When will we have a storage system that isn't susceptible to failure like HDDs and doesn't die after a certain amount...

When will we have a storage system that isn't susceptible to failure like HDDs and doesn't die after a certain amount of reads/writes like SSDs?

Attached: hdd.webm (1280x720, 796K)

Other urls found in this thread:

raidsonic.de/products/external_cases/external_cases_35/index_en.php?we_objectID=3410
twitter.com/SFWRedditImages

stone tablets arn't susceptible to failure
but it's not exacly efficient

Erosion

Considering OP's short, AIDS infested life, that's not something they need to worry about

Just figure out how much reliability you need and then use Bayes' Theorem to calculate how much redundancy you need from the reliability of the hdds you have.

5D optical data storage or DNA

Attached: 1510184434894.jpg (850x553, 49K)

Modern SSDs are more reliable than HDDs were 5 years ago.
What you should do is probably use ZFS/Btrfs or AppleFS to extend their life.

After all, they are fundamentally different from HDDs and filesystems built for HDDs cannot be expected to work flawlessly with SSDs.

Attached: 1534770157504.jpg (761x1024, 51K)

Tape only allows for a few hundred full writes/reads.

I use a 5,000 watt laser to engrave data on the moon surface. I have not figured out a way to read it back though.

omg stop whining. make your own storage system then.

I'm sure the alien jew will find a way to read it after humanity goes extinct.

Never, it's not physically possible. The SSD quality would improve over time, but that's the best you are going to get. Reading and writing to a media is a physical process and it can't happen without side effects.

I've always been very paranoid about storage ever since the first time my hard drive failed. How are SSDs faring? I'm a poorfag so I don't usually follow SSD progress.

Even the desktop ones are more reliable than hard drives, and with their controllers you know when they are failing - from the user point of view the space on the SSD shrinks as it discards the broken cells. They are also much easier to recover in case of hardware failure. Especially the enterprise ones are great. If you are worried about your data just assemble a RAID array, either RAID1 or RAID5, that would skyrocket your storage reliability

Attached: phoneposting.jpg (1194x1280, 207K)

>They are also much easier to recover in case of hardware failure.
I remember trying to recover my data with partial success, though it was an anxiety inducing nightmare.
WD, never again.

>not knowing how block storage works

>fucking WD
I had a bunch of WD's and all of them died within a year or a little more.

Since i go for Seagate, not a single one of their HDD's died for me.
I would say i found a worthy replacement to my old Samsung Spinpoint 2TB HDD's.

DNA

Attached: DNA_storage_graphic_WEB_2.jpg (630x403, 83K)

If I decide to stick with hard drives I'm going for Toshiba. I have a Toshiba hard drive that's god knows how many years old and that one is still going flawlessly. They're all cheaper than the competition too, unlike WD.

>HDDs are susceptible to failures

OP HE HIT THAT HDD WITH A FUCKING HAMMER AND DESTROYED THE PLATTERS DIPSHIT

>DNA data retention
ahahahahahahahah

>implying the webm is related to his point

I totally believe you. I never went for Toshiba disks the last years, but those i had held up pretty well.
Some of my Samsungs are from fucking 2002 and still working perfectly.

HDD brand doesn't mean anything really. The failure rates are so low between brands that it's not worth debating.

Data written to DNA can last 2,000 years. The entire world's data can fit inside an office room user.

I created an archival file system a few years ago that involved writing 1s and 0s on stone tablets. It’s lifespan was essentially infinite, but the I/O ops were painfully slow.

>Since i go for Seagate, not a single one of their HDD's died for me.
Contrarily I've had 2x 3TB Seagates die within weeks from each other. It's a coin toss.

What I do now is buy cheap, refurbished HDD's for $50 and throw them in my RAID. If one of them dies or is failing, then it takes less than 2 days to swap it out. No data loss.

So far surprisingly none of my cheap refurbished drives have died yet and they've been running nonstop for 3 years.

That jpg is so amazingly wrong it's physically painful.

*family pic mutates to cp*

that's not how data or dna works
consider researching the topic

>Contrarily I've had 2x 3TB Seagates die within weeks from each other. It's a coin toss.
yeay, they were horrible. Especially the 3TB ones died like flies.

>What I do now is buy cheap, refurbished HDD's for $50 and throw them in my RAID.
How big are they in size if you buy them for 50 USD?
I entirely go for 8TB Archive disks from Seagate. but haven't bought a single refurbished one yet.
And i actually never looked up prices on refurbished ones too.

>How big are they in size if you buy them for 50 USD?
3TB

>I entirely go for 8TB Archive disks
My file server has a SAS2 backplane which can hold 24x HDD's so I prefer smaller quantities than large chunks. They're cheaper to replace and resilver times are faster.

1x 8TB = $230 x 5 = $1150 for 40TB
1x 3TB = $50 x 14 = $700 for 42TB

I'm also saving $450. I'll move on to larger chunks once the price difference is negligible.

I pay 180 Euros for 8TB ones. They come in USB enclosures and i simply rip them out of the case.

I could not hold the sheer amount of data in smaller disks than that. At least not in a convenient way. So i rather stick with larger ones.
I go for HDD docking stations instead of RAID arrays.

> pay 180 Euros for 8TB ones.
I'd still be saving $335 which could buy me an extra 6x 3TB's.

>I could not hold the sheer amount of data in smaller disks than that
You can get used Supermicro servers for cheap. It has the added benefit that you can use a proper file system for large storage such as ZFS too.

HDD docking stations seem like a poor long term solution for large storage capacities.

Attached: 16-101-796-02.jpg (640x480, 24K)

I would need to run all HDD's all the time with such a rack. The power consumption would be too high for the little i use the stuff.
I am not even running a network here. Just use it on my own computer. And for that, docking stations are the optimal solution. Just one disk running, while the others are off.
Have ran like this since years and always fell back from RAID arrays to my old style USB docking stations because of convenience.

Server solution is overkill for my needs. Also, such systems cost a ton of money too. So i guess i am running cheaper with my way.

A record (like a vinyl) with an optical reading head

tape but its access times are too slow.


modern tape will never rott as well. made of better chemicals and will last like 500years.

And you lose all your data if you hard drive gets an infection. Or it gets irradiated or whatever. DNA data storage could also potentially give you ACTUAL viruses. Shit man, ebola's only like a couple kb, same with smallpox. Well maybe not directly, but widespread DNA data storage would certainly enable people to bring back smallpox or make ebola

youre joking right

Make a backup plan who cares about failure

Tape backups are cheap and come in large capities
External drives are cheap
Some cloud storage isn't bad

Snap is less than 150w iirc
Different poster but its worth mentioning

>proteinic substrate gets a bit dehydrated
*storage dies*

Why the fuck do you idiots assume that the DNA would be stored in living cells? That's like the worst way of doing it.

I have several drives that are 15 years old and they still work 24/7.

I just have three docking stations where i swap out the HDD's after my needs. They take about 15 watts HDD and dock combined.
And these docks are dirt cheap also.

I am opting for a ten bay HDD USB dock:
raidsonic.de/products/external_cases/external_cases_35/index_en.php?we_objectID=3410

Would perfectly fit my needs.

>The entire world's data can fit inside an office room
>A small power surge happens in the office room
*entire world's data disappears*

How about seek/access times? random IOPS? I get that it takes very little time to actually change whatever chemical property is used to store data but how about locating and changing that specific protein/atom/whatever?

not shown here: shards sticking out of that guys arm and face