Personal computing discussed
Moderators: renee, morphine, Steel
just brew it! wrote:Just because it has no moving parts doesn't mean a SSD can't fail or wear out.
Flash cells used in typical consumer SSDs can only be written a few thousand times before they start to degrade and lose their ability to retain data. Wear leveling and error correction algorithms make the effective durability of the drive a lot better than this, but it still has a finite lifetime. A heavily used SSD is probably no more reliable over the long term than a mechanical HDD.
just brew it! wrote:Even if SSDs were as durable as you seem to think they are, what makes you think that the computing devices in use several decades from now will even have SATA ports on them? Storage capacities will also have increased by several orders of magnitude..
just brew it! wrote:From a reliability, compatibility, and capacity standpoint, passing down one of today's SSDs to your grandkids a quarter century from now will probably make about as much sense as your parents giving you a box of old 5-1/4" floppy disks today.
WhatMeWorry wrote:With this assumption (and I hope I am wrong here): that HDDs and SSD are as good as it gets. Then wont SSDs be the only non-volatile storage that consumers (house-holds) will need/use? And HDDs will only be used by the cloud were capacity will always be king.
Ryu Connor wrote:It's worth noting that writes are not the only thing that "wear" away NAND. Reads also take a toll.
sschaem wrote:"Say you can write the cells 5000 times before they become useless; for a 128GB drive, that would be 640,000GB written to it, or 100GB per day for 17 years"
Doesn't work that way. Only the unused space can be used to spread write data.
So if you have a 64GB drive and its 80% full, that mean only 12GB is directly available if the cell are rate at 5000 refresh. Thats 60,000GB, or 32GB a day on average for 5 years.
And BTW, this is conservative because entire block need refresh even if its a partial write. so for small file you can half that number.
If you have a small SSD that is almost full and do allot of IO... the life of the drive might not be as long as you wished.
(Smart controller might do data relocation, but its unclear if this does happen)
absurdity wrote:I think the assumption here is that SSDs are something special, and they're really not. Regardless of how good they are now, the industry will continue to innovate and in 10 years our current SSDs will be clunky and either thrown away or stored in a box somewhere, never to be used.
Jakubgt wrote:absurdity wrote:I think the assumption here is that SSDs are something special, and they're really not. Regardless of how good they are now, the industry will continue to innovate and in 10 years our current SSDs will be clunky and either thrown away or stored in a box somewhere, never to be used.
Hell, in 10 years we'll be thinking "How did we manage with these things?" We'll keep thinking of the days we used our slow, low capacity, and wearing SSDs.
sschaem wrote:Only the unused space can be used to spread write data. So if you have a 64GB drive and its 80% full, that mean only 12GB is directly available if the cell are rate at 5000 refresh. Thats 60,000GB, or 32GB a day on average for 5 years.
flip-mode wrote:sschaem wrote:Only the unused space can be used to spread write data. So if you have a 64GB drive and its 80% full, that mean only 12GB is directly available if the cell are rate at 5000 refresh. Thats 60,000GB, or 32GB a day on average for 5 years.
Are you taking wear-leveling into account? Doesn't seem like you are.
sschaem wrote:I assume that all drive do dynamic wear leveling, but do you know what drive do static wear leveling ?
just brew it! wrote:If your system is hitting the page file at all, and your swap file is on the SSD, there are a lot of writes going on behind the scenes.
Microsoft wrote:In looking at telemetry data from thousands of traces and focusing on pagefile reads and writes, we find that
•Pagefile.sys reads outnumber pagefile.sys writes by about 40 to 1,
•Pagefile.sys read sizes are typically quite small, with 67% less than or equal to 4 KB, and 88% less than 16 KB.
•Pagefile.sys writes are relatively large, with 62% greater than or equal to 128 KB and 45% being exactly 1 MB in size.
just brew it! wrote:If your system is hitting the page file at all, and your swap file is on the SSD, there are a lot of writes going on behind the scenes.
notfred wrote:So from all the stories I've read my data should have been fried by now, but it still runs quite nicely. I do have backups in case...