back
107 comments
Give a man a fish and you feed him for a day; teach him to fish and you feed him for a lifetime. Ban fish and the man gets into farming and animal husbandry and doesn't need you anymore.
The analogy needs something about IP theft of automated fishing pole designs
This misses the point. If you gave all of the F-35 IP to India they would still be incapable of developing the indigenous industries required to build a 5th generation fighter in a similar amount of time.
"Teach a man to fish, he eats for a lifetime. Allow the man to steal your F-35 documents and you must communicate that teaching time is over."
I always don’t understand this argument? By this logic, should every companies publicise their trade secrets and redact all the copyright and patents to maintain a dominant market position?
> Give a man a fish and you feed him for a day; teach him to fish and you feed him for a lifetime. Ban fish and the man gets into farming and animal husbandry and doesn't need you anymore.

Eh, you're misusing the saying in an inappropriate way. It's about fostering self-sufficiency in people you want to help. It's not really applicable to rivals and enemies.

In this case, the hubris of the 1990s cause the US to get into this relationship with China, and now that's going to have to unwind somehow (and doing that in a 5-D chess way is unrealistic). And anyway, it's not like China's goal was to remain dependent on Western tech forever, it's plan is to become self-sufficient on its own terms on its own timetable, and the best the US can hope to achieve is cause some disruption and mess that timetable up.

Up to 2 minutes seek time for cold data. Very niche.
It’s niche enough to be a major industry. Cold storage at extreme scales for very low cost is higher value than it appears - there’s a lot of data that isn’t worth very much individually but has extremely high value if it’s ever needed - but you can’t know ahead of time which data that is. That’s all about ultra low price per gb storage that’s high durable.
Seems like a great way to:

- Archive for the film industry. 2 minutes to retrieve an archived take is a good compromise of cost and time.

- Archive for the recording industry. 2 minutes to recover the "master tapes" when you remaster for the umpteenth time is a good compromise of cost and time.

- The product I work on has requirements to archive data for a year / 10 years / 20 years and then generate a report. A 2 minute delay is perfectly acceptable.

My workplace archived old builds of our product. It’s very rare that we need them, but we do need them occasionally. We’d happily take 1 hour latency to be able to cheaply store more of them.
That is not niche at all. Pretty much any gov't, corp, etc that needs long term achives could do with faster access. 2 min is light speed since it's low(er) cost/decade/gb
If you want to build something like youtube, you could store the data from a start of a video, and data defining a slower resolution stream, and more popular videos on the lower latency parts of tapes in your storage arrays, and while the video plays you could go get the rest from the higher latency places.

You could also predict What is the next video the user will want.

Also, the 2 minute seek time is from one end of the tape to the other. But you could also move accross tracks, vertically so that gives another route for short seek times for some of your data.

So it might be possible to build something with a reasonable performance.

Depends on price. If it was good enough I'd use it for backups.
I see lot of reasonable use cases, say something like Internet Archive or library in general. Keep the frequented stuff in more hot cache and then rest of the stuff in colder storage. Without having to actually think about the process.

Similar could be some new media production company. Archive B-roll or parts of older videos and then have the stuff you use very common available much faster.

This could be easily improved by keeping track of what's on the tape where in part of the SSD, like a length-based filesystem. Then it is just a matter of seeking to that length of tape and finding the actual start of the data. You could take the seek time for cold data to just a couple dozen seconds.
That's plenty quick enough for the Windows Start button.

Zing!

Well yeah, it’s tape storage.
Is it like a hybrid drive, where there's no special hardware, just a controller trying to cache better than the host device would?
I'm not sure why people still need tape storage or hybrid SSD and tape storage, is it because of banking financial transactions recording requirements since the SSD storage now has similar or better storage capacity?

The reported storage size for the hybrid MED is 72TB and Micron now has 60 TB for single SSD drive [1].

Huawei has even higher storage capacity with reported 128 TB in a single SSD drive [2].

