The 2012 Tape Summit was held in San Francisco on March 6 and 7, organized by Greg Duplessie's ExecEvent. The refrains of "tape sucks" and "tape is dead" have been blaring for years, as disk-based technologies have taken over many of the tape functions, such as backup and restore. Statistics like "Gartner says 71% of tape restores fail", or "Yankee Group affirms that 41% of tape restores fail" have been quoted by every disk and de-duplication salesman for years without challenge. The refrain has been loudest from companies without tape solutions.
It seems that a turnaround in the relevance and viability of tape is imminent. The mood at the Tape Summit from the tape specialists was cautiously optimistic. Curtis Preston has been systematically debunking the tape restore figures. Both numbers are fictitious and have been disowned by Gartner and the Yankee Group.
The major tape vendors (HP, IBM, Oracle, Quantum, Spectra Logic, and Tandberg), presented the business and technical case that archive and disaster recovery solutions are better suited to tape or disk-tape hybrids, rather than disk alone. The key technology enablers of this shift in optimism have been:
- The rapid adoption of LTO-5, the latest tape standard from the Linear Tape-Open consortium announced in January 2010;
- The product introduction of the partition feature of LTO-5, and the adoption of the LTFS (Linear Tape File System), which allows tape media to be used in the same way as other removable media (USB flash drive, external hard disk drive, etc.);
- The improvement in tape longevity (15-30 years) and media reliability (100 times greater than disk).
The result of LTO-5 and LTFS is that tape libraries can be designed to be ideally suited for media applications in general, and video (including surveillance) and imaging applications in particular. As video is projected to grow to be 50% of the unstructured storage demand, this increases the potential market for tape-based solutions. With a load time of 60-90 seconds for a file (e.g. a video file), a disk/tape solution with the indexes and buffering in high performance storage, the first 60-90 seconds of video on high density disk and the remainder on tape would produce a very cost-effective solution. The costs would be decreased even more by adding a minute of commercials in front of the video demand.
The potential rise in tape looks to be accelerated by the projected rise of flash-based technologies as the major medium for data that is accessed frequently, and the projected collapse of investment in high-performance disk-based technologies. Figure 1 shows that the disk market as being squeezed between the rise of flash and tape.
Action Item: Storage based on moving magnetic media has been the underpinning of IT for the last half-century. The rise of flash as the medium of choice for active data and introduction of highly available indexed tape solutions means that CTOs and CIOs should revisit the cost benefit analysis of automated tape solutions, especially for media, health-care imaging, and surveillance applications. There is also a strong case for automated tape-based disaster recovery systems.