Jump to content

FourQ: Difference between revisions

From Wikipedia, the free encyclopedia
Content deleted Content added
Citation bot (talk | contribs)
Removed parameters. | Use this bot. Report bugs. | #UCB_CommandLine
 
(8 intermediate revisions by 5 users not shown)
Line 9: Line 9:
| released = {{Start date and age|2015}}
| released = {{Start date and age|2015}}
| discontinued =
| discontinued =
| latest release version = v3.0
| latest release version = v3.1
| latest release date =
| latest release date =
| repo = {{URL|https://github.com/microsoft/FourQlib}}
| repo = {{URL|https://github.com/microsoft/FourQlib}}
Line 20: Line 20:
| website = {{URL|https://www.microsoft.com/en-us/research/project/fourqlib/}}
| website = {{URL|https://www.microsoft.com/en-us/research/project/fourqlib/}}
}}
}}
In [[cryptography]], '''FourQ''' is an [[elliptic-curve cryptography|elliptic curve]] developed by [[Microsoft Research]]. It is designed for key agreements schemes ([[elliptic-curve Diffie–Hellman]]) and digital signatures ([[Schnorr signature|Schnorr]]), and offers about 128 [[Security level|bits of security]].<ref name="4q">{{cite journal |last1=Costello |first1=Craig |last2=Longa |first2=Patrick |title=FourQ: four-dimensional decompositions on a Q-curve over the Mersenne prime |year=2015 |url=https://eprint.iacr.org/2015/565 |access-date=23 May 2019 }}</ref> It is equipped with a [[reference implementation]] made by the authors of the original paper. The [[open source]] implementation is called ''FourQlib'' and runs on [[Windows]] and [[Linux]] and is available for x86, x64, and ARM.<ref name="msf">{{cite web |title=FourQlib |url=https://www.microsoft.com/en-us/research/project/fourqlib/ |website=Microsoft Research |access-date=23 May 2019}}</ref> It is licensed under [[MIT License]] and the source code is available on [[GitHub]].<ref>https://github.com/microsoft/FourQlib</ref>
In [[cryptography]], '''FourQ''' is an [[elliptic-curve cryptography|elliptic curve]] developed by [[Microsoft Research]]. It is designed for key agreements schemes ([[elliptic-curve Diffie–Hellman]]) and digital signatures ([[Schnorr signature|Schnorr]]), and offers about 128 [[Security level|bits of security]].<ref name="4q">{{cite journal |last1=Costello |first1=Craig |last2=Longa |first2=Patrick |title=FourQ: four-dimensional decompositions on a Q-curve over the Mersenne prime |year=2015 |url=https://eprint.iacr.org/2015/565 |access-date=23 May 2019 }}</ref> It is equipped with a [[reference implementation]] made by the authors of the original paper. The [[open source]] implementation is called ''FourQlib'' and runs on [[Windows]] and [[Linux]] and is available for x86, x64, and ARM.<ref name="msf">{{cite web |title=FourQlib |url=https://www.microsoft.com/en-us/research/project/fourqlib/ |website=Microsoft Research |access-date=23 May 2019}}</ref> It is licensed under the [[MIT License]] and the source code is available on [[GitHub]].<ref>{{Cite web|url=https://github.com/microsoft/FourQlib|title = References| website=[[GitHub]] |date = 4 October 2021}}</ref>


Its name is derived from the four dimensional Gallant–Lambert–Vanstone scalar multiplication, which allows high performance calculations.<ref>{{cite journal |last1=Longa |first1=Patrick |last2=Sica |first2=Francesco |title=Four-Dimensional Gallant-Lambert-Vanstone Scalar Multiplication |access-date=23 May 2019 |year=2011 |url=https://eprint.iacr.org/2011/608|arxiv=1106.5149 }}</ref> The curve is defined over a two dimensional [[field extension|extension]] of the [[prime]] field defined by the [[Mersenne prime]] <math>2^{127} - 1</math>.
Its name is derived from the four dimensional Gallant–Lambert–Vanstone scalar multiplication, which allows high performance calculations.<ref>{{cite journal |last1=Longa |first1=Patrick |last2=Sica |first2=Francesco |title=Four-Dimensional Gallant–Lambert–Vanstone Scalar Multiplication |access-date=23 May 2019 |year=2011 |url=https://eprint.iacr.org/2011/608|arxiv=1106.5149 }}</ref> The curve is defined over a two dimensional [[field extension|extension]] of the [[Prime number|prime]] field defined by the [[Mersenne prime]] <math>2^{127} - 1</math>.


