Can we create a true random generator whose entropy source is the loss of electrical signals that a binary architecture cannot detect?

If we have a system that let's say could do this, could we be able to encrypt data in a way that makes people unable to decrypt with the current resources? Of course, as far as security and encryption are concerned it all depends on time. If such a system does exist, could you please point me out as I cannot find it?

  • What you describe is a bad key-gen for One-Time-Pad. Nothing more!
    – kelalaka
    Sep 19, 2022 at 12:46

2 Answers 2


... could we be able to encrypt data in a way that makes people unable to decrypt with the current resources?

Surely one can "encrypt" data so that they cannot be decrypted, like simply replacing the data with random noise. But what is the use case for this? The goal of encryption is to make data available for only specific selected users and nobody else. The goal is not to make data unavailable for everybody.

A true random generator or at least a cryptographically secure random generator (which is not necessarily "true random" but at least sufficiently random) is often used in the context of encryption, like for creating encryption keys. This needs to be sufficiently random and without any bias, so that attackers could not reduce the key space due to broken randomness, which would make brute force attacks faster. See also Predictable random number generator discovered in the Debian version of OpenSSL for an example.

Of course these random keys then need to be shared in a secure way between the users who should be able to decrypt the data. That's were key exchange algorithms fit in.


Two issues:

  • Encryption must be reversible
  • While erasing random bits is obfuscating, it is not encryption. Think of a magazine page where someone cut a few letters out of it. One could probably still read/puzzle out the original text with effort.

So this process would not only need to have a physical property by which bits are erased, but a paired process by which the same bits are restored. That is hurdle one.

Second, output of encryption should be statistically random. If one flipped a coin 50 times, wrote down Heads or Tails for each and then randomly chose a few heads to change to tails, this would skew the distribution to tails. To be random, the process would have to change around as many tail to heads as it changed heads to tails.

So to pull off a encryption by analog means, one would need a process to flip random bits (not just erase), it would need to be reversible (ciphertext goes in and the same bits are magically restored) and that process would need to be repeatable between two participants.

I'll admit, encryption based on analog (even quantum) principles is attractive but I'm not quite bright enough to image how it might be implemented.

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .