Hi,
Just trying to put together a basic layman's guide for what SSDs are, I am by no means any where near being clued up on the subject, but I'm the guy at work who seems to know more about them than anyone else, due only to the fact I've owned one
Anyway! My question is, am I right in saying that cells, are read/written to in blocks of 4KB (called pages?) and erased in blocks of 256KB? As a generic typical figure.
Now am I correct in saying that if one/few page/block fails (their P/E cycles are exceeded), does that mean the drive is dead entirely?
Is this why wear levelling is so important in SSDs? To make sure all cells/pages/blocks are given as much an equal P/E cycle count as possible so as to try and make it so that the cells fail nearly at the same time?
And to see if I have the basic operations of an SSD covered, in terms of writing to the NAND-flash.
You can wrote to individual pages, made up of multiple cells, or 4KB Pages?
You can only erase blocks, which contain multiple pages (so maybe 64 pages of 4KB = 256KB block?)
Data can't be overwritten on a page/cell scale, until a block is filled, then it can be erased?
So with Garbage collection/Trim, say you overwrite a bunch of data in the OS, the SSD copies the pages with valid data to a new section, erased/free pages/blocks etc. then Trim will erase the blocks that the data has come from?
It's the interaction between an everyday use of an OS and how the data is moved about and how things are erased which seems to go over my head.
I've read up on Logical Block Addresses, so this is the means by which things are tracked, the SSD links pages/blocks with LBAs so the OS knows where the data is?
So in terms of SSDs and LBA, if you delete/overwrite a file, being NAND-flash, the data is copied to a new unused cell/page/block with the new data and the LBA is updated.
I apologise to any experts or enthusiasts who read this and simply face palm
I've been reading for about 3 hours now and I shall probably be reading for several more.
Just trying to put together a basic layman's guide for what SSDs are, I am by no means any where near being clued up on the subject, but I'm the guy at work who seems to know more about them than anyone else, due only to the fact I've owned one
Anyway! My question is, am I right in saying that cells, are read/written to in blocks of 4KB (called pages?) and erased in blocks of 256KB? As a generic typical figure.
Now am I correct in saying that if one/few page/block fails (their P/E cycles are exceeded), does that mean the drive is dead entirely?
Is this why wear levelling is so important in SSDs? To make sure all cells/pages/blocks are given as much an equal P/E cycle count as possible so as to try and make it so that the cells fail nearly at the same time?
And to see if I have the basic operations of an SSD covered, in terms of writing to the NAND-flash.
You can wrote to individual pages, made up of multiple cells, or 4KB Pages?
You can only erase blocks, which contain multiple pages (so maybe 64 pages of 4KB = 256KB block?)
Data can't be overwritten on a page/cell scale, until a block is filled, then it can be erased?
So with Garbage collection/Trim, say you overwrite a bunch of data in the OS, the SSD copies the pages with valid data to a new section, erased/free pages/blocks etc. then Trim will erase the blocks that the data has come from?
It's the interaction between an everyday use of an OS and how the data is moved about and how things are erased which seems to go over my head.
I've read up on Logical Block Addresses, so this is the means by which things are tracked, the SSD links pages/blocks with LBAs so the OS knows where the data is?
So in terms of SSDs and LBA, if you delete/overwrite a file, being NAND-flash, the data is copied to a new unused cell/page/block with the new data and the LBA is updated.
I apologise to any experts or enthusiasts who read this and simply face palm
I've been reading for about 3 hours now and I shall probably be reading for several more.