ISTR back in the early days of (build your own) PCs having an old 486 PC with two hard drives, and hacking the "Turbo" switch on the front panel (remember those?) to be a read-protect switch for drive D - i.e. when the switch was one way the hard drive became like a CDROM, and you flipped it the other way when you wanted to write to it.

OS on the r/w C drive, all your data on the write-protected D drive.

Granted in those days drives were IDE 40-way ribbon cables with a separate R/W line, but surely this techniqe is implementable in SATA?

  • I believe most modern SD cards still have these switches, or at least all of mine do. Apr 18 '16 at 15:04
  • 3
    There is no specific wire which is only used for writing data and thus a simple hardware solution can not be done. Apr 18 '16 at 15:09
  • It's probably not a technical issue, more human nature. Hard drives are typically locked away in the insides of a computer and 99% of computer users never see them. Also, you can achieve the same thing with software easily and much more refined (file/folder permissions, ownership etc.). So there really are not enough reasons to support a hardware switch. I'm sure it could be done.
    – Potaito
    Apr 18 '16 at 15:28
  • @SteffenUllrich I imagine there would be a hardware way to actually limit moving the arm (assuming it's an HDD) from reading and writing, but not selectively limiting only one. Apr 18 '16 at 15:31
  • 1
    @WorseDoughnut the SD card switch is a request to not write to the card, not an actual block. See e.g. this question
    – user15392
    Apr 18 '16 at 15:44

SATA has no separate wires for reading and writing so a pure hardware solution is not possible.

You actually have the same problem with Ethernet too: with Fast Ethernet you had separate Tx and Rx (Transmit and Receive) wires and thus could simply built a guaranteed passive network tap by not connecting the Tx lines. With Gigabit Ethernet this is no longer possible so network tapping at this speed will be controlled by software which makes it harder to guarantee that the tap stays passive all the time.

  • 1
    Note that not plugging all wires with 100baseT is nominally out-of-spec and may induce reception trouble -- unconnected wires are not good at absorbing electromagnetic noise. If you remove the Tx lines the cable may not work, or might fall to 10 Mbits/s. It may also prevent link detection from switches. Apr 18 '16 at 15:44
  • @ThomasPornin: in theory you are right. In practice it works usually well. See also Throwing Star LAN Tap. Apr 18 '16 at 15:46
  • 1
    As someone who once replaced a piece of ethernet cable with the core of a screwdriver, I can confirm that ethernet can take considerable abuse. But I just wanted to warn about it in case people want to try to use such "one-way cables" for actual production use. Apr 18 '16 at 16:00

Why aren't hardware R/W switches used to defend hard drives?

Because this is a feature in search of a market. This would be a niche market, at best, because software solutions and other products (like WORM media) exist to provide this functionality already.

Bottom line, not enough people want it (and would be willing to pay enough to make it profitable) for hard drive manufacturers to spend R&D resources on it. Those resources are better spent on developing what consumers do want and will pay for - faster, bigger hard drives.


A drive manufacturer could certainly implement a read-only physical switch by modifying the drive's firmware. As the firmware could have a bug, this wouldn't be as sure-fire as disconnecting wires required for writing to the disk, but, if done carefully, will be very strong. As the implementation would be easy for a drive maker, I have to assume that they feel that there's no market for this type of product.

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.