It is troubling that Optane storage devices are UNUSABLE for long-term cold (unpowered) storage. This has already proven to be the case, though less so, with NAND storage. I for one would find loss of my data, etc., on ANY storage device that has not received power for a period of even a decade to be unacceptable, let alone a period of as little as three months! This is an area where Intel really needs to rethink and improve the product before I would ever use it in a non-enterprise environment. Moreover, should the U.S., for example, ever be subject to a disruptive EMP event where power is out for months or even years, whole companies relying on Optane-based drives for ANY sort of data backup would lose everything....
EMP protected storage at consumer prices means optical media and burying spare PCs somewhere deep, I guess.
Other than that, you simply need to think more than binary about storage: Tiers and characteristics can only evolve in to more complex evolutionary trees, give that everybody wants something different at her price point.
My jaw dropped when I read 3 months. Either this was a printing mistake or Optane cannot be recommended for any consumer systems laptops/desktops. How is that they are calling it to be the same as nand. https://www.anandtech.com/show/9248/the-truth-abou... Note the numbers are for max written drives and not for those which still have a lot of PE cycles left. So the actual retention is much higher..
Yeah, the data retention period is rather bad. Although I wouldn't want to use Optane as a cold storage backup mechanism anyway because the performance and price rule it out as a candidate, I do regularly ignore my computers for weeks or sometimes several months before getting around to noticing them again. My gaming laptop in particular collects a lot of dust in the warmer months of the year. Optane in that case would have to function as a cache system between the SSD and RAM only rather than as the system's only storage device. Since NAND already sucks for cold storage, I guess I'm not going to escape that 2.5" 1TB laptop drive stuffed in a cheap external case as my local backup anytime soon. *sad panda*
The numbers quoted as specified in the answer are for EOL endurance drives just like the nand numbers in the linked article and are basically the same length. Should you use NAND or Optane for archival purposes? No. But for nearline or online storage they are both fine.
For consumers the idea is that you won't be using phase change for storage but for cache, the difference being that the cache will be overwritten continously and therefore never grow stale as long as you use the device.
um... why is this a problem exactly? You expecting to lock your production server up in a closet for a decade and then hope to get something off of it? As with all things in technology, the further you press things, the more specialized it gets. A brand spanking new CPU is only moderately faster on general application performance compared to a similarly clocked CPU from 5 years ago. BUT, for specific tasks like encryption new chips will run circles around the old chips. The new nVidia 20XX GPUs are only slightly faster than their 2 year old counterparts... but they bring with them extra specialization around ray tracing and AI compute loads that the 10XX chips simply don't do well. In the server market these changes are ever more pronounced. Production units tend to be all about RAM, cores, and IOPS. Storage machines then to be more about lots of redundant storage, but less horsepower. For every tool, a purpose. You just cant expect every tool to fit every purpose.
>Having your laptop powered off for 3 months and all of it data maybe unretriavable?
1) You're screwing with your laptop's battery longevity by doing that.
2) If you're not using your laptop for 3 months, then consider throwing it away because it looks like you don't really use it or need it.
3) The compute space still moves fast, despite moore's law being dead. If you're seldomly using a laptop maybe 1 time every 3 months, then why are you investing money in a new storage drive for it when you hardly touch it to begin with?
3a) If you hardly touch your laptop every 3 months, then chances are it's pretty old. Are you sure it even supports an M.2 drive, let alone one that's actually compatible for Intel Optane caching? If it's not that old and you don't even use it, mind mailing it to me so I can smash it into pieces for you? It sounds like you have extra money and electronics you don't need or use.
4) If you need data shared between multiple pc platforms, there are dozens of solutions out there, even free ones, so that in this way you're not potentially losing data you need.
If you need long term cold storage, get a proper solution rather than pretending you're ticked off on some internet news article's comment section just because one (of the dozens of different storage technologies that exist) doesn't tickle your fancy in terms of long term cold storage.
