People have predicted the death of tape for a long time. But the facts are that it remains the most cost-effective, safest way to store data for long periods of time. Even when you figure in the cost of the tape libraries or a media vaulting service, it’s still much cheaper than the alternatives. And despite what you may think or hear, tape is actually better at holding on to bits for long periods of time (i.e. > 5 years) than disk.
Besides its data integrity and cost, the biggest thing tape has going for it is portability, which allows for an “air gap” between the data it’s protecting and the actual media. This is the ultimate safeguard against things like ransomware. Yes, tapes are slow to load, and yes, they are quite finicky when it comes to needing to be streamed to work well. But they’re the only game in town for a last line of defense.
I found myself wondering if there was anything on the horizon that would replace tape. It would have to be as fast as tape, which currently is well in excess of 500 MB/s. If you get 2:1 compression, you’re looking at a Gigabyte of data per second. That’s fast. It would also have to support the concept of an air gap, or something really close to it. And, of course, it would have to be cheap.
Sony’s Optical Disk Archive probably comes closest with products currently shipping, but it still falls way behind in a direct comparison. It writes to multiple optical platters inside a cartridge that looks a lot like a tape. Transfer rate to each platter is slow, but since it’s writing to several of them, it can write about 125 MB/s – if you use the write-once 3.3 TB cartridge, which is $146. If you want a multi-writable cartridge, the biggest one they have is 1.2 TB and it’s the same price for ⅓ as much capacity – and the write speed will be slower. Compare this to a 6TB native (14 TB compressed) LTO-7 cartridge that writes at 500+ MB/s and costs only $99. Like I said, it has a long way to go. It does appear to be able to outlive tape, however, in data durability. It is not a magnetic medium, so it will not suffer the magnetic degradation that tapes or disk have.
In the future category, the University of Southampton is developing 5D storage. Researchers there are using nanostructure glass and femtosecond laser writing techniques to store. We haven’t heard much from them in over a year, but the scientists working on them claim this medium will have a life of 13.8 Billion years. We have no data on its reality, cost, or throughput.
Hitachi claims to be able to print a series of dots upon a sliver of quartz glass which can then be read by a microscope. This would also have a very long shelf life and be safe from fire, chemical and even water threats. They say they will be able to store the data on this while still being able to see through it. But, again, we haven’t heard anything about that since it was announced a few years ago.
And, finally, it appears that the replacement for tape is, well, tape. Scientists at IBM have figured out how to store 330 TB of data on a palm-sized cartridge. The areal density of this medium is 20 times that of current tape drives. So, maybe tape will replace tape.
Tape isn’t going anywhere anytime soon. Sure, there are plenty of reasons to not store at least one copy of your reference data on tape, as my colleague George has been writing about recently. But there doesn’t appear to be anything on the near horizon to replace the cost, durability, and portability of tape.