"Still" use tape for backup? Interesting reading about tape usage in 2020 and in future.


Userlevel 7
Badge +5
  • Veeam Legend, Veeam Vanguard
  • 499 comments

The Tape Storage Council has released summary of 2020’s usage of tapes and a future outlook:

https://blocksandfiles.com/2020/12/07/tape-storage-council-2020-outlook/

Summary:

  • Capacity shipments rose to record amounts in 2019
    • >225 million LTO cartridges
    • >4.4 million drives
    • [IBM and Oracle enterprise tape shipments are not included]
  • Tapes are still cheaper than disk
  • Several public cloud deep archives use tape, such as AWS’s Glacier
  • ESG found a majority increasing their commitment to tape (61%)
  • Tape’s two big advantages:
    • low cost
    • longevity
  • Big drawback:
    • lengthy file access time 

40 comments

Userlevel 6
Badge +2

backup tapes are widely used to make a sort of airgap backup and to fulfill the 3-2-1 backup offsite for the win rules.

Userlevel 7
Badge +5

backup tapes are widely used to make a sort of airgap backup and to fulfill the 3-2-1 backup offsite for the win rules.

Fully agree! Therefore I used the quotes (“) to show the irony of the sentence. Tapes are still extreme widely used and this report shows that this will even increase in the future! A lot of my customers use and will use tapes. Most of them prefer investing in tapes than in cloud-backup.

Userlevel 6
Badge +2

backup tapes are widely used to make a sort of airgap backup and to fulfill the 3-2-1 backup offsite for the win rules.

Fully agree! Therefore I used the quotes (“) to show the irony of the sentence. Tapes are still extreme widely used and this report shows that this will even increase in the future! A lot of my customers use and will use tapes. Most of them prefer investing in tapes than in cloud-backup.

true story friend, imho let's say the best setup for data protection is:

- primary storage backup 7 -10 days

- storage tier cloud medium term backup with data immutability enabled

- Tape for long long retention alternatively

I had designed but I have to test it, storage snapshot primary backup and storage sync volume on a secondary storage (similar to an airgap backup) and subsequent vaulting on tape

Userlevel 7
Badge +4

Tape is a super solution for long retentions.

Some of my clients want to keep yearly backups for several years, so tape is a easy solution to achieve this.

I hope VEEAM keeps up its efforts to improve tape support in their products. And yes, I have seen that there are several improvements in V11 :grin:

Userlevel 7
Badge +5

Tape is a super solution for long retentions.

Some of my clients want to keep yearly backups for several years, so tape is a easy solution to achieve this.

I hope VEEAM keeps up its efforts to improve tape support in their products. And yes, I have seen that there are several improvements in V11 :grin:

Yes, there are a lot of new tape-features in v11, other enterprise-backup solutions had for years. Gap is getting smaller and smaller! :thumbsup_tone4:

Userlevel 7
Badge +3

For us tape is still the number one solution for archiving and offline backups.

I'm wondering if that will change in the near future with the rise of immutable and hardened disk-based backup solutions.

Userlevel 7
Badge +4

@regnor , no, I think for offline - or even offsite - backups we will have tape for a long time… You cannot carry your disk based backups offsite - at least not easy and without problems…

And it is cheap - after you have bought the library and some tape, then addng tapes costs not much...

Userlevel 7
Badge +3

That's true; regularly moving disked based backup offsite could be complicated (at least physically :grin: )

Like @Link State has written, immutable disk based backups will probably complement a backup solution additionally to tape.

Userlevel 7
Badge +4

I think tape is great, I think the stigma it suffers is due to misuse like deduplication appliances. The issue isn’t in the technology but it’s deployment. I don’t believe tape should be used as an air gap solution (or not the primary air gap solution at least), for a single reason, it’s so slow!

 

I was dispatched to a school trust due to a ransomware attack, I discovered due to the attack vector and the entire backup solution being domain joined the primary backup repo had been encrypted, the customer then proceeded to recover from tape but it took between 1-2 weeks for all of their data to be recovered and brought online. It was nearly a week before core services were even restored.

 

With solutions such as Veeam Cloud Connect it makes sense to use VCC for air gap and keep tape for long term archiving purposes only. Most customers will keep their tape systems far longer than necessary and this causes them larger issues in the long run when using this as their DR response.

 

With v11 pushing immutability this will provide a better fault domain for customers not wishing to go to the cloud route.

Userlevel 6
Badge +2

With v11 pushing immutability this will provide a better fault domain for customers not wishing to go to the cloud route.

i love it !!!!

ultimate solution against crypto ransomware

Userlevel 7
Badge +3

Some people miss this factor, more tape generation evolved more they are sensitive to humidity. It’s more than a typical disk.