Most laptop batteries have no problems with being left alone for 3 months. I've been regularly leaving my various gaming laptops alone over the past decade and the only one that gave me any trouble with battery longevity was a Dell Latitude D620. OEM battery replacements at the time were a bit more than I was willing to pay so I purchased 3rd party batteries and went through two of them in two years. The problem speaks more to the lack of battery quality than it does to my treatment.
As for your second and third points (plus 3a since all of them are basically parts of the same argument), I'm not sure I understand why you feel hostility toward people that leave a computer alone for a bit. Either way, I do admit that I feel bad because I see a lot of the same frustrations in you that I do in my son. People shouldn't have to worry about what someone else thinks or get upset about their ownership over an infrequently used computer commented on by some faceless person in the Internet.
If it'd help you feel better, I'd donate my gaming laptop to you as I expect I won't need it in the coming weeks. It's nothing high end, a Dell Latitude e6430 (Ivy Bridge i7) with 8GB of RAM and a Quadro NVS 5200M. I have been thinking about getting rid of it for months and was probably going to just drop it off at a thrift store near my home that sends all its profits to a local animal shelter sometime in the next four months. I'd pull the SSD out to slide into another laptop I have laying around so you'd need to arrange for your own storage solution and, as I have other Latitudes around, I won't include a power supply either. I'd only ask you cover shipping costs for it since a thrift store run is a lot cheaper than packing it up to send to who-knows-where.
Simple solution. DON'T USE OPTANE THEN. How hard is that? Stick to 2.5" HDDs or SSHDs then for your laptop. You're looking at this the wrong way. And if you care about data retention why are you so focused on the integrity of just one medium? Where's your 3-2-1 backup rule? You're doing it wrong.
Optane is just one of dozens of different storage technology solutions. I don't understand why people constantly get bent out of shape that this very specific flavor of storage technology excels in random IOPS but suffers in cold storage use-cases. If your data is specifically needing to go in cold storage, then don't use optane. It's that freaking simple.
I hate to be the idiot that makes car comparisons to technology, but people in racing world don't complain when Hankook releases a new formula of racing slick tires, which excel in giving maximum traction on sticky tarmac road surfaces. If a racing team wanted new tires for their rally cars, they wouldn't look at the racing slicks and whine, they'd be looking at the tire options that excel for offroad use or tire chains for icy road conditions. Racing slicks can be similar to Optane drives, really fast and the best option available when your one goal is maximum speed when the storage conditions are just right for it. Don't use racing slicks when you're driving offroad. Don't use optane drives when you're using systems for cold storage.
It's really that simple. Pretending to be stupid and p'd off in a comments section that EVERY SINGLE STORAGE TECHNOLOGY doesn't excel in cold storage use-cases is amazingly stupid and conceited.
"Pretending to be stupid and p'd off in a comments section that EVERY SINGLE STORAGE TECHNOLOGY doesn't excel in cold storage use-cases is amazingly stupid and conceited."
Calm down over there. It's only computer technology.
I've had work assignments where I'm away from my house and desktop computer for more than 3 months. That's not unheard of. I use it for reasonably disk IO bound tasks for work (VM, DB testing, some image and video processing), so it's reasonable to assume that there are users that want (or need) high speed storage, that also might not use it regularly.
So I don't begrudge any user that has a similar setup with a laptop.
Also, what about IT departments for an engineering company that need hot spares available for staff? I guess you could probably just reimage the machines when delivering a spare. But there's something really nice about having "hot" spares as an IT worker. When you bill per hour, that's a lot of wasted money if your employees have to twiddle their thumbs waiting for a spare to be ready.
>want (or need) high speed storage, that also might not use it regularly. THEN USE IT AS A CACHING DRIVE (rather than installing your OS/software or keeping your data on the drive exclusively) OR SET UP A RAMDISK ON BOOT FOR THESE TASKS.
There's options to have both, but you people are being morons about this. Structure your storage volumes correctly and you shouldn't have issues with data longevity.
The ratings are a worst case scenario for worn out drives. People don’t realize that until recently SSDs had similar ratings, and they’re still not great. In practice it’s rarely a problem.
If you care about long term data retention, why would you ONLY have it on one form of medium? What about the 3-2-1 rule? This is a ridiculous premise if you truly are trying to think about cold storage.
