That study is a bit dated and was somewhat controversial when it came out.
Dated: 2005-2006 - PATA and 1st gen SATA drives from 80 to 400 GB were monitored. Server grade SATA drives didn't exist back then and the price differential per GB with flavours of SCSI and SAS drives was huge. Their results were inconsistent with everyone else when it comes to temperature, they didn't look at vibrations, etc...One of their very strong confirmations was the fact that a drive that gets a single scan error is 39 times more likely to fail within 60 days than a clean drive.
Anyway, here's a much better paper (much bigger sample, less uniform use and conditions) - look at table 1 (summary of findings) and especially figure 10 if you don't have time for the full paper.
http://www.cs.wisc.edu/adsl/Publications/latent-sigmetrics07.pdfHere's a recent tidbit from France, with the 1st year failure rate at the service department of a big e-commerce reseller.
Failure rate in 1st year for 1 To drives
- 5,76% : Hitachi Deskstar 7K1000.B
- 5,20% : Hitachi Deskstar 7K1000.C
- 3,68% : Seagate Barracuda 7200.11
- 3,37% : Samsung SpinPoint F1
- 2,51% : Seagate Barracuda 7200.12
- 2,37% : WD Caviar Green WD10EARS
- 2,10% : Seagate Barracuda LP
- 1,57% : Samsung SpinPoint F3
- 1,55% : WD Caviar Green WD10EADS
- 1,35% : WD Caviar Black WD1001FALS
- 1,24% : Maxtor DiamondMax 23
Failure rate in 1st year for 2 TO drives
- 9,71% : WD Caviar Black WD2001FASS
- 6,87% : Hitachi Deskstar 7K2000
- 4,83% : WD Caviar Green WD20EARS
- 4,35% : Seagate Barracuda LP
- 4,17% : Samsung EcoGreen F3
- 2,90% : WD Caviar Green WD20EADS
It actually correlates well with my experience with flash cards: the latest/highest capacity models always have a much higher failure rate than their lower tech/density siblings. Note: I can't wait to see numbers for the recent hybrid SSD/HD from Seagate.
So basically, the generic advice would be
- never buy the latest generation in storage.
- swap the drive/media as soon as it starts developing an issue. (the problem for a non techie is 1 - to notice the issue 2 - to make the distinction between a hard error and something he simply doesn't understand or did wrongly)
- enterprise HD are significantly more reliable long term than their counterparts (per that study).
- brands are unreliable indicators as a whole. Issues are mostly model dependents.
The EADS drives seem to be currently the best choice in non enterprise drives. They are however rated for 300.000 load cycles compared to 600.000 load cycles for Enterprise drives
Take a look at stats from some of my systems
Modèle: WDC WD10EACS-00ZJB0
Durée de fonctionnement (heures) 8352
Nombre de cycles de chargement 33537
-> consumer drive: 4 load cycles per hour of operation
Modèle: WDC WD2003FYYS-02W0B0
Durée de fonctionnement (heures) 808
Nombre de cycles de chargement 288
-> enterprise drive: 0.3 load cycles per hour of operation
If you take the life expectancy at face value on the basis of cycle counts, the difference is 24x
As all studies say, no single predictor is best, still...