ETSI TS 102 176-1 V2.1.1 (2011-07)
Technical Specification
Electronic Signatures and Infrastructures (ESI);
Algorithms and Parameters for Secure Electronic Signatures;
Part 1: Hash functions and asymmetric algorithms
ETSI
ETSI TS 102 176-1 V2.1.1 (2011-07)2
Reference
RTS/ESI-000080-1
Keywords
e-commerce, electronic signature, security
ETSI
650 Route des Lucioles
F-06921 Sophia Antipolis Cedex - FRANCE
Tel.: +33 4 92 94 42 00 Fax: +33 4 93 65 47 16
Siret N° 348 623 562 00017 - NAF 742 C
Association à but non lucratif enregistrée à la
Sous-Préfecture de Grasse (06) N° 7803/88
Important notice
Individual copies of the present document can be downloaded from:
http://www.etsi.org
The present document may be made available in more than one electronic version or in print. In any case of existing or
perceived difference in contents between such versions, the reference version is the Portable Document Format (PDF).
In case of dispute, the reference shall be the printing on ETSI printers of the PDF version kept on a specific network drive
within ETSI Secretariat.
Users of the present document should be aware that the document may be subject to revision or change of status.
Information on the current status of this and other ETSI documents is available at
http://portal.etsi.org/tb/status/status.asp
If you find errors in the present document, please send your comment to one of the following services:
http://portal.etsi.org/chaircor/ETSI_support.asp
Copyright Notification
No part may be reproduced except as authorized by written permission.
The copyright and the foregoing restriction extend to reproduction in all media.
© European Telecommunications Standards Institute 2011.
All rights reserved.
DECT
TM
, PLUGTESTS
TM
, UMTS
TM
and the ETSI logo are Trade Marks of ETSI registered for the benefit of its Members.
3GPP
TM
and LTE™ are Trade Marks of ETSI registered for the benefit of its Members and
of the 3GPP Organizational Partners.
GSM® and the GSM logo are Trade Marks registered and owned by the GSM Association.
ETSI
ETSI TS 102 176-1 V2.1.1 (2011-07)3
Contents
Intellectual Property Rights................................................................................................................................6
Foreword.............................................................................................................................................................6
Introduction ........................................................................................................................................................6
1 Scope........................................................................................................................................................8
2 References................................................................................................................................................9
2.1 Normative references .........................................................................................................................................9
2.2 Informative references......................................................................................................................................12
3 Definitions and abbreviations.................................................................................................................13
3.1 Definitions........................................................................................................................................................13
3.2 Abbreviations ...................................................................................................................................................13
4 Maintenance of the document ................................................................................................................14
5 Hash functions........................................................................................................................................15
5.1 General .............................................................................................................................................................15
5.2 Recommended one way hash functions............................................................................................................16
5.2.1 SHA-1 is no more recommended................................................................................................................16
5.2.2 RIPEMD-160 is no more recommended.....................................................................................................16
5.2.3 SHA-224.....................................................................................................................................................16
5.2.4 SHA-256.....................................................................................................................................................17
5.2.5 WHIRLPOOL.............................................................................................................................................17
5.2.6 SHA-384.....................................................................................................................................................17
5.2.7 SHA-512.....................................................................................................................................................17
5.2.8 SHA-3.........................................................................................................................................................17
6 Signature schemes..................................................................................................................................17
6.1 Signature algorithms.........................................................................................................................................17
6.1.1 General........................................................................................................................................................17
6.1.2 Recommended signature algorithms...........................................................................................................18
6.1.2.1 RSA.......................................................................................................................................................18
6.1.2.2 DSA.......................................................................................................................................................18
6.1.2.3 Elliptic curve analogue of DSA based on a group E(Fp) ......................................................................19
6.1.2.4 Elliptic curve analogue of DSA based on a group E(F2m) ....................................................................20
6.1.2.5 EC-GDSA based on a group E(Fp).......................................................................................................20
6.1.2.6 EC-GDSA based on a group E(F2m).....................................................................................................20
6.2 Recommended key pair generation methods....................................................................................................21
6.2.1 General........................................................................................................................................................21
6.2.2 Recommended key pair generation methods ..............................................................................................21
6.2.2.1 Key and parameter generation algorithm rsagen1.................................................................................21
6.2.2.2 Key and parameter generation algorithm dsagen1 ................................................................................22
6.2.2.3 Key and parameter generation algorithm ecgen1 for ecdsa-Fp.............................................................22
6.2.2.4 Key and parameter generation algorithm ecgen2 for ecdsa-F2m..........................................................22
6.2.2.5 Key and parameter generation algorithm ecgen1 for ecgdsa-Fp...........................................................22
6.2.2.6 Key and parameter generation algorithm ecgen2 for ecgdsa-F2m........................................................23
7 Signature suites ......................................................................................................................................23
7.1 General .............................................................................................................................................................23
7.2 Padding methods ..............................................................................................................................................23
7.3 Recommended signature suites ........................................................................................................................24
8 Random number generation methods.....................................................................................................25
8.1 General .............................................................................................................................................................25
8.2 Recommended random number generation methods .......................................................................................25
8.2.1 Random generator requirements trueran.....................................................................................................26
8.2.2 Random generator requirements pseuran....................................................................................................27