[1] Micron launches 60TB PCIe gen5 SSD with 12GB/s read speeds:

https://news.ycombinator.com/item?id=42122434

[2] Huawei unveils its OceanStor A800 AI-specific storage solution; announces 128TB high-capacity SSD:

https://www.datacenterdynamics.com/en/news/huawei-unveils-it...

Size per dingus is somewhat irrelevant to the use case: what's it matter if you buy 10 things that take up a rack or 20 things that take up a rack - how much did it cost to store all of your data at the speed you needed? Modern tape still has extremely good $/TB. If you have a LOT (and I don't just mean a fews 100s of TB) of data you need to keep but don't need to be able to access all of it constantly then tape can still save you a metric boatload of money on TCO vs SSD or HDD.

Dual tier systems are for when a small amount of that data may be hot at any given time (like you say, financial often has a lot of active data in a pool of tons of inactive data they can't throw out. Medical is another example). Rather than manual manage this you just run a 2 tier system that automatically bumps things in and out of high $$$, high performance storage banks to low $, high capacity storage banks transparently.

But why use tape storage in this day and age?
Cheap and very very reliable archival storage (essentially no failure modes other than the magnetic media itself), weight and volume wise pretty much still the best density you can get 18TB (up to 45 with compression) in a small 200g package is hard to beat when you need to move peta bytes around.

Also tape is much cheaper to produce than platters especially at large volumes.

In the 1980s or 90s, Jim Gray predicted that in the far future memory hierarchy would be reduced to main memory + tape.

Not sure that prediction holds up (although streaming dbs in the cloud seem awfully "tape-like" to me?), but I did attempt to quickly redo his "5 minute rule" calculations: https://news.ycombinator.com/item?id=42125052

If I did the math right, there's no point to swap or paging anymore, as the break-even point I got was evict after 1 week* (instead of 5 min) for random access patterns, and evict after 5 hours for sequential access patterns.

These days maybe those breakeven calculations would be more useful for deciding how long to keep something in a register and when it'd be better to recalculate or spill to main memory?

* of course, if you keep everything in main memory and forget nothing, we're close to what in 1998 was one of McCarthy's pipe dreams: Elephant2000 http://www-formal.stanford.edu/jmc/elephant/elephant.html ; I guess if I were to attempt something like Elephant2038, it'd wind up looking like a bunch of past-oriented LTL formulae?

(the "speech acts" were probably Winograd's influence?)

There are no downsides to keeping off-cloud physical backup sets of most critical data in a firesafe or two (and still needed for SOC 2 compliance I understand)

> Aug 23: CloudNordic said: “The attackers succeeded in encrypting all servers’ disks, as well as on the primary and secondary backup system, whereby all machines crashed and we lost access to all data.”

"Never underestimate the bandwidth of a station wagon full of tapes hurtling down the highway."

For big data transfers, this still holds true.

I think makes perfect sense for archiving. I keep hearing tapes last 30 years, while in the same room temperature SSDs last a few years. Just wish I could find a solid source for it!
Tape has a tremendous density advantage over disks. Think about how much area a tape occupies if you were to completely unroll it, and compare that to a disk platter. The flip side is access times are extremely long compared to disks, but for the usual application of backup and recovery that doesn't matter much. In this case they're pairing the tape with faster non-volatile storage so writes at least will be very quick (reads not so much, 2 minute access time!)
Backups or satellite imagery is my guess.
Nice idea to integrate all the moving parts in the same cartridge. Probably the economics are favorable now a days. If the components for tape drives are cheap it makes sense to put them in the cartridge and skip all the moving parts in the robot. The question is if the money can be recouped on maintenance saving? Tape robots are quite reliable but comes paired with an expansive support contract to ensure it’s running 24/7. Or maybe the purported cost saving is in the initial procurement? The LTO stuff has always been somewhat pricey.
ELI5: I’ve lost track of tape storage technologies. When does it make sense to use tape backup at home? Most people rely on cloud services or external hard drives for backing up large files like family photos and videos. But is there a point—like a certain amount of data—where it becomes convenient to use tape instead?

