I found this read on the internet interesting.
One of the statements: “In other words: You’ll probably upgrade your entire computer before your SSD fails.”
2 Likes
I think you mean “the USEFUL capacity” otherwise it wouldn’t be an “something” TB SSD.
We already know that some Software is telling you that your SSD’s health is decreasing! But, what makes you think it REALLY is decreasing? Have you noticed anything? Does it have “bad sectors” already?
If you don’t notice anything happening to the SSD… I wold say, leave it alone! and don’t care about any of that.
2 Likes
i’m pretty sure there is a case to be made for trying to keep an eye on system resources and possibly even replace them before they die unexpectedly.
1 Like
The point here is that while the capacity figure might be, say 1TB the useful capacity will be in the 960’s GBs region, no matter if you use an HDD or a SSD it has always been the case for as far back as I can remember. I think @daniel.m.tripp might know the answer as to why this is the case.
Taking what @dreis has to say I have to agree with the point he made about software might be telling you something, but if you really don’t notice a downturn, leave it alone or you’ll end up being totally confused. As @01101111 says there is a case for keeping an eye on thing which is wise to do, but my view is much like Corporal Jones in Dad’s Army - DON’T PANIC DON’T PANIC !
1 Like
I’m not saying an SSD won’t go bad, but as far as a user wearing one out…
“A typical TBW figure for a 250 GB SSD lies between 60 and 150 terabytes written. That means: To get over a guaranteed TBW of 70, a user would have to write 190(!) GB daily over a period of one year (In other words, to fill two thirds of the SSD with new data every day). In a consumer environment this is highly unlikely.”
Of course SSD of 120 GB and 60 GB will wear a lot faster b/c they have less cells. And my guess is that SSD of 480 GB and greater would last a typical user maybe 10 years!
2 Likes
You’re totally right!
I think these questions are due to lack of experience working with SSD
2 Likes
I really have no idea 
- part of it is marketing “engineers” calling 1000 GB a TB, 1000MB a GB etc… when it should be 1024… Also - there’s significant “overhead” when allocating blocks (e.g. number of available 512 byte blocks x 512 bytes is your true size - I believe, don’t quote me - I could be wrong - still remember SunOS [pre Solaris BSD] would always just display the free space [df command] in number of free 512 byte blocks)…
Sheesh - who remembers when your BIOS couldn’t even query or “auto detect” the controller on the HDD - and you had to locate various arcane hard disk databases to specify number of cylinders and platters and sectors (it’s been so long, thankfully I’ve forgotten most of the terminology)??? What a PITA that was! Then there was a whole bunch of different cabling systems, MFM, IDE, ESDI (not to mention variations on SCSI)…
kids today - you got it so easy
3 Likes
For me 1 TB is 1024 GB too but I only realized a couple days ago the system has changed when I searched for more information about SSD
Now we have two multiples of bytes:
-
In decimal system where 1 TB = 1000 GB
-
In binary system where former 1 TB = 1024 GB are now 1 TiB = 1024 GiB because the prefixes are changed to binary system
Binary system prefixes
Former New
K - Kilo Ki - Kibi
M - Mega Mi - Mebi
G - Giga Gi - Gibi
T - Tera Ti - Tebi
Nowadays, if I say 1 TB (terabyte) is really 1000 GB (Gigabyte). If I want to say like I always said in binary system, I need to say 1 TiB (Tebibyte) = 1024 GiB (Gibibyte)
Standard since 2002 - IEE1541-2002
For me 1 TB still is 1024 GB… 
My first HDD was an MFM with 20 MB or I should say 20 MiB?
I remember when I needed to add HDD spec’s (cylinders, heads and sectors) manually in BIOS
Photos from web:
I wasn’t going to write more about this thread, but after reading the post, I need to talk more about it:
Yes, you’re right
Yes, you’re right too
But after reading a little more about SSD I realize if you have, for example, 1/5 of free space, your SSD’s life is shorter than if you have1/2 of free space.
For other side, you can buy specific SSD with different over-provisioning, If you want to use SSD for write intensive or for read intensive, and you can add additional over-provisioning to increase your SSD’s life.
