Tape Is Not Dead, But Dying …

Many people are laughing because they hear since many years that tape is dead, which is not true. Tape is not dead but is dying.

Look at the figures
Below are the figures published by the Santa Clara Consulting Group, the last company analyzing regularly this market (but for how long?):

 
WW Market of Tape Cartridges

 (Source: Back-Up Tape Tracker, Santa Clara Consulting Group)

The SCCG does not calculate the figures for mainframe tapes from IBM and Oracle/Sun but there is no indication that their market is growing, also attacked by VTLs using HDDs, including subsystems recently announced by EMC, but also other ones from B&L Associates with Dynamic Solutions International, Bull, Cybernetics, Fujitsu, Luminex and others.

Look at the vendors
Dell, HP, IBM and Oracle do not revealed any figures on their tape activity. Quantum, Tandberg Data, Overland, etc, all of them are struggling and trying to diversify in disk subsystems.

Media manufacturers like Fujifilm, Maxell, Sony and TDK do not seem to be really happy producing tape cartridges.

In this decreasing niche market only a small firm in the world seems to be successful, Spectra Logic, with record revenues and profitability for FY finishing June 30, 2011. On its side, tiny company Qualstar, also in big tape libraries, finished its fiscal year ending June 30 with $18 million in revenues – including its bigger power supply activity – and $0.7 million net loss.

Look at the applications
Tape are not used anymore on PCs and only on a small number of servers as backup is transitioning completely from optical discs and tapes to SATA hard disk drives with huge capacity and decreasing prices that the competing media cannot follow.

Huge archiving is the last enjoyable market for tapes. At around $60 for native capacity of 1.5TB, one LTO-5 cartridge is less expansive than an hard disk drive of the same capacity. You have to add the pricey tape drives and library, but, for huge capacity, 1PB and more, the final cost is in favor of tape. Read here two studies revealed by LTO consortium demonstrating that tape has significant TCO benefits over disk.

Look at the technology
Theoretically, you can put more data on a magnetic tape with a total surface much larger than the surface occupied by the maximum of four platters inside a HDD. But areal density reached on magnetic disk is much higher as read/write disk heads are more precise than tape heads.

Furthermore, it has to be proved than tape is more reliable than disk with platters enclosed in hermetically sealed volumes. Tape cartridges and their drives are more vulnerable to the environment.

LTO represents 88% of the tape market – without mainframe tapes – with sales of $163 million in 2Q11, according to the SCCG. All the other formats are progressively disappearing: DDS/DAT, DLT, 8mm and QIC. LTO is the only technology evolving with capacity doubling each two to three years. Current LTO-5 is at native 1.5TB and forthcoming LTO-6 will be at 3.2TB, a little more than today’s highest capacity of an HDD (3TB). The roadmap is 6.4TB for LTO-7 and 12.8TB for LTO-8. For mainframes there are currently the IBM 3599 at native 4TB and the Oracle/Sun T10000C at 5TB, being today a record capacity for tape media.