ETSI
ETSI TS 102 176-1 V2.1.1 (2011-07)4
9 Recommended hash functions and key sizes versus time ......................................................................28
9.1 Basis for the recommendations ........................................................................................................................28
9.2 Recommended hash functions versus time.......................................................................................................29
9.3 Recommended key sizes versus time ...............................................................................................................29
10 Time period resistance of hash functions and keys................................................................................32
10.1 Time period resistance for hash functions........................................................................................................32
10.2 Time period resistance for signer's key ............................................................................................................32
10.3 Time period resistance for trust anchors...........................................................................................................32
10.4 Time period resistance for other keys...............................................................................................................32
11 Practical ways to identify hash functions and signature algorithms.......................................................33
11.1 Hash functions and signature algorithms objects identified using OIDs..........................................................33
11.1.1 Hash functions ............................................................................................................................................33
11.1.2 Signature algorithms...................................................................................................................................33
11.1.3 Signature suites...........................................................................................................................................34
11.2 Hash functions and signature algorithms identified objects using URNs.........................................................35
11.2.1 Hash functions ............................................................................................................................................35
11.2.2 Signature algorithms...................................................................................................................................35
11.2.3 Signature suites...........................................................................................................................................35
11.3 Recommended hash functions and signature algorithms objects that do not yet have an OID or a
description........................................................................................................................................................36
11.4 Recommended hash functions and signature algorithms objects that do not yet have a URN or a
description........................................................................................................................................................36
Annex A (normative): Algorithms for various data structures..................................................................37
A.1 Advanced Electronic Signatures based on TS 101 733..........................................................................37
A.2 Advanced Electronic Signatures based on TS 101 903..........................................................................38
A.3 Signer's certificates.................................................................................................................................38
A.4 CRLs.......................................................................................................................................................39
A.5 OCSP responses .....................................................................................................................................39
A.6 CA certificates........................................................................................................................................40
A.7 Self-signed certificates for CA issuing CA certificates..........................................................................40
A.8 TSTs based on RFC 3161 and TS 101 861 ............................................................................................41
A.9 TSU certificates......................................................................................................................................41
A.10 Self-signed certificates for CAs issuing TSU certificates ......................................................................41
A.11 Attribute certificates...............................................................................................................................42
A.12 AA certificates........................................................................................................................................42
Annex B (informative): Recommended key sizes (historical).....................................................................43
B.1 Changes in 2005.....................................................................................................................................43
B.2 Changes in 2007.....................................................................................................................................45
B.3 Changes in 2011.....................................................................................................................................45
Annex C (informative): Generation of RSA modulus.................................................................................46
Annex D (informative): Generation of elliptic curve domain parameters ................................................47
D.1 ECDSA and ECGDSA based on a group E(Fp).....................................................................................47
D.2 ECDSA and ECGDSA based on a group E(F2
m)..................................................................................48
D.3 The class number condition....................................................................................................................49
ETSI
ETSI TS 102 176-1 V2.1.1 (2011-07)5
Annex E (informative): On the generation of random data .......................................................................51
E.1 Classes of random number generators....................................................................................................51
E.2 On tests for NRNGs ...............................................................................................................................52
Annex F (informative): Algorithm identifiers defined in various documents...........................................53
F.1 Algorithm identifiers defined in RFC 3278............................................................................................53
F.2 Algorithm identifiers defined in RFC 3279............................................................................................53
F.3 Algorithm identifiers defined in RFC 3370............................................................................................54
F.4 Algorithm identifiers defined in RFC 3447............................................................................................54
F.5 Algorithm identifier defined in RFC 3874.............................................................................................54
F.6 Algorithm identifiers defined in XML-Signature Syntax and Processing W3C Recommendation.......54
F.7 Algorithm identifiers defined in XML Encryption Syntax and Processing. W3C Recommendation....55
F.8 Algorithm identifiers defined in RFC 4050............................................................................................55
F.9 Algorithm identifiers defined in RFC 4051............................................................................................55
F.10 Algorithm identifiers defined in RFC 4055............................................................................................56
Annex G (informative): Abstracts of ISO/IEC 10118-3 and ISO/IEC 9796-2 ..........................................57
Annex H (informative): Signature maintenance..........................................................................................58
Annex I (informative): Major changes from previous versions .................................................................59
Annex J (informative): National Bodies .......................................................................................................60
Annex K (informative): Bibliography...........................................................................................................64
History..............................................................................................................................................................66
ETSI
ETSI TS 102 176-1 V2.1.1 (2011-07)6
Intellectual Property Rights
IPRs essential or potentially essential to the present document may have been declared to ETSI. The information
pertaining to these essential IPRs, if any, is publicly available for ETSI members and non-members, and can be found
in ETSI SR 000 314: "Intellectual Property Rights (IPRs); Essential, or potentially Essential, IPRs notified to ETSI in
respect of ETSI standards", which is available from the ETSI Secretariat. Latest updates are available on the ETSI Web
server (http://ipr.etsi.org).
