four attributes: price, performance, capacity, and function. Any significant difference in one or more of the four defining attributes can be sufficient to justify a separate storage tier.
Using this slightly broader definition, we realize that tiered storage is already implemented throughout the mainframe environment. In fact, since both tape and disk have recently celebrated their Golden Anniversaries, tiered storage has been deployed for the last fifty years. IBM’s first tape drive, the 726 Magnetic Tape Recorder was announced in 1952, while the first disk product, the 350 RAMAC ("random-access method of accounting and control") was introduced four years later.
For many years, these were thought of as individual and very different storage products, rather than components of a “tiered storage environment”. However, the decisions to deploy these devices were generally influenced by the same four factors (price, performance, capacity, and function) that define tiered storage today.
Variations of both tape and disk storage coexist in an enterprise data center. Tape includes a wide variety of physical tape providing various capacities and performance in addition to virtual tape. Depending upon the mix of tape technologies deployed within an enterprise, it is likely that at least two or three different tiers of tape storage are currently in use.
Disk on the other hand, presents a slightly different picture in many mainframe environments: Additional disk is deployed in response to business needs for additional capacity and performance. The newer disk systems generally provide both capacity and performance increases at lower cost than previous generations of disk, often making the acquisition of new technology the obvious choice.
When this new disk is deployed, it becomes the first tier (“Tier 1”) disk storage because of the performance and capacity improvements delivered by the new technology. If the older disk systems are to remain in the enterprise, they may become a second tier of disk storage as the storage administrators migrate mission-critical data to the new devices. Over time, the older devices are replaced with newer technology, resulting once again in a single tier providing a homogenous level of performance and functional capabilities.
Addressing IT Concerns
Much of the current interest and discussion surrounding tiered storage is directed at reducing costs: IT management must increase service while reducing cost; storage vendors have responded by offering additional enterprise solutions with new combinations of the same four factors: price, performance, capacity, and function.
In many cases, vendor proposals suggesting a tiered storage implementation make reference to the current homogenous disk environment and suggest that a subset of the enterprise data may not derive benefit from the current Tier-1 performance. Instead, the vendor suggests implementing a second tier of disk storage at a reduced price and level of performance. (Capacity and function could also vary, but most discussions center on performance levels and reduced prices.) This marketing effort naturally focuses on the benefits of obtaining additional storage capacity at a reduced cost with less emphasis on performance and management issues.
This is often unfamiliar territory for support staff being asked to consider implementing “slower/cheaper” disk instead of “better/faster” disk.
Often the storage administrators focus their efforts on the data most critical to the enterprise—spending much of their time utilizing their experience and tools to ensure consistent performance and address the less-critical data on an exception basis.
One of the first management issues that must be addressed when implementing a second tier of storage: The storage administrator must continue to focus on the most critical data to be managed and spend additional effort managing data residing in this second tier.
Some portion of the enterprise data (“backups” as one example) may not benefit from the high performance being provided by the Tier-1 storage. Storage vendors are targeting this subset of data when they speak of implementing a tiered-storage approach. By placing less critical data on less costly storage, the acquisition of the higher-priced enterprise storage may be avoided for some time—perhaps even until next year’s budget.
For the mainframe IT manager considering implementing tiered storage, many of the tools and processes necessary to manage this environment already exist in the enterprise. The risks are minimal, providing the challenges are understood:
- Any additional tiers of storage will require additional management effort by the storage administrators
- The data must be correctly identified before it can be successfully managed
- Data classification is largely a collaborative, manual effort
Once the costs of these challenges are understood, the decision to move forward with a tiered-storage implementation can be made on their financial merits.
Next week we will continue this discussion by focusing on five steps that will help you align your storage strategy with business requirements.