Comparing eSATA To USB For External Drives

Posted on April 27, 2008 by Aaron Parker

HardDiskPerformance I’ve just purchased a new SATA-based external hard drive to use with demos and I thought I would share some details about the performance gain over my older ATA-based hard disk.

I’ve been doing a few demonstrations of Microsoft SoftGrid using VMware Workstation on my laptop with an external USB attached hard drive and I’ve also got a couple of user group presentations coming up. So I wanted to squeeze some better disk performance out of this setup because every second counts during demos.

Before I read Mark’s post about USB attached hard drive performance in his notebook, I was hoping to add a second internal hard drive to my Dell laptop, but unfortunately Dell doesn’t provide a media bay type disk enclosure for this model. I was looking into using Firewire until I considered eSATA. I found this review of the Belkin SATA ExpressCard very useful when looking at what I might expect from eSATA.

Here’s what I ended up ordering:

To power this drive I have to connect the included USB power lead. I’m not sure if I used a better quality eSATA cable that it would provide enough power, but carrying the extra cable is not too much of a hassle. I’ve performed these tests on my laptop which has the following hardware:

  • Dell Vostro 1400
  • Intel Core 2 Duo (2.2GHz, 4MB Cache)
  • 4GB RAM
  • Windows Vista x64

To test the performance of my disk setup, I’ve used HD Tune because they offer a free version and it’s easy to use. I tested the performance of my original hard disk (Seagate Momentus 5400.3 80GB ATA/100 5400RPM 16MB 2.5") connected to this laptop via USB. The result stays consistent right up until the end of the test:

Orignal disk performance over USB 

When I performed the same test on a desktop machine the throughput was about 5MBps higher and was consistent to 100%. This is the performance of the new hard disk over a USB connection:

New disk performance over USB 

And the performance of the same hard disk over an eSATA connection using the Belkin ExpressCard:

New disk performance over eSATA 

Overall I’m pretty happy with the performance of the new disk. It more than doubles the performance of my older disk yet still in a portable package and my VMs feel much much snappier.

Original (USB)New SATA (USB)New SATA (eSATA)
Minimum transfer rate21.4MBps27.8MBps33.7MBps
Maximum transfer rate25.2MBps29.5MBps65.4MBps
Average transfer rate24.8MBps29.2MBps52.3MBps
Access time16.5ms15.1ms14.6ms
Burst rate19.6MBps22.9MBps75.1MBps
CPU utilisation13.6%17.1%3.4%
This entry was posted in Hardware and tagged , , . Bookmark the permalink.
  • andy

    I’m looking for a similar product. I think that using my LAN cable will be faster than eSATA. If so I would like a portable NAS device that’s usb powered.

    Any takers?

  • http://stealthpuppy.com Aaron Parker

    At some point that data has to move through a SATA connection, so unless I’m completely wrong, I don’t see how a network connection will make it faster.

  • Jan Ivar Beddari

    Agree with Aron, eSATA should be faster and easier, having read about the new WD VelociRaptor it would be interesting to try that connected to an eSATA expresscard :-)

    Only thing that worries me abit is the fact that you need the extra usb cable for power? Anyone know if this is true for all, it may be the expresscard slot simply does not have enough juice?

  • Josh Straub

    Awesome post – this is exactly the information I was looking for.

    I just got a StarTech USB2SATAIDE converter and tested it with a high speed SATA drive, and was disappointed to see it HDTune bench at about 27MB/sec which my old laptop hard drive did.

    Now I understand that 27-30MB (MegaBytes)/sec is about the absolute limit for USB 2.0. Its a shame.

    But obviously eSATA unlocks the full performance potential of the hard drive, as indicated in the proper curve of your HDTune benchmark.