Stand-Alone Devices

[BitWise]

Sorry but an off-line computer with wireless capability is by definition online at least some of the time. Even a truly off-line computer is vulnerable to a 'sneaker-net' attack if its users transfer applications or data files to from another online device via some storage medium like a flash drive.

The world uses public key encryption because its convenient and reasonably secure. If you publish a public key then anyone that accesses it can send you a secure message that other downloaders of the same key cant use to decode it. Only the holder if the private key can do that. Provided you make the public key big enough that its computationally impossible to crack you're safe.

I agree that encryption with shared secrets is usually uncrackable. Spies have been using one-time pads and book ciphers for centuries for exactly that reason but it requires that the secrets are pre-agreed before communication starts and that's a pain. A lot of the business messages I get come from people I have never met before and may never talk to ever again. Managing secrets for each of them would be a real PITA.

Early computers were more resilient to hacking because they were so much simpler. As software has got more complicated and users have got dumber they have become riddled with security holes, but do you really want to go back to text only messages, no Google and fixed pitch fonts in all your printed documents?

If governments force the removal of end-to-end encryption then the bad guys will just send encoded message over open channels. They will just fall back on the old ways of doing things, like in the 1950's thrillers. The Kremlin apparently has kicked out some of its computers and replaced them with manual typewriters.

'The moon is big tonight in Moscow, Comrade' (trans. Leave the money behind the big rock).

[EugeneNine]

I wouldn't say the earlier computers were resilient to hacking because they were so much simpler. All those cracked commodore programs were the hacking that happened then. The boot sector virus's, etc The hacking today is just a different attack vector but really the same thing.

[whartung]

jamesadrian wrote:

Perfect encryption is not just perfect security, it is the simplest know encryption method. It demands the least of the central processor, (which ever one you choose).

The one time pad has been around forever, and is in use today. The problem is managing key exchange (i.e. exchanging the one time pad).

Quote:

There are three tactics used in encryption algorithms. complexity, obscure mathematics, and shared secrets. Hacking has become a growth industry because shared secrets are rarely used.

Shared secrets are ubiquitous and used everywhere. They're commonly referred to as "passwords".

The issues with encryption center around workflow and convenience, not encryption algorithms.

Quote:

Implementing encryption in a simple device that is ASSURED to be off line is an enormously appropriate and timely application for the 6502.

Having an "off line" encryption device offers little value to anyone, as the simple fact that it's "offline" means it won't be used. People do not use handy "local" (to the machine) encryption options.

Convenient, powerful, secure encryption is ubiquitous today. You may have noticed a few folks in the assorted world governments commenting on this phenomenon.

The modern 6502s are not adequate for the algorithms in wide use today, simple because of the key sizes and data volumes in use today. For encrypting a single message, sure. But at the scale that encryption is done today? No. (I have several TBs of data encrypted on my workstation for example, and the GB's of data on my phone are all encrypted as well, not to mention the PBs of data encrypted across the internets daily).

[jamesadrian]

I suggest everybody read the article before committing themselves to a position.

https://www.futurebeacon.com/perfectencryption.htm

Shared secrets are no longer inconvenient.

There seems to be a misunderstanding. I am proposing that the new device have no ability to go online; that the encryption occurs on that device; that the encrypted message be brought to an online computer by means of a flash drive; and that the online computer never contains a plain-text message.

You should also be aware that there is no such thing as reasonable security. An encryption algorithm has either been cracked or it has not. The people who are hacking are not suffering from idol curiosity. They want the technical insights of entrepreneurs. The law concerning patents has been changed so that the recipient of the patent is the first to file, not the first to conceive. This is convenient for those who can afford a server and a few employees.

The device I propose will enhance upward mobility, which has been very poor since 1975.

To fully understand the magnitude of this opportunity, there are other kinds of facts to consider. I recommend investigating CIA influence through television since 1947.

Your private questions will be answered.

[dwight]

jamesadrian wrote:

I suggest everybody read the article before committing themselves to a position.

https://www.futurebeacon.com/perfectencryption.htm

I read the article but didn't find any part of it that was not already
known.
Most encryption is bases on large prime numbers, these are called
trapdoor type. You can easily go one way but the other is a lot tougher.
PGP encryption is based on this.
They believe that the quantum computers with make these types obsolete.
Truly random keys are used for some types of encryption already.
They still have the risk of the key is being compromised while being
transported.
Such encryption can not be broken without the correct key.
There are trivial ways to create truly random values.
Dwight

[jamesadrian]

dwight,

Use any encryption you like, but the National Institute for Science and Technology (NIST) does not write a standard for an encryption method until it is cracked.

A long time from now, quantum computing will make obsolete those encryption methods that are already made obsolete by now-convenient methods based on shared secrets.