Pursuant to the ETSI IPR Policy, no investigation, including IPR searches, has been carried out by ETSI. No guarantee
can be given as to the existence of other IPRs not referenced in ETSI SR 000 314 (or the updates on the ETSI Web
server) which are, or may be, or may become, essential to the present document.
Foreword
This Technical Specification (TS) has been produced by ETSI Technical Committee Electronic Signatures and
Infrastructures (ESI).
The present document is part 1 of a multi-part deliverable covering the Algorithms and Parameters for Secure
Electronic Signatures, as identified below:
Part 1: "Hash functions and asymmetric algorithms";
Part 2: "Secure channel protocols and algorithms for signature creation devices".
Introduction
The present document provides for security and interoperability for the application of the underlying mathematical
algorithms and related parameters for electronic signatures in accordance with the Directive 1999/93/EC [1] of the
European Parliament and of the Council of 13 December 1999 on a Community framework for electronic signatures.
On the other side the present document is not a legal document answering the question which key lengths or use dates
are sufficient to ensure a certain level of liability. In particular the reader is advised that some national signature laws or
regulations may require a different level of security for qualified electronic signatures than recommended here by the
key lengths and use dates in the present document. The present document is based on cryptographic analysis of the
algorithms and it recommends using at least the parameters given here.
The present document defines a list of hash functions, as well as a list of signature schemes together with the
requirements on their parameters, as well as the recommended combinations of these schemes with hash functions and
padding method in the form of "signature suites" to be used with the data structures defined in the documents developed
under the European Electronic Signature Standardization Initiative (EESSI). The present document contains several
informative annexes which provide useful information on a number of subjects mentioned in the text.
The present document is not a general purpose document dealing with hash functions and asymmetrical algorithms in
general. The goal of the present document is not to list all "good" signature algorithms but those that are most important
to be used in the context of advanced electronic signatures. In addition, the intent of the present document is not to have
a catalog of all algorithms suitable for advanced electronic signatures, but to limit the list to a reasonable set so that
interoperability can be achieved. Interoperability with security is the main issue.
The primary criterion for inclusion of an algorithm in the document is "Secure, widely used and deployed in practice".
Whereas all listed algorithms have been checked for security by cryptographic experts, it cannot be concluded from the
document, that an algorithm not listed would be insecure. Therefore algorithms are not listed as recommended if they
require a restricted environment or remain secure for a short time frame only.
ETSI
ETSI TS 102 176-1 V2.1.1 (2011-07)7
The second part of this multi-part deliverable (protocols and algorithms for SCDev secure channels) is outdated.
It defined at the date of issuance protocols and symmetric algorithms that may optionally be used to construct a secure
channel providing either only integrity or both integrity and confidentiality between an application and a signature
creation device (SCDev). Such a secure channel may be used during the operational phase of a signature creation
device:
• when the key pair is not generated by the SCDev, to remotely download in the SCDev both a private key and
the associated public key certificate;
• when the key pair is generated by the SCDev, to remotely download in the SCDev a public key certificate and
associate it with the previously generated private key.
The protocols and symmetric algorithms in the former scope of part 2 of this multi-part deliverable are defined now in
the European Norm EN 14890-1 [16]. Additionally second part is in preparation.
ETSI
ETSI TS 102 176-1 V2.1.1 (2011-07)8
1 Scope
The present document is targeted to support advanced electronic signatures and the related infrastructure.
The present document defines a list of hash functions and a list of signature schemes, as well as the recommended
combinations of hash functions and signatures schemes in the form of "signature suites".
The primary criteria for inclusion of an algorithm in the present document are:
• the algorithm is considered as secure;
• the algorithm is commonly used; and
• the algorithm can easily be referenced (for example by means of an OID).
This does not mean that other hash functions and signature suites cannot be used, but either they do not correspond to
the above criteria or their security has not been assessed.
The document also provides guidance on the hash functions, signature schemes and signature suites to be used with the
data structures used in the context of electronic signatures. For each data structure, the set of algorithms to be used is
specified. Each set is identified by an identifier which is either an OID (Object IDENTIFIER) or a URI /URN.
The use of such identifiers is necessary so that interoperability can be achieved. In order to allow for data interchange,
the document references algorithms in terms of OIDs and URIs / URNs together with algorithm parameters.
Different requirements apply to the issuers and to the users of the data structures in order to allow for interoperability.