== History ==
== History ==
Line 29: Line 29:
The paper was presented in [[Asiacrypt]] in 2015 in [[Auckland]], New Zealand, and consequently a [[reference implementation]] was published on [[Microsoft]]'s website.<ref name="msf"/>
The paper was presented in [[Asiacrypt]] in 2015 in [[Auckland]], New Zealand, and consequently a [[reference implementation]] was published on [[Microsoft]]'s website.<ref name="msf"/>


There were some efforts to standardize usage of the curve under [[IETF]]; these efforts were withdrawn in late 2017.<ref>{{cite web |title=draft-ladd-cfrg-4q-01 |url=https://datatracker.ietf.org/doc/draft-ladd-cfrg-4q/ |website=datatracker.ietf.org |access-date=23 May 2019}}</ref>
There were some efforts to standardize usage of the curve under [[IETF]]; these efforts were withdrawn in late 2017.<ref>{{cite journal |title=draft-ladd-cfrg-4q-01 |url=https://datatracker.ietf.org/doc/draft-ladd-cfrg-4q/ |newspaper=Ietf Datatracker |date=27 March 2017 |access-date=23 May 2019|last1=Ladd |first1=Watson |last2=Longa |first2=Patrick |last3=Barnes |first3=Richard }}</ref>


== Mathematical Properties ==
== Mathematical properties ==
The curve is defined by a [[twisted Edwards curve|twisted Edwards equation]]
The curve is defined by a [[twisted Edwards curve|twisted Edwards equation]]
:<math>-x^2 + y^2 = 1 + d x^2 y^2</math>
:<math>-x^2 + y^2 = 1 + d x^2 y^2</math>
<math>d</math> is a non-square in <math>\mathbb{F}_{p^2}</math>, where <math>p</math> is the [[Mersenne prime]] <math>2^{127}-1</math>.
<math>d</math> is a non-square in <math>\mathbb{F}_{p^2}</math>, where <math>p</math> is the [[Mersenne prime]] <math>2^{127}-1</math>.


In order to avoid [[Small subgroup confinement attack|small subgroup attacks]],<ref>{{cite journal |last1=van Oorschot |first1=Paul C. |last2=Wiener |first2=Michael J. |title=On Diffie-Hellman Key Agreement with Short Exponents |journal=Advances in Cryptology – EUROCRYPT '96 |volume=1070 |year=1996 |pages=332–343 |doi=10.1007/3-540-68339-9_29 |publisher=Springer Berlin Heidelberg |series=Lecture Notes in Computer Science |isbn=978-3-540-61186-8 |doi-access=free }}</ref> all points are verified to lie in an ''N''-[[Torsion subgroup|torsion]] subgroup of the [[elliptic curve]], where ''N'' is specified as a 246-bit [[prime]] dividing the [[Order (group theory)|order]] of the group.
In order to avoid [[Small subgroup confinement attack|small subgroup attacks]],<ref>{{cite book |last1=van Oorschot |first1=Paul C. |last2=Wiener |first2=Michael J. |title=Advances in Cryptology — EUROCRYPT '96 |chapter=On Diffie-Hellman Key Agreement with Short Exponents |volume=1070 |year=1996 |pages=332–343 |doi=10.1007/3-540-68339-9_29 |publisher=Springer Berlin Heidelberg |series=Lecture Notes in Computer Science |isbn=978-3-540-61186-8 |doi-access=free }}</ref> all points are verified to lie in an ''N''-[[Torsion subgroup|torsion]] subgroup of the [[elliptic curve]], where ''N'' is specified as a 246-bit [[Prime number|prime]] dividing the [[Order (group theory)|order]] of the group.


The curve is equipped with two nontrivial [[endomorphism]]s: <math>\psi</math> related to the <math>p</math>-power [[Frobenius map]], and <math>\phi</math>, a low degree efficiently computable endomorphism (see [[complex multiplication]]).
The curve is equipped with two nontrivial [[endomorphism]]s: <math>\psi</math> related to the <math>p</math>-power [[Frobenius map]], and <math>\phi</math>, a low degree efficiently computable endomorphism (see [[complex multiplication]]).


== Cryptographic Properties ==
== Cryptographic properties ==
=== Security ===
=== Security ===
The currently best known [[discrete logarithm]] attack is the generic [[Pollard's rho algorithm]], requiring about <math>2^{122.5}</math> group operations on average. Therefore, it typically belongs to the 128 bit security level.
The currently best known [[discrete logarithm]] attack is the generic [[Pollard's rho algorithm]], requiring about <math>2^{122.5}</math> group operations on average. Therefore, it typically belongs to the 128 bit security level.
Line 47: Line 47:


=== Efficiency ===
=== Efficiency ===
Most cryptographic primitives, and most notably [[Elliptic Curve Diffie-Hellman|ECDH]], require fast computation of scalar multiplication, i.e. <math>[k]P</math> for a point <math>P</math> on the curve and an integer <math>k</math>, which is usually thought as distributed uniformly at random over <math>\{0, \ldots, N-1\}</math>.
Most cryptographic primitives, and most notably [[Elliptic-curve Diffie–Hellman|ECDH]], require fast computation of scalar multiplication, i.e. <math>[k]P</math> for a point <math>P</math> on the curve and an integer <math>k</math>, which is usually thought as distributed uniformly at random over <math>\{0, \ldots, N-1\}</math>.


Since we look at a [[prime]] order [[cyclic group|cyclic]] subgroup, one can write scalars <math>\lambda_\psi, \lambda_\phi</math> such that <math>\psi(P) = [\lambda_\psi]P</math> and <math>\phi(P) = [\lambda_\phi]P</math> for every point <math>P</math> in the N-torsion subgroup.
Since we look at a [[Prime number|prime]] order [[cyclic group|cyclic]] subgroup, one can write scalars <math>\lambda_\psi, \lambda_\phi</math> such that <math>\psi(P) = [\lambda_\psi]P</math> and <math>\phi(P) = [\lambda_\phi]P</math> for every point <math>P</math> in the ''N''-torsion subgroup.


Hence, for a given <math>k</math> we may write
Hence, for a given <math>k</math> we may write
Line 55: Line 55:
If we find small <math>a_i</math>, we may compute <math>[k]P</math> quickly by utilizing the implied equation
If we find small <math>a_i</math>, we may compute <math>[k]P</math> quickly by utilizing the implied equation
:<math>[k]P = [a_1]P + [a_2] \phi(P) + [a_3] \psi(P) + [a_4] \phi(\psi(P))</math>
:<math>[k]P = [a_1]P + [a_2] \phi(P) + [a_3] \psi(P) + [a_4] \phi(\psi(P))</math>
[[Babai rounding]] technique<ref>{{cite journal |last1=Babai |first1=L. |title=On Lovász' lattice reduction and the nearest lattice point problem |journal=Combinatorica |date=1 March 1986 |volume=6 |issue=1 |pages=1–13 |doi=10.1007/BF02579403 |issn=1439-6912}}</ref> is used to find small <math>a_i</math>. For FourQ it turns that one can guarantee an efficiently computable solution with <math>a_i < 2^{64}</math>.
[[Babai rounding]] technique<ref>{{cite journal |last1=Babai |first1=L. |title=On Lovász' lattice reduction and the nearest lattice point problem |journal=Combinatorica |date=1 March 1986 |volume=6 |issue=1 |pages=1–13 |doi=10.1007/BF02579403 |s2cid=7914792 |issn=1439-6912}}</ref> is used to find small <math>a_i</math>. For FourQ it turns that one can guarantee an efficiently computable solution with <math>a_i < 2^{64}</math>.


Moreover, as the [[Characteristic (algebra)|characteristic]] of the field is a [[Mersenne prime]], modulations can be carried efficiently.
Moreover, as the [[Characteristic (algebra)|characteristic]] of the field is a [[Mersenne prime]], modulations can be carried efficiently.
Line 64: Line 64:
{{Missing information|section|uses|date=July 2019}}
{{Missing information|section|uses|date=July 2019}}


FourQ is implemented in the cryptographic library [https://github.com/cloudflare/circl CIRCL], published by [[Cloudflare]].<ref>{{cite web |title=Introducing CIRCL |url=https://blog.cloudflare.com/introducing-circl/ |website=blog.cloudflare.com |access-date=28 July 2019}}</ref>
FourQ is implemented in the cryptographic library [https://github.com/cloudflare/circl CIRCL], published by [[Cloudflare]].<ref>{{cite web |title=Introducing CIRCL |url=https://blog.cloudflare.com/introducing-circl/ |website=blog.cloudflare.com |date=20 June 2019 |access-date=28 July 2019}}</ref>


== See also ==
== See also ==

Latest revision as of 02:52, 7 July 2023

FourQ
Developer(s)Microsoft Research
Initial release2015; 9 years ago (2015)
Stable release
v3.1
Repositorygithub.com/microsoft/FourQlib
Written inC
Operating systemWindows 10, Linux
PlatformIA-32, x86-64, ARM32, ARM64
TypeElliptic-curve cryptographic library
LicenseMIT License
Websitewww.microsoft.com/en-us/research/project/fourqlib/

In cryptography, FourQ is an elliptic curve developed by Microsoft Research. It is designed for key agreements schemes (elliptic-curve Diffie–Hellman) and digital signatures (Schnorr), and offers about 128 bits of security.[1] It is equipped with a reference implementation made by the authors of the original paper. The open source implementation is called FourQlib and runs on Windows and Linux and is available for x86, x64, and ARM.[2] It is licensed under the MIT License and the source code is available on GitHub.[3]

Its name is derived from the four dimensional Gallant–Lambert–Vanstone scalar multiplication, which allows high performance calculations.[4] The curve is defined over a two dimensional extension of the prime field defined by the Mersenne prime .

History

[edit]

The curve was published in 2015 by Craig Costello and Patrick Longa from Microsoft Research on ePrint.[1]

The paper was presented in Asiacrypt in 2015 in Auckland, New Zealand, and consequently a reference implementation was published on Microsoft's website.[2]

There were some efforts to standardize usage of the curve under IETF; these efforts were withdrawn in late 2017.[5]

Mathematical properties

[edit]

The curve is defined by a twisted Edwards equation

is a non-square in , where is the Mersenne prime .

In order to avoid small subgroup attacks,[6] all points are verified to lie in an N-torsion subgroup of the elliptic curve, where N is specified as a 246-bit prime dividing the order of the group.

The curve is equipped with two nontrivial endomorphisms: related to the -power Frobenius map, and , a low degree efficiently computable endomorphism (see complex multiplication).

Cryptographic properties

[edit]

Security

[edit]

The currently best known discrete logarithm attack is the generic Pollard's rho algorithm, requiring about group operations on average. Therefore, it typically belongs to the 128 bit security level.

In order to prevent timing attacks, all group operations are done in constant time, i.e. without disclosing information about key material.[1]

Efficiency

[edit]

Most cryptographic primitives, and most notably ECDH, require fast computation of scalar multiplication, i.e. for a point on the curve and an integer , which is usually thought as distributed uniformly at random over .

Since we look at a prime order cyclic subgroup, one can write scalars such that and for every point in the N-torsion subgroup.

Hence, for a given we may write

If we find small , we may compute quickly by utilizing the implied equation

Babai rounding technique[7] is used to find small . For FourQ it turns that one can guarantee an efficiently computable solution with .

Moreover, as the characteristic of the field is a Mersenne prime, modulations can be carried efficiently.

Both properties (four dimensional decomposition and Mersenne prime characteristic), alongside usage of fast multiplication formulae (extended twisted Edwards coordinates), make FourQ the currently fastest elliptic curve for the 128 bit security level.

Uses

[edit]

FourQ is implemented in the cryptographic library CIRCL, published by Cloudflare.[8]

See also

[edit]

References

[edit]
  1. ^ a b c Costello, Craig; Longa, Patrick (2015). "FourQ: four-dimensional decompositions on a Q-curve over the Mersenne prime". Retrieved 23 May 2019. {{cite journal}}: Cite journal requires |journal= (help)
  2. ^ a b "FourQlib". Microsoft Research. Retrieved 23 May 2019.
  3. ^ "References". GitHub. 4 October 2021.
  4. ^ Longa, Patrick; Sica, Francesco (2011). "Four-Dimensional Gallant–Lambert–Vanstone Scalar Multiplication". arXiv:1106.5149. Retrieved 23 May 2019. {{cite journal}}: Cite journal requires |journal= (help)
  5. ^ Ladd, Watson; Longa, Patrick; Barnes, Richard (27 March 2017). "draft-ladd-cfrg-4q-01". Ietf Datatracker. Retrieved 23 May 2019.
  6. ^ van Oorschot, Paul C.; Wiener, Michael J. (1996). "On Diffie-Hellman Key Agreement with Short Exponents". Advances in Cryptology — EUROCRYPT '96. Lecture Notes in Computer Science. Vol. 1070. Springer Berlin Heidelberg. pp. 332–343. doi:10.1007/3-540-68339-9_29. ISBN 978-3-540-61186-8.
  7. ^ Babai, L. (1 March 1986). "On Lovász' lattice reduction and the nearest lattice point problem". Combinatorica. 6 (1): 1–13. doi:10.1007/BF02579403. ISSN 1439-6912. S2CID 7914792.
  8. ^ "Introducing CIRCL". blog.cloudflare.com. 20 June 2019. Retrieved 28 July 2019.
[edit]