Paradigm of Post-quantum Cryptography and Crypto-agility:
Strategy Approach of Quantum-safe Techniques
Olaf Grote
1
, Andreas Ahrens
2
and C
´
esar Benavente-Peces
3
1
ETS de Ingenier
´
ıa y Sistemas de Telecomunicaci
´
on, University Polit
´
ecnica de Madrid,
Campus Sur, Ctra. Valencia, 28031 Madrid, Spain
2
Hochschule Wismar, University of Applied Sciences - Technology, Business and Design,
Philipp-M
¨
uller-Straße 14, 23966 Wismar, Germany
3
ETS de Ingenier
´
ıa y Sistemas de Telecomunicaci
´
on, University Polit
´
ecnica de Madrid,
Campus Sur, Ctra. Valencia, 28031 Madrid, Spain
Keywords:
Post-quantum Cryptography, Crypto-agility, Quantum-safe, Quantum Algorithms, Quantum Computer
Science.
Abstract:
Current security protocols use cryptographic methods based on asymmetric and symmetric schemes. These
are used for encrypted communication of information or for general data encryption. The security of these
methods is based on well researched methods and known mathematical problems. These have been developed
for common computing resources and with the certainty that they cannot be broken at a determinable runtime
with finite resources. This excludes novel attack vectors such as those of a quantum computer using quantum
algorithms. These cryptographic methods, especially the asymmetric schemes, are not prepared against quan-
tum attacks and not considered quantum safe. Even the security of current quantum safe symmetric schemes
is not based on proven security against quantum attacks. In order to effectively counter this threat, a new and
effective strategy is necessary. One point of this strategy is the post-quantum cryptography to evaluate new
quantum-safe cryptographic principles. The second point is to research security protocols which are quantum
safe and resistible against quantum attacks. This paper describes the strategy of post-quantum cryptography
and crypto-agility.
1 INTRODUCTION
Current cryptographic schemes must be evaluated to
see how they are still effective against quantum at-
tacks. This affects encrypted communication (data
in transit) and encrypted data that is not continu-
ously changed (data at rest). However, from the
perspective of a quantum computer the symmetric
scheme is considered quantum safe and the asym-
metric as be broken. In order to make this scheme
quantum safe, quantum computer resistant public key
derivatives are sought because this method is cur-
rently needed. The novel contribution of this article
is the review and presentation, which holistic con-
cept is necessary to ward off possible attacks on es-
tablished security mechanisms with quantum algo-
rithms. The paper focusing is on the established cryp-
tographic methods and how current security protocols
can be adapted against quantum attacks. They are two
variants to be resilient against quantum attacks. On
the one hand, research and investigation of quantum
computer-resistant cryptography primitives for post-
quantum cryptography (PQC) is needed to find poten-
tial candidates. On the other hand is the cryptographic
agility (crypto-agility) of security protocols to ex-
change, modify or parameterize public-key cryptog-
raphy algorithms with new cryptographic primitives
which from the perspective of quantum computing
are resilient against quantum attacks. The remaining
parts of this paper are structured as follows: Section
2 shows the state of the art established cryptography
in context of the new approach. Section 3 introduce
the need for use to define a concept and paradigm of
post-quantum cryptography and crypto-agility. The
main section 4 describes the paradigm. Thus section
describes in the first part the new mathematical primi-
tives and cryptographic schemes candidates and in the
second part describes the crypto-agility how it is pos-
sible to integrate this schemes and the way to modify
established security protocol quantum-safe. The re-
sults of this study are presented in section 5. Finally,
section 6 provides some concluding remarks.
Grote, O., Ahrens, A. and Benavente-Peces, C.
Paradigm of Post-quantum Cryptography and Crypto-agility: Strategy Approach of Quantum-safe Techniques.
DOI: 10.5220/0008162800910098
In Proceedings of the 9th International Conference on Pervasive and Embedded Computing and Communication Systems (PECCS 2019), pages 91-98
ISBN: 978-989-758-385-8
Copyright
c
2019 by SCITEPRESS Science and Technology Publications, Lda. All rights reserved
91
2 STATE OF THE ART
In general, the functionality of an algorithm is a
unique, time dedicated, executable sequence of in-
structions with defined length. Such algorithms scale
against the input value exponential. Computer-based
programs deliver the quantified results of this algo-
rithm. The estimated runtime and the efficiency of
cryptographic algorithms are estimated relatively dis-
advantageously. Based on complexity theory, each al-
gorithm is runtime efficient of class P, where P con-
tains all decision problems when the input polyno-
mial is determined and the result is computer-capable
with a polynomial effort runtime. For example, the
following algorithms belong to class P, which have
a constant, logarithmic, linear and quadratic runtime.
