Skip to main content
replaced http://crypto.stackexchange.com/ with https://crypto.stackexchange.com/
Source Link

First, remark that the desired commutativity is incompatible with security under Chosen Plaintext Attack, which (under the name IND-CPA) is considered a requirement for modern encryption systems. Proof, expanded following tylotylo's commentcomment, using the IND-CPA game as played for symmetric encryption (see the CPA indistinguishability experiment in section 3.5 of Katz and Lindell's Introduction to modern cryptography, also here on page 4):

As remarkedremarked by figlesquidgefiglesquidge, XOR with a One Time Pad has the desired commutativity property, in some sense. That's also true for any Stream Cipher with an out-of-band method for synchronization (including a block cipher such as AES in OFB or CTR mode, with out-of-band IV); by out-of-band I mean: not part of the ciphertext for which the commutativity property is thought.

If $\small\text{OAEP}(x)$ designates the padding used for RSA-OAEP with a modulus of $\lceil\log_2(p)\rceil$ bits, then $x\mapsto E(\small\text{OAEP}(x))$ seems to be IND-CPA-secure, and decipherable (the desired commutativity property is lost: that padding must be external to whatever step of the protocol requires commutativity). As pointedpointed by Ricky DemerRicky Demer, OAEP+ is preferable to OAEP.

First, remark that the desired commutativity is incompatible with security under Chosen Plaintext Attack, which (under the name IND-CPA) is considered a requirement for modern encryption systems. Proof, expanded following tylo's comment, using the IND-CPA game as played for symmetric encryption (see the CPA indistinguishability experiment in section 3.5 of Katz and Lindell's Introduction to modern cryptography, also here on page 4):

As remarked by figlesquidge, XOR with a One Time Pad has the desired commutativity property, in some sense. That's also true for any Stream Cipher with an out-of-band method for synchronization (including a block cipher such as AES in OFB or CTR mode, with out-of-band IV); by out-of-band I mean: not part of the ciphertext for which the commutativity property is thought.

If $\small\text{OAEP}(x)$ designates the padding used for RSA-OAEP with a modulus of $\lceil\log_2(p)\rceil$ bits, then $x\mapsto E(\small\text{OAEP}(x))$ seems to be IND-CPA-secure, and decipherable (the desired commutativity property is lost: that padding must be external to whatever step of the protocol requires commutativity). As pointed by Ricky Demer, OAEP+ is preferable to OAEP.

First, remark that the desired commutativity is incompatible with security under Chosen Plaintext Attack, which (under the name IND-CPA) is considered a requirement for modern encryption systems. Proof, expanded following tylo's comment, using the IND-CPA game as played for symmetric encryption (see the CPA indistinguishability experiment in section 3.5 of Katz and Lindell's Introduction to modern cryptography, also here on page 4):

As remarked by figlesquidge, XOR with a One Time Pad has the desired commutativity property, in some sense. That's also true for any Stream Cipher with an out-of-band method for synchronization (including a block cipher such as AES in OFB or CTR mode, with out-of-band IV); by out-of-band I mean: not part of the ciphertext for which the commutativity property is thought.

If $\small\text{OAEP}(x)$ designates the padding used for RSA-OAEP with a modulus of $\lceil\log_2(p)\rceil$ bits, then $x\mapsto E(\small\text{OAEP}(x))$ seems to be IND-CPA-secure, and decipherable (the desired commutativity property is lost: that padding must be external to whatever step of the protocol requires commutativity). As pointed by Ricky Demer, OAEP+ is preferable to OAEP.

