Apparently WD and Seagate stock have already taken a hit at the announcement.
I wonder what the servo writers for these drives look like.
These numbers feel unreal. How long does zeroing a 65TB drive take?
would probably take like 2 to 3 days to write zeros across the entire disk. but hardware or software based full disk encryption can make ‘erasing’ a disk (make the data irretrievable) very fast.
How computing a result from a ?4096?bits key for each byte goes faster than writing 0 for each byte?
You keep the drive encrypted while you are using it, and then when you want to wipe the data, you just wipe the part of the drive holding data related to computing the encryption key. You don’t touch the rest of the drive, but without the encryption key it’s effectively random noise.
It’s slower to encrypt and then delete the keys than it is to zero the drive, yes. But that’s not how it’s supposed to work.
You encrypt your data (based on the pre-generated key) piecemeal as you fill the drive over the course of years, then you delete the key when retiring the drive, so you can make a drive unreadable in seconds.
If the disk has built in encryption, so the storage controller on the disk handles the encrypt/decrypt cycle but presents as an unencrypted disk for all intents and purposes to the user. All thats needed to wipe the disk is to delete the encryption keys. Its a pretty common feature in enterprise disks and i quite like it. Saves a lot of time when decom time rolls arround
All thats needed to wipe the disk is to delete
You are jumping to a point in time after what I am asking:
“all you need” is to delete a key -> No, you need to encrypt the disk before, which is the step prior, and without it, deleting a key change a thing…
So basically, the part “hardware” built-in could make sense, but the guy didn’t specify any if this…
In other words, zeroing the disk is going to be faster “in normal circumstances”.
You don’t need to encrypt empty disk space. Formatting a disk only writes some data structures here and there, leaving vast majority of the disk unchanged.
The whole problem and why people are zeroing is that “empty space” is not really empty.
I don’t know how you manage to keep being confused about this whole deal. When you buy a disk, you don’t care what’s in the empty space. You format it as an encrypted disk by writing a few sectors. When you want to throw it out, you delete the keys that decrypt what you wrote over the life of the disk, and perhaps wipe the sectors containing encryption metadata. What is unclear about this?
Besides, when the disk is configured as encrypted, the junk in the empty sectors can’t be read by decrypting it. Unless it’s software encryption and you bypass it by poking the disk directly, you can’t read anything from those sectors. Which, again, contain either noise, or what the possible previous owner has written there.
The key is in firmware which is setup before you write any data. So to “delete” any data you simply wipe the key.
Even if you use software to encrypt the drive luks would just mean any data written to the drive would be encrypted. Old data my zeroed out could technically be ready unless overwritten.
In either case your weirdly aggressive statement is wrong.
Nice. I switched from WD to Toshiba in 2024
Why?
Not as scummy marketing (hiding SMR among CMR)
Not whom you asked, but in my experience and from what I’ve read on the webs, WD’s reliability is a coin toss. Some drives keep chugging for years and years, other ones croak too soonish.
Trump kept threatening Europe, so I stopped buying US products
I just want affordable storage again in the single digit TB drives, maybe low teens. I waited too long to build a multi drive nas and now I have regerts.
When my 4 TB drive died ~15 years ago I thought to myself I’d just wait a few months and replace it with 8 TB for the same price. Then the floods happened.
Then prices almost were finally at that level again. And then AI started gobbling everything up.
IIRC storage started rapidly rising even before AI, due to those stupid proof-of-storage cryptos.
Yeah, 2TB SSDs are honestly fine for everything except bulk storage.
From my perspective the older slower ssds are fine for bulk storage, you just can’t get them. My ideal would be a 4-8TB SSD that is as cheap as an equivalent SSD, as marginally faster than a HDD
More than marginally, unless HDDs have gotten a LOT better in the last five years, hahah.
Good that Toshiba saw the increase in demand as an opportunity to increase sales volume, rather than price gouge their customers. Hopefully they gain market share.
sales volume
There’s more competition in the storage market than RAM. Someone was going to do it, might as well be them.
RAM is controlled by 3 major companies. Plus Ai companies need as much RAM as possible, unlike HDD.