Nevertheless, all algorithms that are correctly recog-
nized for the solution of a problem in polynomial
time by means of a non-deterministic Turing machine,
but require an exponential computing time in case
of a wrong solution, are called runtime inefficient of
the class NP, where NP define the non-deterministic
polynomial time. Current cryptographic methods
are based on these complexity-theoretical statements,
since P 6= NP is currently assumed to be mathemati-
cal, but has not yet been proven (Eckert, 2014). Thus,
if the complexity class of a problem depends on the
algorithm or is determined by the required resources
such as computing power and memory, quantum al-
gorithms can perform these tasks with optimized run-
times. In contrast to the algorithms that make all prob-
lems in complexity class P decidable to a polynomial
runtime on a deterministic Turing machine, quantum
algorithms include functions rather on questions like
searching in a database, recognizing a global prop-
erty of a function, e.g. period, mean value. Further-
more, quantum algorithms solve runtime-optimized
problems on a non-deterministic Turing machine such
as the number-theoretical problem or the calculation
of the gradient in ndimensions. Two quantum algo-
rithms have received quite a bit of notice: Shor’s al-
gorithm for factoring integers in polynomial time on
a quantum computer (Shor, 1994) and Grover’s algo-
rithm for searching a unsorted database of x elements
with efficient runtime (Grover, 1996). The advantage
and main reason of the high efficiency of quantum
algorithms is the calculation of the periodicity of a
function, which is available as a global property after
the first operation in the quantum register of a quan-
tum computer. With this method, which provides for a
concrete parallelism, the periodic and recurring com-
ponents within a bit sequence can be efficiently fil-
tered and useful for a quantum attack.
3 QUANTUM ATTACK
APPROACHES
Quantum mechanics describes in simplified terms the
microscopic properties of a physical object where the
specific state of this object temporal and physical
space only be determined vaguely. This object (e.g.
atoms, electrons, photons) particle has the same en-
ergetic order of magnitude as the object with which
it is to be measured. This also means that the orig-
inal particle is demonstrably changed after a mea-
surement (Brands, 2011). Quantum computer follow
the laws of quantum physics and performs operations
with three states ”0” or ”1” or ”0 and 1”. This new
computer capability solve mathematical problems of
the complexity class NP in a polynomial time so that
P = NP. Shor’s algorithm (1) which performed on a
quantum computer breaks in polynomial time cryp-
tographic primitives and schemes that based on inte-
ger factorization and discrete logarithms (Shor, 1997).
This algorithm delivers to one natural number N a
nontrivial factor which are defined as:
O((logN)
3
) (1)
All public-key cryptography that use those algebraic
structure are affected like Rivest, Shamir, Adleman
method (RSA), Elliptic Curve Cryptography (ECC)
and Diffie-Hellman (DH). Grover’s algorithm (2) is a
runtime optimized quantum search algorithm to use
brute-force attacking that checks all possible cipher
key by determined time and resources (3) (Grover,
1999), where n is the number of bits that are searched
(see Table 1):
O(
n) (2)
O(logn) (3)
This enables the quantum computer to use the square-
root factor and halved the exponent of time complex-
ity in opposite to a linear search algorithm O(n). The
Table 1 shows the runtime and cost effort for each
algorithm and for the symmetric cryptography Ad-
vanced Encryption Standard (AES).
Table 1: Runtime and cost efforts per algorithm.
n O(n) O(
n) O(
3
n)
128 bit 128 bit 64 bit 42,66
192 bit 192 bit 96 bit 64
256 bit 256 bit 128 bit 85,33
Furthermore, the Grover’s algorithm is useful for at-
tacking the established secure hash algorithm (SHA)
to find preimages (Amy et al., 2017) and hash col-
lisions by modifying the algorithm with a cube-root
factor (Brassard et al., 1998). In general Grover algo-
rithm is a probabilistic algorithm and tries to achieve a
PECCS 2019 - 9th International Conference on Pervasive and Embedded Computing and Communication Systems
92
good or approximately correct average result. To find
the right key it makes sense to use this randomized
algorithm again on the previous result.
4 PARADIGM OF
POST-QUANTUM
CRYPTOGRAPHY AND
CRYPTO-AGILITY
In terms of special requirements, future-proof cryp-
tography is characterized by the ability to adapt to
new technologies and challenges. Today’s cryptog-
raphy achieves this ideal through post-quantum cryp-
tography and crypto-agility. That includes research
new cryptographic primitives and ensuring common
mechanisms to stable current system environment and
the idea to modify security protocols. This will be
necessary because the development of quantum com-
puters is an active field of research. Technical guide-
lines and recommendations for quantum computer re-
sistant cryptographic methods are currently reaching
a stage of development, but no standardization yet.