RFC documents use the terms SHALL, SHOULD, MAY, RECOMMENDED in order to allow for interoperability. The
same terminology is used in the present document (see RFC 2119 [25]).
Issuers of the data structures (e.g. CSPs, CRL Issuers, OCSP responders, TSUs) need to know the algorithms and key
sizes they SHOULD or MAY support. There SHOULD be at least one algorithm recommended to support, but may be
more than one.
Users of the data structures (i.e. signers or verifiers of electronic signatures) need to know the algorithms and key sizes
they SHALL, SHOULD or MAY support. Users may support more than one algorithm for each data structure.
These requirements are listed in annex A.
Annex B provides historical information on the recommended hash functions, algorithms and key sizes for the
generation and verification of electronic signatures. This annex will be periodically updated.
Annex C provides more information on the generation of RSA modulus.
Annex D provides more information on the generation of elliptic curve domain parameters.
Annex E addresses the generation of random data.
Annex F lists the algorithm identifiers defined in various documents.
Annex G provides a short abstract of ISO/IEC 10118-3 [3] and ISO/IEC 9796-2 [17].
Annex H provides some guidance on signature maintenance.
Annex I lists the major changes from the previous versions.
The present document defines a set of algorithms (i.e. hash functions, signature schemes and signature suites) and the
corresponding parameters that are recommended to be used. If such algorithms are used according to the context where
they are expected to be used, then a reasonable security level can be assumed.
The algorithms defined in the present document are usable in particular with the following documents:
• TS 101 733 [18]: "Electronic Signatures and Infrastructures (ESI); Electronic Signature Formats";
• TS 101 903 [19]: "XML Advanced Electronic Signatures (XAdES)";
ETSI
ETSI TS 102 176-1 V2.1.1 (2011-07)9
NOTE: XML language is defined in RFC 3275 [10].
• TS 101 861 [20]: "Time stamping profile";
• TS 101 456 [32]: "Electronic Signatures and Infrastructures (ESI); Policy requirements for certification
authorities issuing qualified certificates";
• TS 102 042 [33]: "Electronic Signatures and Infrastructures (ESI); Policy requirements for certification
authorities issuing public key certificates";
• EN 14169 [34]: Protection profiles for Secure signature creation device;
• CWA 14170 [35]: "Security requirements for signature creation applications";
• CWA 14171 [36]: "Procedures for electronic signature verification";
• CWA 14167-1 [37]: "Security Requirements for Trustworthy Systems Managing Certificates for Electronic
Signatures - Part 1: System Security Requirements";
• CWA 14167-2 [38]: "Security Requirements for Trustworthy Systems Managing Certificates for Electronic
Signatures - Part 2: Cryptographic module for CSP Signing Operations with Backup - Protection Profile";
• CWA 14167-3 [39]: "Security Requirements for Trustworthy Systems Managing Certificates for Electronic
Signatures - Part 3: Cryptographic module for CSP key generation services - Protection profile
(CMCKG-PP)";
• CWA 14167-4 [40]: "Security Requirements for Trustworthy Systems Managing Certificates for Electronic
Signatures - Part 4: Cryptographic module for CSP signing operations - Protection profile - CMCSO PP";
• RFC 5280 [2]: "Internet X.509 Public Key Infrastructure Certificate and Certificate Revocation List (CRL)
Profile";
• RFC 5755 [i.22]: "An Internet Attribute Certificate profile for authorization";
• RFC 3161 [9]: "Internet X.509 Public Key Infrastructure Time-Stamp Protocol (TSP)";
• RFC 2560 [22]: "X.509 Internet Public Key Infrastructure Online Certificate Status Protocol - OCSP".
The CWA documents are processed by CEN/TC224 for re-issuing as European Norms.
Patent related issues are out of the scope of the present document.
2 References
References are either specific (identified by date of publication and/or edition number or version number) or
non-specific. For specific references, only the cited version applies. For non-specific references, the latest version of the
reference document (including any amendments) applies.
Referenced documents which are not found to be publicly available in the expected location might be found at
http://docbox.etsi.org/Reference.
NOTE: While any hyperlinks included in this clause were valid at the time of publication, ETSI cannot guarantee
their long term validity.
2.1 Normative references
The following referenced documents are necessary for the application of the present document.
[1] Directive 1999/93/EC of the European Parliament and of the Council of 13 December 1999 on a
Community framework for electronic signatures.
ETSI
ETSI TS 102 176-1 V2.1.1 (2011-07)10
[2] IETF RFC 5280 (2008): "Internet X.509 Public Key Infrastructure Certificate and Certificate
Revocation List (CRL) Profile". .
[3] ISO/IEC 10118-3 (2004): "Information technology - Security techniques - Hash functions Part
3: Dedicated hash functions".
NOTE: See annex G for main content description.
[4] FIPS Publication 180-3 (2008): "Secure Hash Standard (SHS)".
