Picking A Hard Drive For Your NAS: New Green Beats Old Speed

Features
By Marcel Binder
published

Test Setup And Results

As in our previous NAS tests, all measurements were taken with jumbo frames disabled in our Gigabit Ethernet LAN. Our NAS used firmware version DSM 2.1-0844 for all benchmarks.

Swipe to scroll horizontally
System Hardware
MotherboardAsus P5E3 Deluxe, Rev.1.03GIntel X38, BIOS: 0810 (02/11/2007)
CPUIntel Core 2 Duo E6750 (65nm Conroe core) @ 2.26 GHz
RAM2 x 1024MB Crucial Ballistix DDR3-1600
eSATA ControllerJMicron JMB363
System DriveSeagate Barracuda 7200.9, 160GB7,200 RPM, SATA/300, 8MB Cache
NAS Drives4 x Samsung Spinpoint F2 HD103SI, 1,000GB 5,400 RPM, SATA/300, 32MB Cache4 x Samsung Spinpoint HD321KJ, 320GB7,200 RPM, SATA/300, 16MB Cache
Optical DriveSamsung SH-D163A , SATA/150
Graphics CardGigabyte Radeon HD 3850 GV-RX385512HGPU: 670 MHz Memory: 512 MB DDR3 (830 MHz, 256 Bit)
NICMarvell Yukon 88E8056 PCI-E Gigabit Ethernet Controller
AudioOn-Board
Power SupplyCoolerMaster RS-850-EMBA, ATX 12V V2.2, 850W
System Software and Drivers
Operating SystemWindows Vista Enterprise SP1
DirectX 10DirectX 10 (Vista default)
DirectX 9Version: April 2007
Graphics DriverATI Radeon Version 7.12
Networking Driver9.0.32.3 (Vista-Standard)
Intel Chipset DriverVersion 6.9.1.1001 (20/02/2008)
JMicron Chipset DriverVersion 1.17.15.0 (24/03/2007)

You can find additional Benchmark graphs in our Image Gallery.

There is hardly a difference between the drive models where write speeds are concerned.


Aside from its lower power consumption and heat dissipation, Samsung’s newer drive, the Spinpoint F2 EcoGreen HD103SI, enjoys other benefits as well. For example, despite its lower spindle speed, it can read data faster than its older sibling, the Spinpoint T166 HD321KJ.

Marcel Binder
Marcel Binder
28 Comments Comment from the forums
  • Is it a good comparison between 320 GB drives and 1 TB ones? Of course de 1 TB will perform better: much higher data density means less movement of heads and platters for the same amount of data. Gotta be faster. Nothing to do with the enclosure.

    Also I would like to mention from my own experience that heat can be a problem. I have a D-Link DNS323 and it cannot handle two server edition WD drives - it overheats.

    - Bertus.
    Reply
  • evongugg
    If you buy a NAS, it's better to use the recommended drives, which have been put through hard tests.
    Other drives may experience RAID problems, burnout problems, etc.
    If you want to salvage a NAS hard drive, you may be dealing with a Linux type partition.
    Reply
  • Weird concept to compare an "old" performance drive with a new eco drive, where by default the new eco drive is outperforming the old performance drive in all categories.
    What is the expected result? Yes, we guessed right, the new eco drive is outperforming the old performance drive in a NAS environment.
    Unfortunately that comparison does not allow any conclusion whether a NAS needs a performance drive or not, it only tells us that a new eco drive is better than an old performance drive.
    Sad...
    Reply
  • awaken688
    Totally agree with "theaxe" on this one. Not sure what this proved other than if you choice is an old performance drive vs. a new green one, choose the green one. Show me the WD 1TB Caviar Black vs. an equivalent "green" 1TB drive and then we can see more of a comparison that means something.
    Reply
  • cknobman
    good thing we have the comments section so I dont wast my time reading a pointless article. first thought is "why in the f would you compare a 320gb to a 1gb?" reading the comments let me know my thought is right and not to waste time reading a pointless article.
    Reply
  • chunkymonster
    Our comparison showed that the difference in write performance between faster and slower spinning hard drives is minimal to nonexistent when they are used in a NAS. Instead, the NAS’ RAID engine becomes the limiting factor.
    Ok, I can understand the RAID engine being the limiting factor when using off the shelf NAS solutions, especially considering most off the shelf NAS solutions use software RAID. But, would a 5400RPM drive become the limiting factor, compared to a 7200RPM, when using a hardware RAID solution like an Areca, 3Ware, or HighPoint controller card? Any guesses?
    Reply
  • davidhbrown
    I've been reflexively buying 7200RPM drives for everything (notebooks, NAS, desktops) for so long now, that just the *concept* of the article is a useful reminder that in some settings, other issues might be important. Too bad the specific comparison may be flawed, but it's not like exactly that hardware is going to be relevant in six months anyway.
    Reply
  • Even with a 10 watt difference, the electricity savings is negligible. At the nation average of 12 cents per kWh, using 10 less watts 24/7 for a year works out to be a whole $10.50. On the year. Less than $1 a month. A far cry from what the article stated: "the difference jumps into the two-figure range. And that can definitely make a difference when the time comes to pay your electric bill" I'll take that 88 cents I save a month and go buy...wait, what can you get for 88 cents these days?
    Reply
  • farrellj
    Just one thing...many consumer NAS boxes run a cut-down form of Linux...and Samba! And that makes it much more stable than some proprietary software that hasn't been as extensively tested as Linux/Samba has.
    Reply
  • farrellj
    Just found out...the Synology DS409+ does use Linux!
    Reply