According to the current state of research, one can
classify symmetric encryption as a quantum computer
resistant. The turn-based method (4), as well as the
key length (5) of one of its key phrases |K|, allows
this assumption:
|K| = 10/128,12/192,14/256 (4)
|K| = 2
128
,2
192
,2
256
(5)
These properties make the symmetric scheme resis-
tant to brute-force attacks by a quantum computer. A
complete analysis of the keyspace is considered inef-
fective. An attacker could scan the ciphertext about 4
times faster and break it if he has little plaintext to the
ciphertext than to perform an analysis of the keyspace
(Bogdanov et al., 2011). Furthermore, the higher the
key, the fewer steps are required for decryption. For
example, a ciphertext created with AES-192 requires

steps for decryption and only
,
steps for AES-
256 (Biryukov and Khovratovich, 2009). Neverthe-
less, in brute-force attacks of quantum computers on
the key space of AES thanks to turn-based methodol-
ogy and the key length as a quantum-safe. As an alter-
native to AES, the Camellia procedure recommended
by the European ENISA can be used. Both methods
are based on block encryption with the same block
cipher. Camellia is slower in direct comparison to
AES, but uses a block cipher directly of 128 bits for
more laps (/,/,/). For stream ci-
phers, the method Salsa20 (Snuffle 2005) with a 256
bit key length could be used. The standard version
Salsa20/20 uses an encryption of 20 rounds. The al-
gorithm is extremely efficient and resistant to possible
side channel attacks. The Salsa20/12 and Salsa20/8
variants can be used for time-critical as well as jitter-
and latency-relevant applications. In contrast to sym-
metric methods, the asymmetric method is consid-
ered non-quantum-safe. Nevertheless, this scheme
still needs to be used for key exchange. With asyn-
chronous encryption algorithms, it makes sense to use
a RSA key length of 3000 bits to achieve a secu-
rity level that can react to possible attacks from quan-
tum computers. In order to attack a 3000 RSA
key a quantum computer must have a few thousand
logical Qubits for the mathematical operation. Fur-
thermore, correction or cache Qubits are necessary,
which hold intermediate results. To compare, state-
of-the-art fully controlled quantum computer operates
of 20 Qubits with entangled states (Friis et al., 2018).
But it makes sense to search for other public-key al-
ternatives to replace this asymmetric methods. The
McEliece method can be used as an RSA alternative,
since no quantum algorithm has yet been found that
can effectively break this cryptographic method but
use of the large matrices for this method is rarely used
in practice. Quantum Key Distribution (QKD) is a
possible solution for the key exchange. The QKD
belongs to quantum cryptography and is a method
for secure key exchange. The method guarantees its
security with the physical fact of quantum mechan-
ics. This method is not suitable as an encryption
method for messages, but for the secure exchange of
keys. A man in the middle attacker would have to
actively measure the quantum particles. This gener-
ates high error rates and measurement inaccuracies
on the quantum channel. Furthermore, the charge
of the quantum particles would collapse. This gives
you physical evidence of safety. With this method
even very long symmetric keys can be transmitted in
a relatively short time. An application of the Ver-
nam’s One Time Pad (OTP) method, which is con-
sidered quantum-safe, would be conceivable. ECC
equivalents like Elliptic Curve Digital Signature Al-
gorithm (ECDSA) and Elliptic Curve Diffie-Hellman
(ECDH) must be considered broken and not useful
for the PQC. Nevertheless, one can provide secu-
rity protocols with a hybrid procedure. The Super-
singular Isogeny Diffie-Hellman (SIDH) is defined
as quantum computer resistant and could be a can-
didate for hybride implementation of ECC with e.g.
Montgomery and Twisted Edwards Curves (Boureanu
et al., 2014). For digital signature procedures, new
such hash-based eXtended Merkle Signature Scheme
(XMSS) and SPHINCS can be used for PQC. This
hash methods are high-security post-quantum state-
Paradigm of Post-quantum Cryptography and Crypto-agility: Strategy Approach of Quantum-safe Techniques
93
less hash-based signature schemes. The established
secure hash algorithm SHA-2/-3 is currently believed
secure and quantum computer resistant but not to be
resilient and proved. However, the 384 hash value
(bit) should already be used for long-term storage. Ta-
ble 2 illustrates the recommended key length for the
cryptographic scheme.
Table 2: Recommended key length and method.