[5] IEEE P1363 (2000): "Standard Specifications for Public-Key Cryptography".
[6] FIPS Publication 186-3 (2009): "Digital Signature Standard (DSS)".
[7] ANSI X9.62 (2005): "Public Key Cryptography for the Financial Services Industry: The Elliptic
Curve Digital Signature Algorithm (ECDSA)".
NOTE: This standard is consistent with Elliptic Curve Cryptography (SEC 1) Part 1, Version 1.0, Certicom, 2000
(http://www.secg.org/collateral/sec1_final.pdf). The current draft is Version 1.7
(http://www.secg.org/download/aid-631/sec1_1point7.pdf).
[8] ISO/IEC 14888-3 (2006) "Information technology - Security techniques - Digital signatures with
appendix - Part 3: Discrete logarithm based mechanisms".
[9] IETF RFC 3161 (2001): "Internet X.509 Public Key Infrastructure Time-Stamp Protocol (TSP)".
[10] IETF RFC 3275 (2002): "(Extensible Markup Language) XML-Signature Syntax and Processing".
[11] IETF RFC 5752 (2010): "Use of Elliptic Curve Cryptography (ECC) Algorithms in Cryptographic
Message Syntax (CMS)".
[12] IETF RFC 3279 (2002): "Algorithms and Identifiers for the Internet X.509 Public Key
Infrastructure Certificate and Certificate Revocation List (CRL) Profile".
NOTE: Updated by RFC 4055, RFC 4491, RFC 5480, RFC 5758.
[13] IETF RFC 3370 (2002): "Cryptographic Message Syntax (CMS) Algorithms".
NOTE: Updated by RFC 5754.
[14] IETF RFC 3447 (2003): "Public-Key Cryptography Standards (PKCS) #1: RSA Cryptography
Specifications Version 2.1".
[15] IETF RFC 4055 (2005): "Additional Algorithms and Identifiers for RSA Cryptography for use in
the Internet X.509 Public Key Infrastructure - Certificate and Certificate Revocation List (CRL)
Profile".
[16] CEN EN 14890-1 (2008): "Application Interface for smart cards used as Secure Signature
Creation Devices - Part 1: Basic services".
[17] ISO/IEC 9796-2 (2002): "Information technology - Security techniques - Digital signature
schemes giving message recovery - Part 2: Integer factorization based mechanisms".
NOTE 1: Not to be confused with ISO/IEC 9796-2 (1997): Mechanisms using a hash-function.
NOTE 2: See annex G for main content description.
[18] ETSI TS 101 733: "Electronic Signatures and Infrastructures (ESI); CMS Advanced Electronic
Signatures (CAdES)".
[19] ETSI TS 101 903: "Electronic Signatures and Infrastructures (ESI); XML Advanced Electronic
Signatures (XAdES)".
[20] ETSI TS 101 861: "Electronic Signatures and Infrastructures (ESI); Time stamping profile".
[21] IETF RFC 3281 (2002): "An Internet Attribute Certificate profile for Authorization".
ETSI
ETSI TS 102 176-1 V2.1.1 (2011-07)11
[22] IETF RFC 2560 (1999): "X.509 Internet Public Key Infrastructure Online Certificate Status
Protocol - OCSP".
[23] IETF RFC 3852 (2004): "Cryptographic Message Syntax (CMS)".
[24] Void.
[25] IETF RFC 2119 (1997): "Key words for use in RFCs to Indicate Requirement Levels".
[26] Common PKI Specifications for Interoperable Applications, Version 2.0 (2009),
TeleTrusT e.V. Deutschland.
NOTE: See http://www.common-pki.org.
[27] IETF RFC 3874 (2005): "A 224-bit One-way Hash Function: SHA-224".
[28] IETF RFC 4050 (2005): "Using the Elliptic Curve Signature Algorithm (ECDSA) for XML
Digital Signatures".
[29] IETF RFC 4051 (2005): "Additional XML Security Uniform Resource Identifiers (URIs)".
[30] W3C Recommendation - 12 February 2002: "XML-Signature Syntax and Processing".
NOTE: See http://www.w3.org/TR/2002/REC-xmldsig-core-20020212/.
[31] W3C Recommendation - 10 December 2002: "XML Encryption Syntax and Processing".
NOTE: See http://www.w3.org/TR/2002/REC-xmlenc-core-20021210/.
[32] ETSI TS 101 456: "Electronic Signatures and Infrastructures (ESI); Policy requirements for
certification authorities issuing qualified certificates".
[33] ETSI TS 102 042: "Electronic Signatures and Infrastructures (ESI); Policy requirements for
certification authorities issuing public key certificates".
[34] CWA 14169: "Secure Signature-Creation Devices "EAL 4+"".
[35] CWA 14170: "Security requirements for signature creation applications".
[36] CWA 14171: "Procedures for electronic signature verification".