It is not the under-used perfect encryption which will be new. It is the low-cost second computer that CANNOT BE ONLINE that will be new.

Those who join the effort will discover that even the NSA will be pleased by our business plan.

This has to do with liberty and upward mobility, not hiding criminality.

Please tell me more about generating random numbers.

[whartung]

jamesadrian wrote:

dwight,
Use any encryption you like, but the National Institute for Science and Technology (NIST) does not write a standard for an encryption method until it is cracked.

This is tin foil black helicopter nonsense. Bruce Schneier and his bevy of Cypherpunks, cryptography researchers, white hack, grey hat, and black hat hackers, have been, are still, and will be all over the algorithms and exploits of open cryptography systems. Everyone invested in that community are motivated by its integrity. Not all of these people work for Governments, not even necessarily the best people.

The state actors have a vast array of other approaches more effective, and far cheaper, than simply cracking cryptography. As we've already stated, perfect cryptography exists and is used today by those who wish to use it as long as they can get the key transferred and kept out of the hands of those prying eyes they're trying to protect it from. https://xkcd.com/538/

When the FBI arrested Ross William Ulbricht, operator of the Silk Road, a famous marketplace of illicit goods on the "Dark Web", one team went for him while another seized his computer while it was open and booted. This conveniently threw all of his hard work surrounding cryptographic systems out the window.

Quote:

It is not the under-used perfect encryption which will be new. It is the low-cost second computer that CANNOT BE ONLINE that will be new.

Funny, the Iranian nuclear centrifuge control systems were air gapped and never online either, yet Stuxnet still managed to silently destroy them while all of their monitor panels read "green".

[Tor]

My gut feeling is that this thread is derailing with that (repeated) link offsite.
Any possibility of getting the thread back on-track again, and talking about 6502 computers? No 6502 that I'm aware of can handle modern encryption loads anyway.

[jamesadrian]

Tor wrote:

My gut feeling is that this thread is derailing with that (repeated) link offsite.
Any possibility of getting the thread back on-track again, and talking about 6502 computers? No 6502 that I'm aware of can handle modern encryption loads anyway.

Tor,

This is the simplest algorithm. No history or keys are stored in it. The keys are in external flash drives. One will last you a long time.

This is about the 6502 in an application that will do everybody a lot of good. What could be more on-topic than that?

There is also a sound-file application. From the beginning, I have been advocating a small machine with a USB post. Can anybody here do that?

[BigEd]

I think as soon as you add a USB port then you will have some kind of controller for it. It then becomes an interesting choice not to use that controller as the main CPU! As a hobbyist, I'd do it for the interest, but for a product, I'm not sure how to justify using two, three or four chips when one would do.

[BigEd]

(Actually, reading and writing low-capacity SD Cards is something which a 6502 can do unassisted. So, if you choose carefully, you can have swappable solid-state storage without a controller chip. Carry on!)

[Alarm Siren]

(Notwithstanding that an SD card itself contains a controller chip. I've read about MicroSD cards which contain fully fledged ARM cores to implement the interface. Obviously it'd be a pain in the bum, but I've always wondered if it would be possible to reprogram an SD card's controller to make it into a general purpose computer - with a LOT of built in flash!)

[BigEd]

(I think it may be possible! And it's a good point - there are very few dumb devices and very few whose implementations could be known. So you'd need to know what your threat model is, to decide who and what to trust, and how much, before embarking on a system design.)

[dwight]

There are several methods. They mention using a radioactive decay but that
requires something that is radioactive.
With a geiger counter tube you can also do cosmic ray counts.
You run a counter that counts between count hits. The time between
is random.
Another way is to use something that generates white noise. A 7.2v zener
will do that if biased right at the edge of the zener voltage. As silicon
transistor can also be used with reverse bias on the base to emitter.
You tap off some random part of the frequency with a receiver. The noise
will be random.
You have a really fast counter and a button for the person to stop that count.
You use the LSBs of the counter.
There are others as well.
Dwight

[sark02]

A product requiring a display, USB, mass storage, keyboard falls out of the practical profile of a 6502, I think. With that level of complexity, although technically achievable, the 6502 is a force-fit. A Raspberry Pi Zero (no wireless) has the requisite I/O for $5. You might begin to put a discrete 6502 design together, with the required I/O for 20x that; but I doubt it.

I don't see a market for an encrypting typewriter where you have to take the encrypted message and send it via some other connected machine. Maybe I just don't live in the same circles as James and so don't see the need. I certainly don't see the liberty and upward mobility aspect... unless this is in the context of subversives wanting to securely communicate; in which case I suspect there are already established tools... Or are the armies of change waiting in the shadows for perfect untraceable, uninterceptable communication?

The morning goose sees three marks on the red panda. Two three one two seven zero four zero zero seven two five one zero two six three.