So, I can have a SSD with 256 GB but the useful capacity depends on these items:
- How much space I will use in SSD
- How much space I want to leave free
- How much is the over-provisioning of my SSD
- How much I want to add additional over-provisioning
This is an extract from Seagate about over-provisioning:
In practice, an SSD’s performance begins to decline after it reaches about 50% full. This is why some manufacturers reduce the amount of capacity available to the user and set it aside as additional over-provisioning. For example, a manufacturer might reserve 28 out of 128 GB and market the resulting configuration as a 100 GB SSD with 28% over-provisioning. In actuality, this 28% is in addition to the built-in 7.37%, so it’s good to be aware of how vendors toss these terms around. Users should also consider that an SSD in service is rarely completely full. SSDs take advantage of this unused capacity, dynamically using it as additional over-provisioning.
OK, maybe I’m getting crazy with this subject and still haven’t accepted that it is a new technology and that I need to forget what I know about the HDDs and accept the SSD technology as it is but, in my opinion, the SSD storage capacity is a utopian value.
Links:
https://www.seagate.com/pt/pt/tech-insights/ssd-over-provisioning-benefits-master-ti/
https://www.kingston.com/en/ssd/overprovisioning
Sorry about my English
Moderators. Please feel free to correct my text. Thank you
3 Likes
I love it! Someone who seems to know a bit about hardware (H/W). Now to mix things up a bit. No expert, gather from my reading.
On the market for the consumer here are 2 types of SSD to choose from. Most of us buy the less expensive one labeled TLC. The other type of SSD is MLC which is harder to find and a bit more expivse. Why? MLC SSD’s will place 2 bytes per cell and TLC SSD puts 3 bytes per cell. So for MLC you need more cells. 120 GB of cells x 2 = 240 GB disk. For TLC, 80 GB of cells x 3 = 240 GB. So, as you can see, 3D uses the cells more MLC and will wear fastrer. There is also SLC SSD’s (one byte per cell) but only used in a bussiness.
I have read the same thing, but I wonder if the life is really shorter or whether you just get a warning sooner that the SSD is wearing? If 4/5 of the disk (80%) of the SSD is not being used, how can that part wear out?
I don’t know about you, but with my HDD, I would only check SMART data maybe 3 or 4 times a year. Today’s HDD are very reliable. So now with SSD so much faster (“Even the worst SSD is at least three times as fast as a hard drive”) and even more reliable, why do we worry about them?
|||
A little ancient history. My first exposure to a HDD was at work back in 1970. It was an IBM 2311. You could actually see the platters! It had a capacity of 7.25 MB. The mainframe ran IBM DOS. To write a file to the disk, you had to actually tell the computer where to start on the disk (track number) and how many tracks you wanted with punched cards called a DLBL (for file name) and an ‘extent’ card. So an example might look like;
// DLBL DSN=‘My first file on disk’,Disp=SHR,Exp=99/365
// EXTENT sys013,wssc43,1122,120
|
2 Likes
Nice reply Howard, although looking at the two types of SSD I have I can not find either of the things you mention here. Nor could I find reference to them on line anywhere, but then it is only 04:00 as I write this. I too remember the platters and having to actually tell the computer where to start, which goes to show how far we’ve come, but also a reminder that HDDs still use the platter.
2 Likes
@easyt50,
I don’t know about hardware, I just try to get more information about a SSD which is a new technology for me.
For example, I learned on this forum that I needed to turn on the SSD setting. I thought It was automatic and not manually.
About SSD types
Yes I know that and you’re right
In my example, I said only 1/5 of SSD is free
@easyt50,
like I said in last post I think maybe I still haven’t accept the SSD technology and it’s secure to use it to store our data instead HDDs. Possibly, in 6 months, I’ll say SSDs are excellent and very reliable
About History
I can only speak up to ZX Spectrum, my friend. 