[37] CWA 14167-1: "Security Requirements for Trustworthy Systems Managing Certificates for
Electronic Signatures - Part 1: System Security Requirements".
[38] CWA 14167-2: "Security Requirements for Trustworthy Systems Managing Certificates for
Electronic Signatures - Part 2: Cryptographic module for CSP Signing Operations with Backup Protection
Profile".
[39] CWA 14167-3: "Security Requirements for Trustworthy Systems Managing Certificates for
Electronic Signatures - Part 3: Cryptographic module for CSP key generation services - Protection
profile (CMCKG-PP)".
[40] CWA 14167-4: "Security Requirements for Trustworthy Systems Managing Certificates for
Electronic Signatures - Part 4: Cryptographic module for CSP signing operations - Protection
profile - CMCSO PP".
NOTE: The CWA documents are processed by CEN/TC224 for re-issuing as European Norms.
[41] DIN 66291-1: "Chipcards with digital signatur application/function according to SigG and SigV Part
1: Application interface".
[42] ANSI X9.175: "Pseudo Random Number Generator (RNG)".
[43] IETF RFC 3278: "Use of Elliptic Curve Cryptography (ECC) Algorithms in Cryptographic
Message Syntax (CMS)".
ETSI
ETSI TS 102 176-1 V2.1.1 (2011-07)12
[44] Required Canonical XML (omits comments).
NOTE: Available at: http://www.w3.org/TR/2001/REC-xml-c14n-20010315.
[45] Recommended Canonical XML with Comments.
NOTE: Available at: http://www.w3.org/TR/2001/REC-xml-c14n-20010315#WithComments.
2.2 Informative references
The following referenced documents are not necessary for the application of the present document but they assist the
user with regard to a particular subject area.
[i.1] Void.
[i.2] Void.
[i.3] Void.
[i.4] Void.
[i.5] DigitPA. Determinazione n. 69/2010: "Modifiche alla Deliberazione n. 45 del Centro Nazionale
per l'Informatica nella pubblica Amministrazione".
NOTE: Available at: http://www.digitpa.gov.it/sites/default/files/normativa/DTcommissario%2069-
2010%20Modifiche%20del%20%20Cnipa%20n%2045-
2009%20su%20regole%20tecniche%20documento%20informatico%20.pdf
[i.6] ECRYPT II, Yearly Report on Algorithms and Keysizes (2010), Rev. 1.0, ICT-2007-216676,
2010-03-30.
NOTE: See http://www.ecrypt.eu.org/documents/D.SPA.13.pdf.
[i.7] NIST, Transitions: "Recommendation for Transitioning the Use of Cryptographic Algorithms and
Key Lengths", NIST Special Publication 800-131A, 2011-01-13.
NOTE: See http://csrc.nist.gov/publications/nistpubs/800-131A/sp800-131A.pdf.
[i.8] ISO/IEC 18032 (2005): "Information technology - Security techniques - Prime number
generation".
[i.9] Wiener, M.: "Cryptanalysis of short RSA secret exponents", IEEE Transactions on Information
Theory, Vol. 36 1990.
[i.10] Boneh, D. and Durfee, G.: "Cryptanalysis of RSA with private key less than N0.292", Proc.
Eurocrypt '99, LNCS, J. Stern (ed.), Springer-Verlag, 1999. Final version in IEEE Trans.
Information Theory, Vol. 46 2000.
[i.11] Durfee, G. and Nguyen, P.: "Cryptanalysis of RSA with short private exponent", Proc. Asiacrypt
'99, LNCS, Springer-Verlag, 1999.
[i.12] R.L. Rivest, R.D. Silverman: "Are 'Strong' Primes needed for RSA?", Peprint 1999 December 1,
1998.
[i.13] IETF RFC 5639 (2010): "Elliptic Curve Cryptography (ECC) Brainpool Standard Curves and
Curve Generation".
[i.14] ISO/IEC 18031 (2005): "Information technology - Security techniques - Random bit generation".
[i.15] ANSI X9.82 (2006) Random Number Generation Parts 1.
[i.16] AIS 20: "Application Notes and Interpretation of the Scheme: Functionality classes and evaluation
methodology for deterministic random number generators", Version 1.
NOTE: Available at:
https://www.bsi.bund.de/SharedDocs/Downloads/DE/BSI/Zertifizierung/Interpretationen/ais20e_pdf.pdf.
ETSI
ETSI TS 102 176-1 V2.1.1 (2011-07)13
[i.17] ANSI X9.17-1985: "Financial Institution Key Management (wholesale)".
[i.18] Ruhkin, A. et al.: "A Statistical Test Suite for Random and Pseudorandom Number Generators for
Cryptographic Applications", NIST Special Publication 800-22 with revisions dated 15.05.2001;
available at Menezes, A., van Oorschot, P. and Vanstone, S., "Handbook of Applied Cryptography
(chapter 5)", CRC Press, 1997.
NOTE: See http://www.cacr.math.uwaterloo.ca/hac.
