h a l f b a k e r yThe Out-of-Focus Group.
add, search, annotate, link, view, overview, recent, by name, random
news, help, about, links, report a problem
browse anonymously,
or get an account
and write.
register,
|
|
|
This chip has built-in flash RAM which can be disabled for very long (and short) periods of time. These periods could be many years, if desired.
It is made to be virtually impossible to physically tamper with it without destroying the data nor is it possible to circumvent its protection electronically.
To ensure the failure of all physical attempts at circumvention, the chip is fragile on the inside, and therefore has a hard, durable shell plus padding on the outside.
It saves its countdown progress periodically in case of power failure. Once a delay is set:
* It requires a constant external signal to enable the count-down.
* It doesn't respond to any request for the stored data until that time period has elapsed.
* It allows restarting the current delay.
* It allows increasing the current delay.
While it's in locked mode, there is a data-integrity check available at all times so people wouldn't wait for perhaps years, just to be disappointed by corrupted or destroyed data. One can also view all of the internal settings (delay, count-down progress, password enabled, and so on...)
The counter would save the count of minutes to the flash RAM so that a complete restart would never be needed after an interruption.
This could be useful in emergencies when it's more important that absolutely no one can get the data than for the owner to be able to access it. For example, if one's computer is stolen, he could now have a chance of changing his passwords before the thief could get his original ones.
(edited)
[link]
|
| |
Just for clarity, if I set it for 2 years, immediately turn it off, then turn it on again 1 year later, do I now have to wait 2 years (for a total of 3), or just 1 year (for a total of 2)? |
|
| |
spidermother,
Yes, you would have to allow the full, preset
delay to elapse. There are no shortcuts, by design. |
|
| |
So a total of 3 years? Fair enough; that's the high security (but low convenience) scenario. Julian Assange would buy one. |
|
| |
spidermother,
It could of course, support a super-secure password which could be stored in a bank vault or something.
Certainly many would want to disable this feature for
some situations though. |
|
| |
+
Yes, make the security fully configurable to expand your market. |
|
| |
- Optional password to unlock immediatly. |
|
| |
- Optional password required even after timout expires. |
|
| |
- Configurable setting for whether the locking perid can be extended or reset with or without a password. |
|
| |
All these settings are of course stored with the same security as the data. |
|
| |
Unfortunately the actual physical security suggested sounds a little WIBNI to me. I work for a silicon chip company. When one of our chips fails unexpectedly in the field, we fairly routinly will disolve the package with some very nastsy chemicals then use a focused ion beam to cut into the circuits on the chip and lay down additional metal wires to allow us to isolate parts of the circuit for testing. This process puts practically no stress on the chip, so it will be hard to make it fragile enough to break in this case, but not break during normal use. Our chips aren't specifically designed to be physically secure, so it could clearly be made more difficult, but it would never be impossible. If the chip was designed so that when it looses power the data is destroyed, that might enable more techniques for intentionally destroying the data if any attempts to cut into the chip are detected, but just like standard encryption algorithms: it's not a matter of "if" they can be broken, but "how much time" and "how expensive" it is to do that. |
|
| |
That being said, this time-release option is not something that can be implemented in an encryption algorithm, so this could have at least a niche market. It would be much easier to implement this type of functionality as a web service, but in favor of your chip solutions, corrupting an employee at that web service might be significantly easier (and cheaper) than cutting open a chip designed to be tamper resistant. Also, the chip can be designed to continue functioning for 50 years, but there's no guarantee that the web service won't go out of business. |
|
| |
scad mientist,
All data is stored in flash RAM, and the chip only requires power when it is counting down. The count-down process can be started, paused and resumed at any time with no loss of any kind.
|
|
| |
It is expected that there would be many other special considerations involved in the design.
|
|
| |
It is also expected that there will always be some way to circumvent it, but hopefully at such a high cost that no one would bother.
|
|
| |
One of the intended uses is to allow safe storage of data in or near a computer and to never need to subject it to the dangers of transmission over networks.
|
|
| |