On the surface, this might seem like a simple question, but let me elaborate with an example: think about storing unavailable movies on high-quality 2K/4K/8K content. Would tape backup be a good fit for this kind of use at home?

The main problem with tape storage at small scales (including home use) has always been that tape drives are relatively expensive (several thousand dollars new), and buying just one makes it a single point of failure for your backup system. If your data set can fit on one or two dozen hard drives, then hard drives are the cheaper option (unless you keep them powered up and spinning 24/7).
> When does it make sense to use tape backup at home?

Never. Tape only makes sense for true cold storage-ie. business-critical backups that are only going to be read back in case of emergency, where a couple days to retrieve the data is an acceptable compromise. "Movies I don't watch very often" are not 'cold' enough- it's just not worth the hassle of switching tapes and maintaining the entirely different infrastructure for even finding the right tape and reading the data. And while the price of the media is attractive, even second hand drives are prohibitively expensive and packed with moving parts to break down. If you really have a use case for offline backups at home BD-R is much more accessible (and I did look into this recently for the same use-case; unfortunately optical disk robots are still out of the reach of enthusiasts too. Could be a cool DIY project though...)

The tape market is highly price-fixed and innovation-restricted. Intel has more or less of a total monopoly in this space. It's essentially LTO or go home. Your one and only option is enterprise grade hardware at enterprise prices. Even on eBay, prices are well beyond what I'd consider consumer-accessible. Your best bet is to piece together a fiber channel system and hope you get lucky on a cheap used drive.

IBM is wholly uninterested in selling tape to consumers, so you aren't going to have a good time trying to use it anyway.

Throughout my career, folks have asked me to look at tape for a wide range of uses, and even after investing a bunch of time and money, my conclusion is: few users need tape. For the vast majority of users, hard drives represent the pinnacle of tradeoffs in capacity, performance, and long-term reliability.
tape storage would be perfect for storing movies. There is not much random access and it's cheap. Also need it to last 100 years would be a good selling point. DVDs work but they do fail over time. Still waiting for that holographic or DNA storage.
Has it ever been revealed what Amazon is using to back Glacier. I had heard a lot of speculation including some sort of custom optical disk. Anyone know?
China is a country of over 1 billion people. It has advanced and surpassed many Western countries in the span of just 40 years. Trying to put a leash on them with sanctions is futile, hostile, and will only increase their determination. We should be trying to find common ground and working with them instead of fabricating another boogeyman for no reason.
>We should be trying to find common ground and working with them

Why should we, after what they did to the rights and freedoms of the people of Hong Kong?

Look at China's neighbors. I'm sure the solution to their conflicts with China is "finding common ground and working with them". Would you have told Ukraine the same thing about Russia before Russia invaded? Hell, do you say it now?
+500 social credits

Indeed, why not find common ground with a totalitarian regime that conducts genocides and violently suppresses democracy and dissent

take it from an old man that was born in eastern europe in another totalitarian regime, I've seen it in practice, I've seen the horrors behind closed doors: appeasement never works and only prolongs the suffering of the people.

I like your optimism but this ship sailed a long time ago.

If you are outside the US and China it is hard not to empathize with the Chinese annoyance with the do-as-we-say-not-as-we-do “rules based order”, and deciding to play the exact game on the US that the US has played with everyone else.

Reminds me of how some manufacturers tried to combine a small SSD with a large HDD the same way in the early 2010s.
Haha, Apple was still selling iMacs equipped with Fusion Drives in 2021. Really cool tech for the time, but should've stayed in the 2010s
This is great. I’ve been looking for more local multistage storage: with fast frequent access degrading to slow large capacity. In the future with large volumes I can store the text of the Internet for a personal archive.
How are you squeezing the market when your stats say 2Ms(minutes) seek time. If to becomes popular WD can easily copy that tech.
When Chinese implements an existing feature: stealing. When WD implements an existing feature: innovation
Chinese manufactured SSDs all use YMTC NAND and have serious problem with data retention. https://goughlui.com/2023/10/10/psa-ssds-with-ymtc-flash-pro...

Tape might be a bodge to make product usable in enterprise.