[i.19] AIS 31: "Functionality Classes and Evaluation Methodology for Physical Random Number
Generators", Version 3.1 (25.09.2001).
NOTE: Available at:
https://www.bsi.bund.de/SharedDocs/Downloads/DE/BSI/Zertifizierung/Interpretationen/ais31e_pdf.pdf.
[i.20] Maurer, U.: "A universal statistical test for random bit generators", Advances in Cryptology Crypto
'90, LNCS 537, pp. 409-420, 1991.
[i.21] ISO/IEC 15946-2 (withdrawn): "Information technology - Security techniques - Cryptographic
techniques based on elliptic curves -- Part 2: Digital signatures".
[i.22] IETF RFC 5755: "An Internet Attribute Certificate Profile for Authorization".
[i.23] "Elliptic Curve Cryptography (ECC)"
Bundesamt für Sicherheit in der Informationstechnik, TR-03111(2006).
[i.24] J.S. Coron: "Optimal Security Proofs for PSS and Other Signature Suites", EUROCRYPT 2002,
LNCS 2332, Springer, Berlin, 2002.
[i.25] H. Cohen: "A Course in Computational Algebraic Number Theory", Springer, Berlin, 1993.
[i.26] IETF RFC 2459: "Internet X.509 Public Key Infrastructure Certificate and CRL Profile".
3 Definitions and abbreviations
3.1 Definitions
For the purposes of the present document, the following terms and definitions apply:
bit length: bit length of an integer p is r if 2r-1 ≤ p < 2r
signature policy: set of rules for the creation and validation of an electronic signature, that defines the technical and
procedural requirements for electronic signature creation and validation, in order to meet a particular business need, and
under which the signature can be determined to be valid
signature scheme: triplet of three algorithms composed of a signature creation algorithm, a signature verification
algorithm and a key generation algorithm
NOTE: The key generation algorithm generates the keys for the two others algorithms.
signature suite: combination of a signature scheme with a padding method and a cryptographic hash function
3.2 Abbreviations
For the purposes of the present document, the following abbreviations apply:
AA Attribute Authority
AC Attribute Certificates
ANSI American National Standards Institute
CA Certification Authority
CRL Certificate Revocation List
ETSI
ETSI TS 102 176-1 V2.1.1 (2011-07)14
CRT Chinese Remainder Theorem
CSP Certification-Service-Provider
CWA CEN Workshop Agreement
DRNG Deterministic Random Number Generator
DSA Digital Signature Algorithm
ECC Elliptic Curve Cryptography
ECDSA Elliptic Curve Digital Signature Algorithm
ECGDSA Elliptic Curve German Digital Signature Algorithm
IETF Internet Engineering Task Force
ISO International Organization for Standardization
MGF Mask Generation Function
NIST National Institute of Standards and Technology
NRNG Non-deterministic Random Number Generator
OCSP Online Certificate Status Protocol
OID Object IDentifier
RfC Request for Comments
RNG Random Number Generator
RSA Rivest, Shamir and Adleman algorithm
SCDev Secure Signature Creation Device
TST Time-Stamp Token
TSU Time-Stamping Unit
URI Uniform Resource Identifier
URN Uniform Resource Number
4 Maintenance of the document
As a response to relevant developments in the area of cryptography and technology, activities for the maintenance of the
present document shall enable dynamic updating of the lists of recommended algorithms and signature suites. The
current list of recommended cryptographic hash functions and signature algorithms is given in the present document.
The maintenance activities will introduce new cryptographic hash functions and signature algorithms and will lead to
remove cryptographic hash functions and signature algorithms from the list and need to respond to the following
situations:
1) The need to introduce new algorithms and relevant parameters will call for a mechanism that is rather
dynamic. Since it is important to maintain interoperability, updates may result from the adoption or removal of
an algorithm in a document on which an ETSI TS or CWA is based upon.
2) Advances in cryptography will call for a phasing out of some algorithms or parameters. Such phasing out will
normally be known well in advance.
3) In the case of new attacks the immediate need to remove an algorithm could arise.
The maintenance activity will be carried by ETSI. For cases 1 and 2, ETSI shall review the present document at regular
intervals of at most two years. In case 3, ETSI shall update the present document as soon as possible. In that case,
affected algorithms should no longer be used even before the present document will be revised. However, immediate
response is up to legislation and/or supervision of CSPs. For that reason an informative annex J is included that list
National Bodies, known at the issuing time of this Technical Specification.
The maintenance activity for cases 1 and 2 allow for some transition period during which the algorithm/parameter
values to be revised may be used in already issued certificates and electronic-signature products. However, the revised
items should not be considered for use in certificates to be issued and new electronic-signature products. The transition
period shall allow vendors using the revised items to alter their production process. If the security implications of a
revision are considered very serious, the certificates and electronic-signature products using the revised item should be
withdrawn before their planned expiry date and other measures, like archival time stamps according to TS 101 733 [18]
should be undertaken to preserve the security status of electronic signatures based on it.