The level of chaos resulting from a massive emp that disrupts power on a months-long scale will put most companies out of business regardless of whether they can recover their software.
That said, every large enterprise has a tiered backup plan that eventually gets down to emp immune media. Most would lose less than a month of data.
Geez this was hard to read. It feels very sterile. I am 100% certain that robots exist today that could provide the same responses except for the last few that have actual content.
What I read between the lines of Intel answers: Hardware RAID no longer makes any sense. XOR-Offload and other logic are better dealt with by the CPU. What remained was draining a BBU backed DRAM cache after a power recovery, because that either involves heavy OS rework or the smart controller.
With writeback cache fully committing to cache line phase change NV-RAM instead of BBU backed DRAM, any capacitors are really just about finishing the current block and perhaps another n-copy block of initial metadata for the NV-RAM journal: That is OS transparent enough and were no longer dealing with IDE interfaces even at boot.
missing "granularity": 'cache line granularity phase change', want edit. Just want to high-light, it may be byte addressable but AFAIK all memory controller operations are actually at cache line level, including some write amplification for a theoretical byte update.
Byte-addressable does not mean individual byte-alterable directly because there exists a minimum write granularity for ECC parity and encryption. Altering a byte means reading that whole thing, changing the byte and reapplying ECC and encryption.
And since write granularity is large, possibly needing a read, some amount of power loss protection is required, no matter how fast or good the media. Nothing can be instantaneous. And I would bet there is a pipeline of data flow through the controller so it's not just one thing at a time to be concerned with. That can be seen in the reviews on this site that show Q depth still affects bandwidth on the optane SSDs... Matters a lot less than a nand SSD but still matters.
nice summary! Some links to intel are broken like in the answers to Q: Are there any plans to create Optane cache RAID controllers? Q: I use Optane quite a bit but in configurations that are not officially supported. ...
3 month unpowered retention time: I guess the major takeaway is that surprises tend to be least appreciated in storage. And storage which forgets when it’s not used is far more ‘biological’ than most of us like in a computer, which is mostly about complementing our talents.
While I have abstractly known that consumer and enterprise SSDs have rather distinct standardized retention times, I have always wondered how you would have to compensate for that: Should you actively re-write logical data blocks at a minimum of said retention time? Because in non-cache usage there would be blocks, that might never get rewritten normally during normal usage e.g. the master boot block or similar r/o master data even in an otherwise very active database.
Should you do patrol reads or at least regular full disk reads to catch failing blocks and have them rewritten or is that something that an SSD would do on its own?
Would it have to be powered on to do that sort of background maintenance or does it need to be told that six months have passed since the last power-on so it can gauge its housekeeping accordingly?
Sounds like a really nice topic for an in-depth coverage of the type that sets Anandtech apart :-)
We’ve updated our terms. By continuing to use the site and/or by logging into your account, you agree to the Site’s updated Terms of Use and Privacy Policy.
30 Comments
Back to Article
thuckabay - Thursday, October 11, 2018 - link
It is troubling that Optane storage devices are UNUSABLE for long-term cold (unpowered) storage. This has already proven to be the case, though less so, with NAND storage. I for one would find loss of my data, etc., on ANY storage device that has not received power for a period of even a decade to be unacceptable, let alone a period of as little as three months! This is an area where Intel really needs to rethink and improve the product before I would ever use it in a non-enterprise environment. Moreover, should the U.S., for example, ever be subject to a disruptive EMP event where power is out for months or even years, whole companies relying on Optane-based drives for ANY sort of data backup would lose everything....abufrejoval - Thursday, October 11, 2018 - link
EMP protected storage at consumer prices means optical media and burying spare PCs somewhere deep, I guess.Other than that, you simply need to think more than binary about storage: Tiers and characteristics can only evolve in to more complex evolutionary trees, give that everybody wants something different at her price point.