Scheme Recommendation
AES 256 bit
Camellia 192 bit
Salsa20 256 bit stream chiffre
RSA 3000 bit
McEliece method 128 bit
SIDH 128 / 192 bit
SHA-2 /-3 256 / 384 Hash value (bit)
XMSS/ SPHINCS SHA-256 / AES-128
4.1 Post-quantum Cryptography
Increase of the key space alone is not enough to be
quantum-safe. For example, If the symmetrical en-
cryption AES was attacked by the Grover quantum
algorithm, the key strength would still be 50%. An
attack via Shor quantum algorithm on asymmetric al-
gorithms such as RSA or ECC could no longer nearly
guarantee the same security that an asymmetric en-
cryption would provide. Even increasing the size of
the key space would not be an adequate measure and
would also make the cryptographic process inefficient
(Fumy, 2017). Table 3 shows the effectively capacity
of the key size and key data and define the efficient
properties of the established not quantum-safe public-
key schemes.
Table 3: Key properties of no quantum-safe schemes.
Key-Length (bit) Size (bytes) Data (bytes)
RSA-2048 256 256
RSA-3072 384 384
RSA-4096 512 512
ECC-256 32 32
ECC-512 64 64
The Table 4 shows schemes which are believed
quantum-safe and proofed secure level for public-key
and key exchange method. This methods operates
with a transmitted data capacity that is significant
larger then established methods e.g. public-key sig-
natures XMSS (Butin, 2017), SPHINCS (Bernstein
et al., 2015) and HFE* (Petzoldt et al., 2015). The
same situation with the public-key encryption e.g.
the coding-based scheme McEliece (Bernstein et al.,
2008) and grid-based scheme NTRU (IEEE, 2008),
and the key exchange as well like NewHope (Alkim
et al., 2015) and SIDH (Costello et al., 2016).
Table 4: Key properties of quantum-safe schemes.
Key-Length (bit) Size (bytes) Data (bytes)
HFE*/PMI* = 500,000
¯
30
NTRU = 1,500 = 1,500
McEliece = 800,000 = 180
SIDH n/a = 500
XMSS 64 = 2,000
SPHINCS = 1,000 = 40,000
However, the challenge is to update the established se-
curity protocols and cryptographic methods in a man-
ner that is useful and beware against possible quan-
tum attacks. Here are ve potential candidates of
quantum-safe primitives.
4.1.1 Multivariate Cryptography
The multivariate cryptography is mathematically
based on a quadratic polynomial equation. The func-
tionality is based on the idea that figures on sets of n
quadratic polynomials p
1
,..., p
n
over finite elements
K
n
with more then n variables χ
1
,...,χ
n
can be repre-
sented in different ways. For function P, this results
in:
K
n
K
n
(6)
(χ
1
,...,χ
n
) (p
1
(χ
1
,...,χ
n
,..., p
n
)(χ
1
,...,χ
n
)) (7)
and P(x) can therefore be calculated if x K
n
applies.
The inverse way for a y K
n
, thus a x K
n
with func-
tion P(x) = y by means of a quadratic equation sys-
tem. The difficulty of this variant and at the same
time the safety of the procedure is to solve the equa-
tion system, as the following applies:
p
1
(χ
1
,...,χ
n
) y
1
= 0 (8)
p
n
(χ
1
,...,χ
n
) y
n
= 0 (9)
For encryption the evaluation is performed by a poly-
nomial. Decryption, on the other hand, is performed
by the inverse polynomial mapping using knowledge
of a mapping structure. The weakness of this pro-
Table 5: Character of Multivariate cryptography.
Property Specification
based on quadratic equations
methods HFE, HFEv, HFEv-, PMI+
benefits efficient, ready for use
disadvantages mathematically not proven
cedure lies in the mathematical verifiability as to
whether the function P is a one-way function, and to
ensure that it is not effectively reversible. But secu-
rity in cryptographic processes is based on this. Table
5 shows the properties of this method.
PECCS 2019 - 9th International Conference on Pervasive and Embedded Computing and Communication Systems
94
4.1.2 Grid-based Cryptography
The mathematical approach of grid-based cryptogra-
phy is based on creating a grid as a discrete subset
of a ndimensional real vector space. It is thus di-
vided in a ndimensional space like a grid and di-
vided into a defined number of cells. Each cell con-
tains objects with statistical values such as number,
average, deviations, min./max. values. The advantage
results from the low complexity, since a runtime does
not depend on the contained objects, but only on the
cells to be considered. The values for these are saved
as soon as the data has been loaded into the cells.
Grid-based cryptographic methods are rather limited
to solving mathematical problems in such grids e.g.
shortest vector problem, close vector problem, Ring-
Learning With Errors (Ring-LWE). Thus, for exam-
ple, the sum of two grid points again results in a grid
point in the grid where no other grid point exists. Fur-
thermore, the runtime of the algorithm is exponen-
tially, the more granular the more accurate the number
of dimensions of the grid. This generally means that
problems on standard grids will cause a strong safety
proof but run slower. Problems in ideal grids are used
for encryption and decryption at the expense of a low
safety proof of runtime efficiency. Table 6 shows the
general properties of the common grid-based method.
