Following is a sentence from the book "Cryptography and Network Security - Principle and Practices" by William Stallings [Fourth Edition]

Sentence - " ... the 'one-time pad' is of limited utility, and is useful primarily for low bandwidth channels requiring very high security. "

Can anyone please elaborate on low bandwidth , because I thing low bandwidth will make transfer of the key and ciphertext slower instead they should use high bandwidth isn't?

Book Link - http://uru.ac.in/uruonlinelibrary/Cyber_Security/Cryptography_and_Network_Security.pdf Page Number - 79/983


For an OTP you need to distribute the OTP key stream out of band. That can be performed by e.g. bringing a data storage device to the other side. Let me explain.

The problem with an OTP key stream is that you cannot transport it over a secure transport channel without removing the claim of theoretical security. Similarly, you cannot simply distribute a seed for a PRNG, because the PRNG would have similar security as a stream cipher.

Because you cannot distribute an OTP without compromising its security it is limited to the amount of information that you distributed out of band.

How limiting this is depends on the application of course. Nowadays you can store several TB in a tiny flash disk. That's enough for many, many hours of video. Still, securely distributing such a disk would be troublesome; just sending it by regular post is obviously not a good option.

Another option would be to use quantum key distribution. However, implementations of quantum key distributions are not necessarily secure, and the speed may be a severe limit.

Fortunately, we've got a lot of ciphers that we don't know how to break, key ratchets and whatnot to give us practical security even for much smaller key sizes.

  • $\begingroup$ "The problem with an OTP key stream is that you cannot transport it over a secure transport channel without removing the claim of theoretical security." I means we can asymmetric key to solve this right? Honestly I am novice started crypto course recently ,if u can add little more context to your answer, that would be really helpful. $\endgroup$ – jiya singh Mar 1 at 14:51
  • 1
    $\begingroup$ Asymmetric cryptography is by definition not theoretically secure. We can transport anything with unproven, practical security. An OTP is random, so you cannot find (much) about an encrypted OTP key - until you use that OTP to secure plaintext. In that case the security of the OTP degrades to the security of the transport protocol. If you want practical security you generally don't need an OTP - you'd use a stream cipher to fill the same requirements. $\endgroup$ – Maarten Bodewes Mar 1 at 16:52
  • 1
    $\begingroup$ @jiya singh: the One Time Pad assumes its pad is (1) perfectly random (2) somewhat made securely available to both parties (3) at least as large as the data encrypted (4) used at mot once. If (2) is achieved by cryptographic means, the security claims falls. (2) used to be achieved by transporting punched paper tape in a sealed envelope physically carried by a diplomat. Nowadays, we could use an USB key. But to secure a high-bandwidth link (say 10 Tbit/s, which is feasible for a fiber optic bundle), we'd need in the order of one USB key every second of use. That's a logistic issue. $\endgroup$ – fgrieu Mar 1 at 20:09
  • $\begingroup$ @fgrieu what, you don't always use a low orbit USB drive cannon in conjunction with your high-speed fiber-optic links? $\endgroup$ – user253751 Mar 3 at 18:27

Let me re-quote:-

" ... the 'one-time pad' is of limited utility, and is useful primarily for low bandwidth channels requiring very high security. "

This raises four issues that divide opinion:-

  1. What is low bandwidth? No one knows, yet OTPs are making quite a resurgence these days for things like secure phone communications and video conferencing systems, both detailed in this answer. That seems fast to me, and they're just getting faster every day. Channels already operate at Mbps. The raison d'être of a OTP is for ultra secure strategic communications, typically in textual form. It is not to encrypt 8K UHD movies. That is your limited utility.

  2. The security is the highest theoretically possible given a correct implementation. One time pad material is securely transmitted theses days in-band using quantum key distribution networks (QKDNs). See this answer as to how, and how theoretic security is maintained based on quantum physics. I have to reiterate, please do not conflate the theoretical security of a OTP/QKDN with how it can be attacked if someone is looking over your shoulder. That's a common denominator that undermines all cryptography (AES etc.) not just OTPs/QKDNs.

  3. There is an every growing market for OTPs distributed via QKDNs. Many organisations, nations and trading blocks are convinced that they're suitable and practical. This is a OTP distribution device machine thingie.

  4. OTPs were deemed suitable for use in three world wars, which you can explore here and here.

And yes, you can pre-distribute OTP material on a flash drive provided you have a means of making the random numbers in a truly random manner. That's not hard, and getting easier. And yes too, you can send it in the post provided you do it before your operation gets onto the radar. A meeting would be preferable, but that's almost implicit as if you're going to exchange ultra secure material, you might want to meet the recipients.


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.