sharath.naik - Thursday, October 11, 2018 - link
My jaw dropped when I read 3 months. Either this was a printing mistake or Optane cannot be recommended for any consumer systems laptops/desktops. How is that they are calling it to be the same as nand. https://www.anandtech.com/show/9248/the-truth-abou... Note the numbers are for max written drives and not for those which still have a lot of PE cycles left. So the actual retention is much higher..PeachNCream - Thursday, October 11, 2018 - link
Yeah, the data retention period is rather bad. Although I wouldn't want to use Optane as a cold storage backup mechanism anyway because the performance and price rule it out as a candidate, I do regularly ignore my computers for weeks or sometimes several months before getting around to noticing them again. My gaming laptop in particular collects a lot of dust in the warmer months of the year. Optane in that case would have to function as a cache system between the SSD and RAM only rather than as the system's only storage device. Since NAND already sucks for cold storage, I guess I'm not going to escape that 2.5" 1TB laptop drive stuffed in a cheap external case as my local backup anytime soon. *sad panda*ats - Thursday, October 11, 2018 - link
The numbers quoted as specified in the answer are for EOL endurance drives just like the nand numbers in the linked article and are basically the same length. Should you use NAND or Optane for archival purposes? No. But for nearline or online storage they are both fine.abufrejoval - Friday, October 12, 2018 - link
For consumers the idea is that you won't be using phase change for storage but for cache, the difference being that the cache will be overwritten continously and therefore never grow stale as long as you use the device.CaedenV - Thursday, October 11, 2018 - link
um... why is this a problem exactly? You expecting to lock your production server up in a closet for a decade and then hope to get something off of it?As with all things in technology, the further you press things, the more specialized it gets.
A brand spanking new CPU is only moderately faster on general application performance compared to a similarly clocked CPU from 5 years ago. BUT, for specific tasks like encryption new chips will run circles around the old chips. The new nVidia 20XX GPUs are only slightly faster than their 2 year old counterparts... but they bring with them extra specialization around ray tracing and AI compute loads that the 10XX chips simply don't do well.
In the server market these changes are ever more pronounced. Production units tend to be all about RAM, cores, and IOPS. Storage machines then to be more about lots of redundant storage, but less horsepower.
For every tool, a purpose. You just cant expect every tool to fit every purpose.
halcyon - Thursday, October 11, 2018 - link
He wrote: "I would ever use it in a non-enterprise environment"Having your laptop powered off for 3 months and all of it data maybe unretriavable?
This product is not ready for consumer mass market.
24/7 production environments maybe, consumer use, no.
JoeyJoJo123 - Thursday, October 11, 2018 - link
>Having your laptop powered off for 3 months and all of it data maybe unretriavable?1) You're screwing with your laptop's battery longevity by doing that.
2) If you're not using your laptop for 3 months, then consider throwing it away because it looks like you don't really use it or need it.
3) The compute space still moves fast, despite moore's law being dead. If you're seldomly using a laptop maybe 1 time every 3 months, then why are you investing money in a new storage drive for it when you hardly touch it to begin with?
3a) If you hardly touch your laptop every 3 months, then chances are it's pretty old. Are you sure it even supports an M.2 drive, let alone one that's actually compatible for Intel Optane caching? If it's not that old and you don't even use it, mind mailing it to me so I can smash it into pieces for you? It sounds like you have extra money and electronics you don't need or use.
4) If you need data shared between multiple pc platforms, there are dozens of solutions out there, even free ones, so that in this way you're not potentially losing data you need.
If you need long term cold storage, get a proper solution rather than pretending you're ticked off on some internet news article's comment section just because one (of the dozens of different storage technologies that exist) doesn't tickle your fancy in terms of long term cold storage.
PeachNCream - Friday, October 12, 2018 - link
Most laptop batteries have no problems with being left alone for 3 months. I've been regularly leaving my various gaming laptops alone over the past decade and the only one that gave me any trouble with battery longevity was a Dell Latitude D620. OEM battery replacements at the time were a bit more than I was willing to pay so I purchased 3rd party batteries and went through two of them in two years. The problem speaks more to the lack of battery quality than it does to my treatment.As for your second and third points (plus 3a since all of them are basically parts of the same argument), I'm not sure I understand why you feel hostility toward people that leave a computer alone for a bit. Either way, I do admit that I feel bad because I see a lot of the same frustrations in you that I do in my son. People shouldn't have to worry about what someone else thinks or get upset about their ownership over an infrequently used computer commented on by some faceless person in the Internet.
If it'd help you feel better, I'd donate my gaming laptop to you as I expect I won't need it in the coming weeks. It's nothing high end, a Dell Latitude e6430 (Ivy Bridge i7) with 8GB of RAM and a Quadro NVS 5200M. I have been thinking about getting rid of it for months and was probably going to just drop it off at a thrift store near my home that sends all its profits to a local animal shelter sometime in the next four months. I'd pull the SSD out to slide into another laptop I have laying around so you'd need to arrange for your own storage solution and, as I have other Latitudes around, I won't include a power supply either. I'd only ask you cover shipping costs for it since a thrift store run is a lot cheaper than packing it up to send to who-knows-where.
CheapSushi - Friday, October 12, 2018 - link
Simple solution. DON'T USE OPTANE THEN. How hard is that? Stick to 2.5" HDDs or SSHDs then for your laptop. You're looking at this the wrong way. And if you care about data retention why are you so focused on the integrity of just one medium? Where's your 3-2-1 backup rule? You're doing it wrong.JoeyJoJo123 - Monday, October 15, 2018 - link
^^^ Basically this.Optane is just one of dozens of different storage technology solutions. I don't understand why people constantly get bent out of shape that this very specific flavor of storage technology excels in random IOPS but suffers in cold storage use-cases. If your data is specifically needing to go in cold storage, then don't use optane. It's that freaking simple.
I hate to be the idiot that makes car comparisons to technology, but people in racing world don't complain when Hankook releases a new formula of racing slick tires, which excel in giving maximum traction on sticky tarmac road surfaces. If a racing team wanted new tires for their rally cars, they wouldn't look at the racing slicks and whine, they'd be looking at the tire options that excel for offroad use or tire chains for icy road conditions. Racing slicks can be similar to Optane drives, really fast and the best option available when your one goal is maximum speed when the storage conditions are just right for it. Don't use racing slicks when you're driving offroad. Don't use optane drives when you're using systems for cold storage.
It's really that simple. Pretending to be stupid and p'd off in a comments section that EVERY SINGLE STORAGE TECHNOLOGY doesn't excel in cold storage use-cases is amazingly stupid and conceited.
PeachNCream - Monday, October 15, 2018 - link
"Pretending to be stupid and p'd off in a comments section that EVERY SINGLE STORAGE TECHNOLOGY doesn't excel in cold storage use-cases is amazingly stupid and conceited."Calm down over there. It's only computer technology.
erple2 - Saturday, October 13, 2018 - link
I've had work assignments where I'm away from my house and desktop computer for more than 3 months. That's not unheard of. I use it for reasonably disk IO bound tasks for work (VM, DB testing, some image and video processing), so it's reasonable to assume that there are users that want (or need) high speed storage, that also might not use it regularly.So I don't begrudge any user that has a similar setup with a laptop.
Also, what about IT departments for an engineering company that need hot spares available for staff? I guess you could probably just reimage the machines when delivering a spare. But there's something really nice about having "hot" spares as an IT worker. When you bill per hour, that's a lot of wasted money if your employees have to twiddle their thumbs waiting for a spare to be ready.
JoeyJoJo123 - Monday, October 15, 2018 - link
>want (or need) high speed storage, that also might not use it regularly.THEN USE IT AS A CACHING DRIVE (rather than installing your OS/software or keeping your data on the drive exclusively) OR SET UP A RAMDISK ON BOOT FOR THESE TASKS.