The methods BLISS and Tesla are mentioned as grid-
based methods but not presented.
Table 6: Character of Grid-based cryptography.
Property Specification
based on Grid calculation
methods New Hope, Frodo, Kyber, NTRU
benefits Runtime efficient
disadvantages No experience by cryptanalysis
A possible compromise would be a procedure based
on module grids. However, the proven speed over
RSA encryption can be noticed. Furthermore, there
are some implementation of security protocols with
New Hope that based on grid-based cryptography
(Alkim et al., 2015). The NTRU algorithm which
is based on the principles of the Ring-LWE prob-
lem, requires a runtime of O(N
) for an encryp-
tion/decryption operation of a message of length N,
while a conventional RSA method requires a run-
time of O(N
) (Hoffstein et al., 1998). NTRU algo-
rithm is thus depending on the three integer param-
eters N, p, q and four polynomials p
1
,..., p
4
with a
degree of N and an integer coefficient. The pa-
rameters p, q not necessarily need be prime numbers,
where gcd(p,q) = and q > p apply. Thus, the ring
R is calculated as follows if Z is the quantity of the
total Numbers:
R = Z[X]/(X
N
1) (10)
4.1.3 Coding-based Cryptography
The coding-based cryptography is the oldest approach
towards quantum computer resistant public-key meth-
ods and is based on the difficulty to decode gen-
eral error correcting codes efficiently. The binary
Goppa code is preferred for this, on which the meth-
ods McEliece and Niederreiter are based on. The
property of error correction codes is that the decoding
algorithm can correct a maximum of errors r, which
is encoded and/or transmitted in a ciphertext c:
y = c + e (11)
If thus e is regarded as error vector to r, one can
conclude from y to c. An attacker could not deter-
mine c from y in an acceptable time. The basis for
the security of the procedure is the secrecy of the de-
coding matrix. Only the coding matrix from which
no conclusions about the decoding matrix can be de-
rived may be published. The Goppa code forms such
a large class of algebraic error correcting codes to en-
sure the security of this procedure. The Goppa code
Γ(L,g(z)) is represented by the Goppa polynomial
g(z) having a degree t defined over a finite body hav-
ing a finite number GF(q
m
) in conjunction with a gen-
erated matrix G and vector parameters. Based on the
Goppa code, the McEliece and Niederreiter methods
are used. Table 7 shows the character of this method.
Table 7: Character of Coding-based cryptography.
Property Specification
based on Efficient decoding
methods McEliece, McBits, Niederreiter
benefits McEliece method well studied
disadvantages Large public-key, low power
4.1.4 Isogeny-based Cryptography
The isogeny-based cryptography based on the alge-
braic geometry and describes the state of two groups
that produce a structure-preserving image when cer-
tain variety properties of a homogeneity are fulfilled.
Restricted to cryptography, the elliptic curves form
over an algebraically closed body. Thus a morphism
of elliptic curves is considered to be isogenic if fol-
lowing is valid:
Φ :E
E
(12)
Φ(O) = O (13)
Paradigm of Post-quantum Cryptography and Crypto-agility: Strategy Approach of Quantum-safe Techniques
95
If such a Φ exists, then E E is called iso-
genic. Thus, mathematical mappings between el-
liptical curves are generated, which can be used as
a basis for the Diffie–Hellman key exchange (DH).
The difficulty lies in finding this isogenies between
the elliptic curves. Current implementations of DH
are based on isogenies between supersingular elliptic
curves such as the Montgomery curves. For exam-
ple the Supersingular Isogeny Diffie Hellman key ex-
change (SIDH) is a possible method. Table 8 shows
the character of this method.
Table 8: Character of Isogeny-based cryptography.
Property Specification
based on Elliptic curves principles
methods SIDH
benefits Elliptic curves well-researched
disadvantages Less researched
4.1.5 Hash-based Cryptography
Cryptographic hash procedures or one-way hash
functions are based on the difficulty of calculating
hash collisions. Digital signatures are to be issued us-
ing collision-resistant hash algorithms such as XMSS
(Buchmann et al., 2011) and SPHINCS (Bernstein
et al., 2015). Since 2018, a first draft of the IETF stan-
dard has been available (Arrow et al., 2018). XMSS
procedure is well developed and standardized and of-
fers an efficient calculation runtime, since the signa-
ture length and the verification time are linear to the
message length and independent of the number of sig-
natures - in contrast to algorithms of signature chains,
whose runtime is linear to the number of signatures on
the signature length and verification time. The XMSS
procedure is based on a stateful hash-based signature
scheme, which is mapped on a single or multi tree
(hash tree) and is based on the W-OTS+ variant. With
the pure OTS procedure, only one message can be
digitally signed with a key. With the extended W-
OTS+ variant, a limited number of digital signatures
can be generated for a XMSS public-key. Thus the
secret key must be updated after each signing and the
key pair is limited with regard to the signature pro-
cess. The number of messages m to be signed and
verified in a key pair is mathematically dependent on
the height H of the hash tree. Thus, the number of
H
messages applies if:
H N (14)
H = (15)
Furthermore, the limiting factor of
H
can be the
signature-generating device or a guideline. There is
currently no quantum algorithm that allows finding a
hash collision in polynomial time, so they are consid-
ered to be resistant to quantum computers.