1 Like
My first IT job was as mainframe operator in 1992 - it was an IBM 4381, the DASD (IBM get religious on the naming, HDD’s are “Direct Access Storage Devices”) cabinets (I think we had 3 or 4) were the size of giant fridges - if you opened up the cabinet, it was a huge pyrex/glass doohicky with belts driving the platters, and when the whole mainframe was booted up (an IML) you coudn’t stay in the room, was like standing next to a 737 :
By 1992 - we didn’t have any reel to reel tape systems, it was all cartridge… I’ve still got a couple of badges off the old girl, when IBM upgraded the “release” - they replaced the IBM 4381 badges (I have one stuck to my BananaPi with bluetack)…
Job didn’t last too long - 1992 - 1994, when the place I was working got privatized, and the new owner didn’t want a mainframe, so on the last day there, we powered it down never to IML again 
… big crowd in the computer room and champagne corks a popping (I’ve got a video on youtube of that last day - I should edit it out so it’s only got the mainframe shutdown section)…
2 Likes
The same reason why you need special SSD eraser programs to actually delete the deleted files. SSDs are based on an entirely different concept. HDDs are purely mechanical, in a sense, whereas SSDs are a different matter. I’m not an expert on SSDs, as I almost always deal with HDDs, so I won’t dive into explaining too deeply.
The point of what I’m saying is that if you use an SSD, any part of the SSD might have remnants of whatever was ever written to it. That’s why deleting 100% by far does not guarantee deleted data. In fact, you may delete your SSD entirely a 100 times and there is still no guarantee that it actually has overwritten all the parts of that SSD. That is why you need the aforementioned special eraser programs and that is also why the entire SSD “wears out” even if you never have more than let’s say 20% usage in relation to the SSD’s total storage capacity.
1 Like
Sorry, I was not clear. What I was trying to say was if 80% of the SSD is full with data that’s not changing meaning the cells are only being read (not erased and re-used) then that part of the SSD could not wear or wear very slowly. It the number of writes (which is actually an erase then a write) that wears a cell out. So in your example, only 1/5 or 20% of the cells are free to be R/W to.
1 Like
"Types of NAND Flash
NAND flash is a type of non-volatile storage architecture used in SSDs and memory cards. It gets its name from the type of the logic gate (NOT-AND) used to determine how digital information is stored in a flash device’s chips.
SLC
Single-Level Cell SSDs store one bit in each cell, a design that yields enhanced endurance, accuracy and performance. For critical enterprise applications and storage services, SLC is the go-to flash technology. Of course, it carries the highest price tag.
MLC
Considered the consumer-grade flavor of flash, Multi-Level Cell architectures can store two bits per cell. Although it may seem like packing more than one bit into a memory cell is a good use of space, it comes at the cost of a lower useful life and decreased reliability. MLC SSDs make it possible to economically add flash storage to PCs and laptops, relatively speaking.
eMLC
Enterprise Multi-Level Cell is a hardier version of MLC NAND flash that somewhat bridges the performance and endurance gap between SLC and MLC. eMLC drives costs more than MLC drives, but much less than SLC their counterparts. Although it still stores two bits per cell, an eMLC drive's controller manages data placement, wear leveling and other storage operations in a way that extends an eMLC SSD's useful life.
TLC
The least expensive of the bunch, Triple-Level Cell NAND flash stores three bits per cell and is typically used in consumer-grade electronics with comparatively low performance and endurance requirements. Best suited for read-heavy applications, TLC-based storage components were rarely, if ever, used in business environments but recent improvements in flash architectures, including 3D NAND (more on that later), and endurance-enhancing data placement and error correction techniques have earned the technology a place in read-intensive enterprise storage applications."
Reference: https://www.enterprisestorageforum.com/storage-hardware/slc-vs-mlc-vs-tlc-nand-flash.html
I said byte per cell and I was wrong. Only bits are storage in them. 1, 2, or 3. Again, only my opinion from what I have read.
3 Likes
| SLC | eMLC | MLC | TLC | |
|---|---|---|---|---|
| Bits per Cell | 1 | 2 | 2 | 3 |
| Max Program/Erase (PE) Cycles | 100,000 | 30,000 | 10,000 | 5,000 |
| Reliability/Endurance | Excellent | Very Good | Very Good | OK |
| Cost | $$$$$ | $$$$ | $$$ | $$ |
| Use Case | Enterprise | Enterprise | Enthusiast/Consumer | Consumer |
If you been reading these post I have a correction to make. TLC and 3D is not the same. 3D is a manufacturing process and not a way to store bits. I corrected my earier post.
Also on the way is QTL, 4 bits per cell.
2 Likes