Link to reference paper on Mental Poker
Source Link
fgrieu
  • 145.4k
  • 12
  • 319
  • 611

We study the simplest commutative cipher in the cryptographic folklore, sometime used for Mental PokerMental Poker:

We study the simplest commutative cipher in the cryptographic folklore, used for Mental Poker:

We study the simplest commutative cipher in the cryptographic folklore, sometime used for Mental Poker:

fix location in reference article
Source Link
fgrieu
  • 145.4k
  • 12
  • 319
  • 611

For appropriate $p$ (making the Discrete Logarithm Problem hard), the cipher is conjectured secure for multiple random plaintexts related to the same random key [that's statestated before (1415) in the original article, linked to the title of this section, withsection; and justified by an argument I do not get]. However, it has a number of other properties that a random commutative cipher would not obviously have, including the multiplicative property $\pmod p$:

For appropriate $p$ (making the Discrete Logarithm Problem hard), the cipher is conjectured secure for multiple random plaintexts related to the same random key [that's state before (14) in the article linked to the title of this section, with an argument I do not get]. However, it has a number of other properties that a random commutative cipher would not obviously have, including the multiplicative property $\pmod p$:

For appropriate $p$ (making the Discrete Logarithm Problem hard), the cipher is conjectured secure for multiple random plaintexts related to the same random key [that's stated before (15) in the original article, linked to the title of this section; and justified by an argument I do not get]. However, it has a number of other properties that a random commutative cipher would not obviously have, including the multiplicative property $\pmod p$:

State the conjectured security of the Pohlig-Hellman Exponentiation Cipher
Source Link
fgrieu
  • 145.4k
  • 12
  • 319
  • 611
Loading
Expand proof, make notation more consistent
Source Link
fgrieu
  • 145.4k
  • 12
  • 319
  • 611
Loading
Mod Removes Wiki by Paŭlo Ebermann
Remark for probabilistic cipher
Source Link
fgrieu
  • 145.4k
  • 12
  • 319
  • 611
Loading
Change references to the IND-CPA game to a most accepted one
Source Link
fgrieu
  • 145.4k
  • 12
  • 319
  • 611
Loading
Polish
Source Link
fgrieu
  • 145.4k
  • 12
  • 319
  • 611
Loading
Back to "commutativity is incompatible with security under CPA".
Source Link
fgrieu
  • 145.4k
  • 12
  • 319
  • 611
Loading
It may be that even a probabilistic encryption scheme can't be both commutative and IND-CPA secure
Source Link
fgrieu
  • 145.4k
  • 12
  • 319
  • 611
Loading
Correct per comment
Source Link
fgrieu
  • 145.4k
  • 12
  • 319
  • 611
Loading
Separate the standard Pohlig-Hellman Exponentiation Cipher from the variant I propose; improve the later to remove symetry.
Source Link
fgrieu
  • 145.4k
  • 12
  • 319
  • 611
Loading
Remove the three key-indenpendent fixed points. Explain why Pohlig-Hellman Exponentiation Cipher is seldom used.
Source Link
fgrieu
  • 145.4k
  • 12
  • 319
  • 611
Loading
Explain why OTP and friends is not usable for Mental Poker
Source Link
fgrieu
  • 145.4k
  • 12
  • 319
  • 611
Loading
Make the proposed distinguisher work for our E
Source Link
fgrieu
  • 145.4k
  • 12
  • 319
  • 611
Loading
Polish
Source Link
fgrieu
  • 145.4k
  • 12
  • 319
  • 611
Loading
That's the Pohlig-Hellman Exponentiation Cipher
Source Link
fgrieu
  • 145.4k
  • 12
  • 319
  • 611
Loading
That's the Pohlig-Hellman Exponentiation Cipher
Source Link
fgrieu
  • 145.4k
  • 12
  • 319
  • 611
Loading
IND-CPA security is incompatible with the commutativity property
Source Link
fgrieu
  • 145.4k
  • 12
  • 319
  • 611
Loading
Align segments in ke. Fix modulus size for padding.
Source Link
fgrieu
  • 145.4k
  • 12
  • 319
  • 611
Loading
Add intro. Make k a more standard key. Update per Ricky Demer's comment.
Source Link
fgrieu
  • 145.4k
  • 12
  • 319
  • 611
Loading
Source Link
fgrieu
  • 145.4k
  • 12
  • 319
  • 611
Loading