4.2 Crypto-agility
Crypto-agility describes the ability to continue us-
ing existing cryptographic methods and security pro-
tocols with adapting for PQC. Non quantum resis-
tant cryptographic algorithms will be replaced PQC
proved algorithms. Furthermore, there is the need to
achieve an affinity to modifying cryptographic sys-
tems by manufacturers. Proactive adapting and con-
tinuous improvement of the cryptographic methods
and security protocols are the effectively countered
against the quantum attacks. Three examples with
X.509v3, Secure / Multipurpose Internet Mail Exten-
sions (S/MIME) and Secure Shell (SSH) show a pos-
sible approach of crypto-agility.
4.2.1 Modifying X.509v3 Certificate
The X.509v3 certificate is a compilation of gen-
erally valid data formats and algorithms and not
a native security protocol. The integral compo-
nents of the X.509v3 certificate are Secure Socket
Layer/Transport Layer Security (SSL/TLS), S/MIME
and Extensible Markup Language (XML) Digital Sig-
natures. Based on the format description and the
data structure of the X.509v3 certificate a Public-Key-
Infrastructure (PKI) specification can be defined to
access digital certificates with quantum resistant al-
gorithm. There is no need to adapt the RFC compli-
ant standards. However, further adjustments must be
made at the security protocol level, such as upgrading
the TLS version to version 1.3 to ensure a hybrid ap-
proach. Furthermore, the very large key room for the
public-key must be reckoned to realized the protec-
tion against quantum attack (Campagna et al., 2015).
4.2.2 Modifying S/MIME Protocol
The S/MIME protocol is used via the PKI procedure
to ensure that digital signatures guarantee authentica-
tion, data integrity, binding nature and secure e-mail
encryption. A similar scheme and protocol is Open
Pretty Good Privacy (OpenPGP), which is based on
Web of Trust (WOB). To generate and verify a key
pair, a digital signature is required for S/MIME ver-
sion 3.2, which requires an asymmetric key of at least
1024 bits. The asymmetric algorithms Digital Sig-
nature Algorithm (DSA), RSA and RSA Probabilis-
tic Signature Scheme (RSA-PSS) with SHA-256 each
are available. The content of the message is exe-
cuted with the symmetric algorithm AES. Thus the
content is secure against quantum algorithms with
PECCS 2019 - 9th International Conference on Pervasive and Embedded Computing and Communication Systems
96
AES encryption, but the key exchange is not quan-
tum computer resistant. Note that the S/MIME proto-
col supports extended key sizes and encryption meth-
ods. Thus, the signature and key generation of the
S/MIME protocol can be updated with quantum com-
puter resistant algorithms. Therefore, the S/MIME
protocol supports elliptic curve algorithms for Cryp-
tographic Message Syntax (CMS), which is compat-
ible with the Public Key Cryptography Standards#7
(PKCS#7) certificate regarding to data format. The
implementation of CMS is possible for most applica-
tions. The strength of the security comes from the
encapsulation. Furthermore, many security features
are based on the parameters of CMS, which allow
the modification and selection of algorithms. Some
S/MIME versions prior to version 3.2 may only use
RSA (Campagna et al., 2015).
4.2.3 Modifying SSH Protocol
The SSH architecture consists of TLS, a user au-
thentication protocol and a connection protocol, each
of which uses its own algorithm in different net-
work layers. The TLS protocol ensures the protec-
tion goals of confidentiality, authentication and in-
tegrity and can be designed to be quantum-safe by a
parameterized adaptation. In contrast, the user au-
thentication and connection protocol with its cryp-
tographic primitives such as RSA, Digital Signature
Algorithm (DSA), ECDSA is not quantum secure.
Fundamentally, however, the SSH protocol is cryp-
tographically agile, since the SSH TLS protocol ne-
gotiates the first algorithm for key exchange between
client and server when a connection is to be estab-
lished. Thus, a quantum-safe algorithm can be used
for key exchange, if the protocol flexible enough to
cover the properties and requirements of PFS. Fur-
thermore, the use of the non-quantum-safe signature
algorithms DSA, ECDSA and RSA Probabilistic Sig-
nature Scheme (RSASSA-PSS) for the authentication
of the host can be replaced by quantum-safe digital
signatures. Thus, the standard as specified in RFC
4253 does not need to be fundamentally changed
when making the appropriate modifications to the
SSH protocol and its underlying layers (Campagna
et al., 2015). The adapting already existing encryp-
tion methods is a common way to be quantum-safe.
