Tape storage an asset for high leveragability

Computer Technology Review, Feb, 2004 by Richard Harada

One of the primary benefits of today's storage networks--whether they are based on Fibre Channel, iSCSI, InfiniBand, or any other architecture--is the ability to attain better utilization from storage resources. The basic idea is that all of the available disk storage can be pooled together into one big repository and then shared by all users and applications. This is also known as disk virtualization. Without virtualization, applications can only use the storage that is directly attached to their particular server(s), and when they need more capacity it must be physically purchased and added--even though there may be other servers with an excess of available capacity. The net result of storage networking and disk virtualization is a large reduction in the costs of storage hardware and administration.

This storage-networking paradigm can be applied to automated tape libraries. Today's tape products are highly reliable and extremely cost-effective when compared to other storage technologies. But in most organizations, these important resources are limited to single applications, much the same way that server-attached disks are used in a non-networked environment. Tape systems are more flexible and scalable, and can be leveraged across any number of storage applications. For example:

Backup and Restore

Tape is the most ideal storage technology for making backup copies of important data assets, and this has been true since computers first hit the corporate world more than 40 years ago. Due to its high capacity, high data transfer speed, and low cost, tape technology is ideal for backup and restore applications. And, since tapes are also reusable, the costs are considerably lower.

"Tape has shown a better return on investment than disk-based systems," explains Jeff Ash, Fujifilm's vice president of marketing. "Companies who evaluate disk-based backup have found its management to be complex, and adding capacity to be more difficult and costly than their tape storage solution."

Mirroring of storage assets is also a popular method for providing fast recovery following a failure of the primary storage hardware. But, keep in mind, mirroring does not protect against data loss--if a file is deleted from the primary disk it will also be immediately deleted from the mirror. While mirroring can augment data protection, it does not replace backup.

Disaster Recovery

Another longstanding application for tape has been disaster recovery. Tape allows a large amount of data to be copied and stored on inexpensive, removable cartridges that can be moved to a safe, off-site location. These tapes can be used in the unlikely but unfortunately possible event that an organization's primary data center is damaged or destroyed. Disaster recovery should be a crucial component of a company's business continuance plan.

And on-line systems designed to handle disaster recovery have not proven to be an economical and secure alternative to tape storage. At first glance, transmitting data to a remote site in the background might appear financially feasible. However, the cost of managing and storing data is higher than the cost of storing a tape at an off-site location.

"Ultimately, the company offering this service who truly cares about the client's data will store data onto tape to ensure a reliable backup copy of the data," adds Ash.

Information Lifecycle Management

Information Lifecycle Management (ILM) is an evolution of the Hierarchical Storage Management (HSM) model. ILM promises to automatically store data in the most appropriate place, at any given time in the data's lifetime. Like HSM. ILM is built on policies that define the routing of data. For example, the movement of data and its current location is linked to its original location, making it transparent to users and applications. The difference between them is that HSM is linear: it moves data from primary storage to secondary and perhaps tertiary--storage at given data age points. ILM can move data anywhere, at any time, based on any number of parameters that can be application-specific.

In both HSM and ILM, the idea is to keep less-valuable information on the least-expensive storage repository that meets the retention and performance requirements for that data type at that time in its life. Automated tape-based systems are excellent repositories for storing these less frequently accessed data files.

"As a business grows, the storage requirements for archived data will expand far faster than the needs for new data," says Paul Scheuer, IBM's strategy manager for tape storage. "Tape media is well-suited, since hardware changes are not required other than to add a new media cartridge when there is a need to expand capacity."

Long-Term Archival

The difference between archival and ILM is that the archive repository is a discrete resource that applications and users access directly to store and access information that is in its final form and is not actively used. The benefit to ILM is that data can be both archived and actively accessed.