There's options to have both, but you people are being morons about this. Structure your storage volumes correctly and you shouldn't have issues with data longevity.
sor - Sunday, October 14, 2018 - link
The ratings are a worst case scenario for worn out drives. People don’t realize that until recently SSDs had similar ratings, and they’re still not great. In practice it’s rarely a problem.CheapSushi - Friday, October 12, 2018 - link
If you care about long term data retention, why would you ONLY have it on one form of medium? What about the 3-2-1 rule? This is a ridiculous premise if you truly are trying to think about cold storage.surt - Sunday, October 14, 2018 - link
The level of chaos resulting from a massive emp that disrupts power on a months-long scale will put most companies out of business regardless of whether they can recover their software.That said, every large enterprise has a tiered backup plan that eventually gets down to emp immune media. Most would lose less than a month of data.
willis936 - Thursday, October 11, 2018 - link
Geez this was hard to read. It feels very sterile. I am 100% certain that robots exist today that could provide the same responses except for the last few that have actual content.PeachNCream - Thursday, October 11, 2018 - link
The answers did have a "sanitized-by-legal-and-marketing-departments" feel to them.surt - Sunday, October 14, 2018 - link
Intel is at the cutting edge of AI research. There's actually a pretty solid chance these answers _were_ provided by robots.abufrejoval - Thursday, October 11, 2018 - link
What I read between the lines of Intel answers:Hardware RAID no longer makes any sense. XOR-Offload and other logic are better dealt with by the CPU. What remained was draining a BBU backed DRAM cache after a power recovery, because that either involves heavy OS rework or the smart controller.
With writeback cache fully committing to cache line phase change NV-RAM instead of BBU backed DRAM, any capacitors are really just about finishing the current block and perhaps another n-copy block of initial metadata for the NV-RAM journal: That is OS transparent enough and were no longer dealing with IDE interfaces even at boot.
abufrejoval - Thursday, October 11, 2018 - link
missing "granularity": 'cache line granularity phase change', want edit. Just want to high-light, it may be byte addressable but AFAIK all memory controller operations are actually at cache line level, including some write amplification for a theoretical byte update.woggs - Thursday, October 11, 2018 - link
That all seems about right.Byte-addressable does not mean individual byte-alterable directly because there exists a minimum write granularity for ECC parity and encryption. Altering a byte means reading that whole thing, changing the byte and reapplying ECC and encryption.
And since write granularity is large, possibly needing a read, some amount of power loss protection is required, no matter how fast or good the media. Nothing can be instantaneous. And I would bet there is a pipeline of data flow through the controller so it's not just one thing at a time to be concerned with. That can be seen in the reviews on this site that show Q depth still affects bandwidth on the optane SSDs... Matters a lot less than a nand SSD but still matters.
hpvd - Thursday, October 11, 2018 - link
nice summary!Some links to intel are broken like in the answers to
Q: Are there any plans to create Optane cache RAID controllers?
Q: I use Optane quite a bit but in configurations that are not officially supported. ...
Ryan Smith - Thursday, October 11, 2018 - link
Thanks!abufrejoval - Friday, October 12, 2018 - link
3 month unpowered retention time: I guess the major takeaway is that surprises tend to be least appreciated in storage. And storage which forgets when it’s not used is far more ‘biological’ than most of us like in a computer, which is mostly about complementing our talents.While I have abstractly known that consumer and enterprise SSDs have rather distinct standardized retention times, I have always wondered how you would have to compensate for that: Should you actively re-write logical data blocks at a minimum of said retention time? Because in non-cache usage there would be blocks, that might never get rewritten normally during normal usage e.g. the master boot block or similar r/o master data even in an otherwise very active database.
Should you do patrol reads or at least regular full disk reads to catch failing blocks and have them rewritten or is that something that an SSD would do on its own?
Would it have to be powered on to do that sort of background maintenance or does it need to be told that six months have passed since the last power-on so it can gauge its housekeeping accordingly?
Sounds like a really nice topic for an in-depth coverage of the type that sets Anandtech apart :-)
sendai - Friday, October 12, 2018 - link
"Our Optane drives do have power loss protection to protect customer data. We have released a 100GB M.2."Are they talking a out the 800P here or another product? As far as I've heard the smallest 905P will be 380GB.
emvonline - Monday, October 15, 2018 - link
question: is optane pm read latency close to dram? can i write to pm directly using rdma?answer: we do think the redsox will win the world series. and we like the cloud
seriously ... its like watch a political debate