5 RESULTS
It generally applies: Without any proof that a crypto-
graphic algorithm via quantum algorithm is vulnera-
ble or to broken by a quantum attack, a cryptographic
method could be with sufficient research presumed to
be quantum attack resistant. Furthermore, every cryp-
tographic algorithm that based of mathematical com-
plexities like factorization and discrete logarithms are
believed as broken from the perspective of quantum
computer. This applies also to all security protocols
that use thus cryptographic algorithm. Typical asym-
metric encryptions such as RSA, ECC, public key
methods and their security protocols are not consid-
ered PQC ready. The asymmetric method can only re-
act to potential quantum attacks by increasing the size
of the key. Furthermore, the security of the method
and the key length can be guaranteed until the next
quantum attack and is dependent on the next higher
quantum computer generation. However, a large RSA
or ECC key affects the transfer of the key. Cryptosys-
tems which based on symmetric encryption like AES
are considered quantum resistant because the method
is turn-based and the key size can be increased. The
common security protocols, cryptosystems and appli-
cations use public-key method for the key exchange
and symmetric key for the data payload. In order
to be quantum-safe the public-key method must be
adapted by PQC approved cryptographic algorithm
and method. Furthermore, it must be ensure that all
security protocols are easy to substituted and be cryp-
tographic agility. The research for new cryptographic
primitives compliment the strategy of crypto-agility
and post-quantum cryptography.
6 CONCLUSIONS
At present, no quantum computers can break estab-
lished encryption. Furthermore, the perception of the
feasibility of a quantum computer is becoming more
concrete and this is the reason why it is comprehensi-
ble that quantum computer has such an enormous im-
pact on all areas of our society. The further develop-
ment of quantum algorithms are no longer just a topic
for universities or institutions of research and teach-
ing. Especially larger IT companies such as Google,
IBM, Microsoft, etc. are investing a great deal of time
and effort in creating a powerful quantum computer.
Just because we are currently still secure with our es-
tablished encryption mechanisms does not mean that
we do not have the cryptographic future to plan. It
must rather be accompanied by the timely and par-
allel development of the quantum computer and the
same procurement of resources.
Paradigm of Post-quantum Cryptography and Crypto-agility: Strategy Approach of Quantum-safe Techniques
97
REFERENCES
Alkim, E., Ducas, L., P
¨
oppelmann, T., and Schwabe, P.
(2015). Post-quantum key exchange - a new hope.
Cryptology ePrint Archive, Report 2015/1092.
Amy, M., Di Matteo, O., Gheorghiu, V., Mosca, M., Parent,
A., and Schanck, J. (2017). Estimating the cost of
generic quantum pre-image attacks on sha-2 and sha-
3. In Avanzi, R. and Heys, H., editors, Selected Areas
in Cryptography SAC 2016, pages 317–337, Cham.
Springer International Publishing.
Arrow, A., Butin, D., Gazdag, S., Rijneveld, J., and Mo-
haisen, A. (2018). Xmss: extended merkle signature
scheme. Request for Comments rfc8391, Corporation
of TU Darmstadt, Radboud University, University of
Central Florida, genua GmbH.
Bernstein, D. J., Hopwood, D., H
¨
ulsing, A., Lange, T.,
Niederhagen, R., Papachristodoulou, L., Schneider,
M., Schwabe, P., and Wilcox-O’Hearn, Z. (2015).
Sphincs: Practical stateless hash-based signatures. In
Oswald, E. and Fischlin, M., editors, Advances in
Cryptology EUROCRYPT 2015, pages 368–397,
Berlin, Heidelberg. Springer Berlin Heidelberg.
Bernstein, D. J., Lange, T., and Peters, C. (2008). Attacking
and defending the mceliece cryptosystem. In Buch-
mann, J. and Ding, J., editors, Post-Quantum Cryp-
tography, pages 31–46, Berlin, Heidelberg. Springer
Berlin Heidelberg.
Biryukov, A. and Khovratovich, D. (2009). Related-key
cryptanalysis of the full aes-192 and aes-256. In
Proceedings of the 15th International Conference on
the Theory and Application of Cryptology and In-
formation Security: Advances in Cryptology, ASI-
ACRYPT ’09, pages 1–18, Berlin, Heidelberg, Ger-
many. Springer-Verlag.
Bogdanov, A., Khovratovich, D., and Rechberger, C.