IDK, it’s pretty risky. Doubling production means youre taking a bet that the bubble is not going to burst.
Cory Doctorow is making the argument that the reason RAM prices are high is the producers aren’t building new fabs for the data centre demand because those are expensive and hence strategic long term investments. If AI is a bubble they would have a lot of product and no one to sell it to, cratering the market.
Prices are so up that even when it bursts, they will go to normal ones. I mean price per GB was normal to go down, and now for 1st time in history it went up.
Same and worse for RAM.
Its like we are 15 years back for RAM and like 10 years back for HDDs.
Great point! Sextuple production!
Hard disks will always have demand, we need more every day and they have a limited lifespan
Great point, quintuple production!
Ok, give me the money
??? I thought it was a risk free investment! You got this!
It may be risk-free, but you still need money to spend money. That’s what the asshats called ‘investors’ are for.
Should be an easy sell, someone on the internet found an infinite demand glitch!
I’m pretty behind what I want for drives, and modern dev practices of storing tons of intermediate container build artifacts, and a resurgence of arrrborists will probably soak up drive demand even if prices fall.
Great point! Septuple production!
That is a fair point, especially with things that are very hard to set production for like processors.
Why not both ™ ?
Toshiba are the real ones.
The good, more HDDs and bigger drives.
The bad, they will directly to data centers.Straight to the datacenters. And once they get their five years out of them, straight into to the shredder.
Wouldn’t want any normies getting their hands on them, and having ideas of not using whatever subscription cloud computer bullshit they’ve forced on us by then.
Consumers aren’t buying large drives like this, those are intended for data centers. And if they’re increasing production aimed at those capacities, that means smaller capacity stock opening up for consumers.
I hope it alleviates the pressure. I need to upgrade the 4TB in my array, badly. It’s all on borrowed time.
I have 4x8tb drives that have way too many hours on them
I don’t want to look. I think mine might be getting towards 10 years old.
I don’t hoard video files, so 1TB still seems unimaginably vast. I was shocked when I learned my entire CD collection, ripped to FLAC, would fit on my phone. 100TB is like a billion dollars, my mind has difficulty understanding numbers that large.
You get games that are several hundred gigabytes these days - high resolution art assets are big.
Filling up a 1TB drive is easy. 100TB, yeah, that’s a lot of space.
Don’t count bytes, look at what they can store: 1TB will hold like four to eight recent AAA games.
Even if you do need 100 TB of storage, I feel like it’d be a lot better to have multiple smaller drives rather than one massive one. To be fair, though, I’d expect such large drives to be targeted mostly to commercial customers.
My 116TB pirate ship could use a few more 100TBs to keep afloat.
I’m in a smaller, but similar boat with my 40TB. A few of the drives are pretty old now, and i would prefer swapping them out before they die.
Large drives are convenient for 3 2 1 backup. Not all your storage needs or even should be ssd or raid 5.
Yeah, if get a few bad sectors suddenly you’re spending most of a day backing things up.
You should’ve been backing things up already.
“If you have one, you have none”.
Without proper backup your data is just waiting to disappear at the first unrecoverable failure.
Backing up with an increased frequency to mitigate the lost time.
And it can be fairly invisible in day-day use with 4 drives set up in RAID5.
Really depends on what you store. If you archive videogames like i do, 100TB is a good place to be, and redundancy means you need more than one.
Larger storage is good for videos (think Blu-rays, 4K videos are very big! Security camera feeds could also take up lots of space), and you might also want to host for multiple users (for family, friends, or a business)
But yes, _100_TB is a lot. 8-24TB (before memory became super expensive) was reasonable for most consumers, usually in a RAID configuration for redundancy. Multiple smaller drives are more reliable and modular than a single big one!
Without going into video media, I’ve got around 2-3TB of photos (in RAW format) and my ripped/downloaded music collection is about the same size. Probably got about 5TB of device backups and a similar amount of retro games ROMs/disk images
I think between my desktop PC and server I’ve got about 70-80TB of storage (though not all of it is usable for redundancy)
I could see myself building a new NAS in a couple of years that would make use of a few of those drives if the economics work out
I’ve had good reliability from Toshiba in the past few years. I hope they can keep it up.
