Storage management is a difficult business. Almost all applications and users have an insatiable need for data storage. Exponential data growth coupled with storage prices (as little as $.08 per megabyte) result in a mound of company storage—a management problem that's like an ever-growing hill that IT administrators must climb daily. What can companies do about the data storage problem?
Many situations call for adding additional Direct Attached Storage (DAS), which works until the computer system reaches the server's supported capacity. At that point, IT administrators must decide whether to implement additional servers to support the environment or to implement a Storage Area Network (SAN), which is a more costly, short-term decision in terms of capital investment and personnel costs. Although SANs and Network Attached Storage (NAS) have shown their ability to lower a company's storage total cost of ownership (TCO) over time, much of their value comes from how well the IT administrator can manage these technologies.
Fortunately, new developments in hardware and software promise to make IT administrators' lives easier—developments such as integrated management of application snapshots, remote mirroring and replication, and other wide-area, data-storage management technologies promise to make the IT administrators life easier. Even as these technologies evolve, however, well-known existing storage-management functions such as policy-based archiving, backup, and Hierarchical Storage Management (HSM) continue to provide tremendous value in lowering storage TCO. Implementing software to address storage management is only half the battle, however. To successfully manage a well-run storage infrastructure, IT administrators must take time to develop carefully thought-out policies for the functions they decide to implement.
Archiving is a process of identifying data that no one has accessed for a certain period of time and moving that data to secondary, long-term, inexpensive-media storage. An archive is a set of data that a company needs to retain but probably won't need to access often, if at all. A company generally archives data for historical reporting or legal purposes. Also, a company sometimes needs to be able to recover an environment—in this case, an archive serves the same purpose as a backup.
The process of data backup is similar to archiving, but has an altogether different purpose. IT departments back up data for the sole purpose of potential recovery. Backed-up data is an insurance policy that provides the ability to restore and recover anything—from individual files that a user erroneously deleted to the recreation of a business environment in the event of a disaster.
HSM is the process of recognizing the frequency of access of any given set of data and moving that data to the most appropriate media (in terms of cost, performance, and accessibility) for the lowest storage TCO. HSM also provides transparent access to data, regardless of where that data resides—one doesn't need to initiate a formal recovery of required data.
Sophisticated software tools can perform all these storage functions, but the key to successful storage management lies in IT administrators' understanding of how to build policies to use these tools together. IT administrators develop an archiving policy based on the requirements of corporate applications, and archiving policies generally are less flexible than those for backup and HSM. Periodically, a company has a special need to create a copy of data that it puts in long-term storage. The key to managing archive functions successfully is to be sure to complete the process when there's little activity on the systems and to use resources (tape, optical, CPU) effectively.
IT administrators should think through HSM policies according to the application's need for data access. Based on frequency of access and latency requirements, IT administrators can implement appropriate secondary-storage technologies (e.g., optical, tape) to cost-effectively add storage capacity. In an optimal configuration, HSM integrates with backup so that the system doesn't migrate data to secondary storage until there is at least one backup copy and so that the backup of migrated data doesn't cause retrieval of the data from secondary storage. When a company understands how to use storage applications effectively, IT administrators can build policies that let HSM act as a powerful tool to decrease the amount of data that they must back up. IT administrators should interweave archiving policy with HSM and backup policies so that the system moves data to long-term storage only if necessary and only after that data has passed a certain threshold of potential need for either transparent access or recovery.
In the optimal IT environment, administrators fully understand the functionality they provide with the archive, backup, and HSM functions, and IT administrators implement integrated policies for these tools based on their corporate applications' needs for data.