(2011). Biclique cryptanalysis of the full aes. In
Proceedings of the 17th International Conference on
The Theory and Application of Cryptology and In-
formation Security, ASIACRYPT’11, pages 344–371,
Berlin, Heidelberg, Germany. Springer-Verlag.
Boureanu, I., Owesarski, P., and Vaudenay, S., editors
(2014). Applied Cryptography and Network Security
- 12th International Conference, ACNS 2014, Lau-
sanne, Switzerland, June 10-13, 2014. Proceedings,
volume 8479 of Lecture Notes in Computer Science,
Berlin, Heidelberg, Germany. Springer.
Brands, G. (2011). Einf
¨
uhrung in die quanteninformatik.
In Quantenkryptografie, Teleportation und Quanten-
computing, pages 17–18, Berlin, Heidelberg. Springer
Berlin Heidelberg.
Brassard, G., HØyer, P., and Tapp, A. (1998). Quantum
cryptanalysis of hash and claw-free functions. In Luc-
chesi, C. L. and Moura, A. V., editors, LATIN’98:
Theoretical Informatics, pages 163–169, Berlin, Hei-
delberg. Springer Berlin Heidelberg.
Buchmann, J., Dahmen, E., and H
¨
ulsing, A. (2011). Xmss -
a practical forward secure signature scheme based on
minimal security assumptions. In Yang, B.-Y., editor,
Post-Quantum Cryptography, pages 117–129, Berlin,
Heidelberg. Springer Berlin Heidelberg.
Butin, D. (2017). Hash-based signatures: State of play.
IEEE Security Privacy, 15(4):37–43.
Campagna, M., Chen, L., Dagdelen,
¨
O., Ding, J., Fer-
nick, J. K., Gisin, N., Hayford, D., Jennewein, T.,
L
¨
utkenhaus, N., Mosca, M., Neill, B., Pecen, M., Perl-
ner, R., Ribordy, G., Schanck, J. M., Stebila, D., Wa-
lenta, N., Whyte, W., and Zhang, Z. (2015). Quantum
safe cryptography and security: An introduction, ben-
efits, enablers and challengers. Technical report, ETSI
(European Telecommunications Standards Institute).
Costello, C., Longa, P., and Naehrig, M. (2016). Efficient
algorithms for supersingular isogeny diffie-hellman.
In Robshaw, M. and Katz, J., editors, Advances in
Cryptology CRYPTO 2016, pages 572–601, Berlin,
Heidelberg. Springer Berlin Heidelberg.
Eckert, C. (2014). IT-Sicherheit: Konzepte - Verfahren -
Protokolle. De Gruyter Oldenbourg, Munich, 9nd edi-
tion.
Friis, N., Marty, O., Maier, C., Hempel, C., Holz
¨
apfel, M.,
Jurcevic, P., Plenio, M. B., Huber, M., Roos, C., Blatt,
R., and Lanyon, B. (2018). Observation of entangled
states of a fully controlled 20-qubit system. Phys. Rev.
X, 8:021012.
Fumy, W. (2017). Quantencomputer und die zukunft
der kryptographie. Datenschutz und Datensicherheit,
41(1):13–16.
Grover, L. K. (1996). A fast quantum mechanical algorithm
for database search. In Proceedings of the Twenty-
eighth Annual ACM Symposium on Theory of Com-
puting, STOC ’96, pages 212–219, New York, NY,
USA. ACM.
Grover, L. K. (1999). Quantum computing. The Sciences,
39(4):24–30.
Hoffstein, J., Pipher, J., and Silverman, J. H. (1998). Ntru:
A ring-based public key cryptosystem. In Buhler, J. P.,
editor, Algorithmic Number Theory, pages 267–288,
Berlin, Heidelberg. Springer Berlin Heidelberg.
IEEE (2008). Ieee draft standard specification for public-
key cryptographic techniques based on hard prob-
lems over lattices. IEEE Unapproved Draft Std
P1363.1/D12, Oct 2008.
Petzoldt, A., Chen, M.-S., Yang, B.-Y., Tao, C., and Ding,
J. (2015). Design principles for hfev- based multivari-
ate signature schemes. In Iwata, T. and Cheon, J. H.,
editors, Advances in Cryptology – ASIACRYPT 2015,
pages 311–334, Berlin, Heidelberg. Springer Berlin
Heidelberg.
Shor, P. W. (1994). Algorithms for quantum computation:
discrete logarithms and factoring. In Proceedings 35th
Annual Symposium on Foundations of Computer Sci-
ence, pages 124–134.
Shor, P. W. (1997). Polynomial-time algorithms for prime
factorization and discrete logarithms on a quantum
computer. SIAM J. Comput., 26(5):1484–1509.
PECCS 2019 - 9th International Conference on Pervasive and Embedded Computing and Communication Systems
98