Interesting comment on a question re Seagate Ironwolf drives on Reddit data hoarders forum. It seems to spell it out well.
level 1
stoatwblr
8 points ·
15 days ago · edited 15 days ago
There's no such thing as "Parallel Magnetic Recording"(+)
BEWARE - HERE BE DRAGONS
- The preferred marketroid term for "normal" drives is now "CMR" - Conventional Magnetic Recording. BE VERY CAREFUL TO USE IT WHEN ASKING QUESTIONS
- PMR stands for "Perpendicular Magnetic Recording" (the way the magnetic domains are packed on the platter, vs Longitudinal magnetic domains of yore)
- Shingling (SMR) is Shingled PMR. It's built on top of PMR
- TGMR is a head technology needed for PMR and is the same for SMR or CMR
- TDMR( Two Dimensional Magnetic Recording ) is a weaselly way of saying "zoned media" and the only reason you have zoned media is.... because it's SMR (you need the zones to be able to do rewrites)
- Helium does not imply CMR or SMR. You can get squeaky drives in both formats,- HOWEVER you can infer probable SMR for air-filled drives above 4TB/3.5" 2TB/2.5"
So:
- If you are told a drive is PMR, therefore not shingled.... BZZT: see above.
- If you are told a drive is TGMR technology, therefore not shingled, see above
- If you ask "is this drive SMR or PMR?", the correct answer is "yes"
(All mechanical drives on the market use PMR, whether CMR or SMR and regardless of speed. This is because all 3 makers use the same platters and heads - each made by only one manufacturer)
Some other points:
- Drives whose datasheets show them as being substantially lighter than their predecessors are likely to be SMR
- Drives with large cache (256MB+) are extremely likely to be SMR - they need it to work (there are data tiering functions in Drive Managed SMR drives (DM-SMR) akin to those found in TLC/QLC SSDs - but much MUCH slower....)
- Drives reporting "trim" functions _ARE_ SMR, even if they don't report zoned media (Reporting zoning is a ACS-4 function - whilst that was codified in 2016, many drives are still only ACS-3 which doesn't support this reporting function.... hmmmmm... I wonder why that might be? Nothing suspicious to see here citizen, move along!)
- HOWEVER - Not all SMR drives report trim functions and the only way to find out is to benchmark the things - (Seagate barracuda ST3000DM003 being a good example)
- PAY ATTENTION TO THE SUFFIXES, READ THE SPEC SHEETS.
- COMPARE WITH PAST MODELS
- a generational change (from 003 to 004, or from EFRX to EFAX) may well mean a change has been submarined into the channel. Check the drive mass. Losing a platter at the same size is a sure sign of it having happened.
Above all: Complain loudly to your local marketing regulators about this misleading behaviour. WD and SG KNOW we don't want SMR, which is why they're going out of their way to keep it off their datasheets (European and Chinese regulators should be especially receptive to fraudulent marketing and cartel behaviour complaints)
(+) see the IEEE article below. The only references to "Parallel Magnetic Recording" anywhere on the web are on review sites and it's 100% clear they're a result of an utterly clueless reviewer guessing what the initials meant, then that guess spreading meme-style amongst other cluess reviewers who apparently don't know how to read tech documents (ie: You can pretty much surmise that such reviews are paid-for, as anyone independent will be doing their homework - and murketers aren't going to be falling over themselves to correct such an error)
(there ARE reference to lab work for recording calibration and sync tracks in parallel, but what THAT means is that they're activating multiple heads at once instead of the normal sequential switching between individual heads. (Attempts at parallelising heads in-service have never worked well due to differing thermal expansion of platters due to slight temperature changes across the case internals))
EDIT: This IEEE article is worth reading.
https://ieeexplore.ieee.org/stamp/stamp.jsp?arnumber=4694034