Don’t miss this because some Datacenter don’t measure this. It’s crucial to don’t have a big surprise

Userlevel 7
Badge +3

With v11 pushing immutability this will provide a better fault domain for customers not wishing to go to the cloud route.

i love it !!!!

ultimate solution against crypto ransomware


I hope people won’t be too greedy with immutability. It’s really concrete when the linux host who host XFS is enough hardened or you will always exposed to side attack (voluntary degradation)

Userlevel 7
Badge +5

Some people miss this factor, more tape generation evolved more they are sensitive to humidity. It’s more than a typical disk.

Don’t miss this because some Datacenter don’t measure this. It’s crucial to don’t have a big surprise

Yes, you should schedule tape-to-tape copy after some years.

Userlevel 7
Badge +4

IBM and Fujifilm have set yet another new world record in tape storage – the sixth since 2006. Pushing the limits, they achieved 317 GB/in2 (gigabits per square inch) in areal density on a prototype strontium ferrite (SrFe).

https://www.ibm.com/blogs/research/2020/12/tape-density-record/

 

So there will be a lot of development and increase in capacity in the future...

Userlevel 7
Badge +4

IBM and Fujifilm have set yet another new world record in tape storage – the sixth since 2006. Pushing the limits, they achieved 317 GB/in2 (gigabits per square inch) in areal density on a prototype strontium ferrite (SrFe).

https://www.ibm.com/blogs/research/2020/12/tape-density-record/

 

So there will be a lot of development and increase in capacity in the future...

It’s amazing the density increases we’ve achieved by sticking to specific form factors such as tape, 3.5” HDD and microSD. Comparing where they started to now and the lack of any signs of slowing down is a testimony to what the human race can do!

Userlevel 7
Badge +5

Yes, density increases very well! I hope throughput will increase even faster. Years ago I can remember we had some challenges to keep tapes streaming, now we have problems to put data fast enough to tape at maximum speed.

Userlevel 7
Badge +5

Could be interesting as well: What is the physical speed of tape:

  • LTO-6 @ 6.83 m/sec when reading or writing
  • LTO-7 @ 5.01 m/sec 
  • LTO-8 @ 4.731 m/sec

Interestingly, speed is decreasing from gen to gen.

https://blocksandfiles.com/2021/01/04/incremental-storage-media-density-magic-meant-flash-chipped-away-at-disk-in-2020/

Userlevel 7
Badge +4

Yes, but the data density increases from generation to generation and “compensates” this… :grin:

And… the tape gets thinner and thinner in each generation, so it makes sense to decrease the strain for the  tape.

Userlevel 7
Badge +4

Yes, but the data density increases from generation to generation and “compensates” this… :grin:

And… the tape gets thinner and thinner in each generation, so it makes sense to decrease the strain for the  tape.

I wonder if we’ll reach a point where we have deliberately blank borders on the tape to increase the material strength. Like how we saw iPhones getting thinner each generation, then we had the “Bend-Gate” drama and now iPhones appear to have stopped becoming thinner and in some cases slightly thicker again.

Userlevel 7
Badge +5

Yes, but the data density increases from generation to generation and “compensates” this… :grin:

And… the tape gets thinner and thinner in each generation, so it makes sense to decrease the strain for the  tape.

When the speed decreases and the density increases from gen to gen, we will have - at a far away future - a tape backup at a point. This will be the backup-data-black-whole :joy:

Userlevel 7
Badge +4

Yes, but the data density increases from generation to generation and “compensates” this… :grin:

And… the tape gets thinner and thinner in each generation, so it makes sense to decrease the strain for the  tape.

When the speed decreases and the density increases from gen to gen, we will have - at a far away future - a tape backup at a point. This will be the backup-data-black-whole :joy:


A tape singularity…  :grin: 

Perfect, this is the solution for all ransomware problems, you can go back in time through the tape singularity and fix it...

Userlevel 7
Badge +5

Yes, but the data density increases from generation to generation and “compensates” this… :grin:

And… the tape gets thinner and thinner in each generation, so it makes sense to decrease the strain for the  tape.

When the speed decreases and the density increases from gen to gen, we will have - at a far away future - a tape backup at a point. This will be the backup-data-black-whole :joy:


A tape singularity…  :grin: 

Perfect, this is the solution for all ransomware problems, you can go back in time through the tape singularity and fix it...

So a tape will look like:

LTO Gen:infinity

:smile: It will be available at the end of time.

Userlevel 7
Badge +3

We are in the process right now of testing a few tape solutions that front end to S3 and then tape out to a library.  We do have a solution but reliability has been an issue so I am doing PoCs for new ones to replace this.  We also integrate this with our in-house tool to use 4 factor authentication.

Userlevel 7
Badge +4

I have looked at such a solution just this week. I think it's very interesting.

Just collecting resources for a test 😁

Userlevel 7
Badge +3

tape is still the more secure way vs side attack if it out the tape library in a strong box in a faraday cage :japanese_ogre:

Comment