Insider Insights: z/Journal Interviews Storage Specialist Fred Moore

by Bill Carico
December 1, 2005

z/Journal: Virtual tape for open systems is hot now. What does that mean for the market?

FM: The movement of the virtual tape solution into the non-mainframe markets, Unix, Windows, and Linux, has been fast-paced in 2005. At present, there are 26 companies offering a variety of virtual tape solutions. No clear leader has emerged in the non-mainframe market. The major benefits for virtual tape are obvious and appealing and include significantly increased tape-cartridge utilization, higher performance (as data can most frequently be accessed or recovered from the disk arrays), defining more tape drives than physically exist to improve backup performance, and a financial savings, resulting from the reduction in the number of tape drives and media. Virtual tape also is hot, since making a disk array appear as several tape drives allows disks to be implemented quickly as a backup solution without changing existing backup/recovery procedures that are written to use tape for backup. True integrated virtual tape libraries enable direct data movement between Tier 2 and Tier 3 storage systems, avoiding the overhead server I/O tax.

z/Journal: Over the next five years, what are your hot prospects for new/successful/innovative storage architectures?

FM: New technologies rarely change the world overnight. It’s hard to predict when a new technology will reach the inflection point that signals success. I think the storage industry is examining and digesting several new offerings that could make a notable difference. Several new and beneficial offerings are entering the storage industry. MAID (Massive Arrays of Idle Disks) libraries consisting of economy SATA disks offer the lowest price point for any disk subsystem, bringing the TCO of disk close to tape libraries, while potentially shrinking the time it takes to perform data recovery. Expect magnetic storage to dominate the storage landscape for the next decade, as no other significant alternatives are imminent.

Ultimately, I expect the backup/recovery application will give way, favoring keeping additional data copies in executable formats. Why recover data at all?

Intelligent switches have been discussed since the early ’90s and now appear ready for introduction. These switches enable a more centralized and automatic approach to storage management. Device-to-device data movement offloads the I/O traffic from the servers and allows the storage subsystem to perform data transfer functions. Several SAN and fabric-based solutions are becoming available but none has been widely implemented, as customers remain cautious.

The requirement or desire to keep all data forever now looms. A new storage management discipline that establishes lifetime policy management capabilities is unfolding. Archival science will involve media life, devices, security, and data migration capability to enable infinite archive data preservation for tens, if not hundreds, of years. Security has become the newest storage management discipline and is quickly driving development of security appliances, surveillance technologies, and encryption capabilities for data.