In order to allow an easy follow up of the present document, a history of the tables provided in the main body of the
document will be maintained and kept as annexes.
ETSI
ETSI TS 102 176-1 V2.1.1 (2011-07)15
5 Hash functions
5.1 General
A hash function takes as input a variable-length message and produces as output a fixed-length hash value.
NOTE: In the present document, "hash function" means a hash function with the three properties defined in this
clause (i.e. clause 5.1).
Hash functions may be used in a variety of cases, such as:
• Advanced Electronic Signatures include the identifier of the hash function used to compute the digital
signature.
• Time-Stamp tokens include the identifier of the hash algorithm used to compute the hash value for the
time-stamped data.
• Public key certificates include the identifier of a signature suite which defines the hash function used to
compute the digital signature.
For the purpose of generating signatures the following (informally defined) three properties are required from the hash
function h:
1) Pre-image resistance: Given y = h(m) (but not m) it is practically infeasible to find m. Without this property,
a signature scheme may otherwise be vulnerable to an attack based on generating the signature "backwards",
applying the verification function to a randomly chosen signature value.
2) 2nd pre-image resistance: Given h(m) and m, it is practically infeasible to find another m' ≠ m such that
h(m) = h(m'). For signatures, this property protects from re-using an already existing signature for another
message.
3) Collision resistance: It is practically infeasible to find any pair of distinct values m, m' such that h(m) = h(m').
This property is obviously needed to protect signature against chosen message attacks.
While one can construct examples of functions that are collision resistant, but not pre-image resistant, one would for
practical purposes nevertheless expect that the above list of properties is ordered by difficulty for an attacker,
i.e. breaking pre-image resistance is the most difficult. New attacks against hash function MD5 and SHA-1 succeeded,
it was shown that MD5 is not collision resistant by constructing classes of messages-pairs with the same hash value.
Whereas the loss of collision resistance does not imply that a pre-image or second pre-image can easier be constructed,
it is recommended to migrate to other hash functions, if the collision resistance becomes weaker.
In addition to this, more subtle properties are often required as a consequence of mathematical properties of the
signature scheme itself. For instance, "h" should not preserve algebraic structure. The perhaps best known example is
the multiplicatively of the (naive) RSA scheme, which would otherwise give a valid signature for a x b from two valid
signatures of a and b.
The above properties have led to some signature schemes being defined and proven secure in the so-called Random
Oracle Model, where one assumes "h" "behaves" like a completely random function. Intuitively, a completely random
function should have all of the above properties as long as the range of the function is large enough.
The list of currently recommended hash functions is given in table 1. Each hash function has a unique entry index
represented by a string beginning with "1." followed by a two-digit entry number.
ETSI
ETSI TS 102 176-1 V2.1.1 (2011-07)16
Table 1: The list of recommended hash functions
Hash function entry index Short hash function entry name Adoption date Normative references
1.03 sha224 2004 FIPS Publication 180-3 [4]
1.04 sha256 2004 ISO/IEC 10118-3 [3] and
FIPS Publication 180-3 [4]
1.05 whirlpool 2004 ISO/IEC 10118-3 [3]
1.06 sha384 31.03.2007 FIPS Publication 180-3 [4]
1.07 sha512 31.03.2007 FIPS Publication 180-3 [4]
NOTE: Additional secure hash algorithms, beside the SHA-2 family, were needed. For that reason the Whirlpool
algorithm has been added. This algorithm has been reviewed by NESSIE experts.
5.2 Recommended one way hash functions
5.2.1 SHA-1 is no more recommended
SHA-1 MAY be used to hash a message, M, having a length of up to 264-1 bits.
The final result of SHA-1 is a 160-bit message digest. The SHA-1 algorithm is described in ISO/IEC 10118-3 [3] and
FIPS Publication 180-3 [4].
NOTE 1: Several attacks against SHA-1 have been discovered since February 2005. Although a collision of the full
80-round SHA-1 has not yet been published, at least SHA-224 and SHA-256 SHOULD be implemented
for any new product for electronic signatures.
NOTE 2: All known collision attacks on SHA-1 require full control of certain substrings within the data to be
hashed and knowledge of the data bits prior to these strings. This is be considered as a realistic attack
scenario for documents signed by signers (in particular, when a kind of "active" program element may be
hidden in the document). On the other hand for X.509 certificates such attacks can be prevented by the
CA by including a reasonable amount of entropy (i.e. data bits neither known to nor predictable by the
attacker) in the certificate string prior to any data bits controllable by the attacker. This method leads to a
considerably higher resistance of certificates against collision attacks. Collision attacks on SHA-1 were
recently improved and colliding messages may be constructed in the near future. Therefore SHA-1
SHOULD NOT be used for document signatures. SHA-1 remains resistant against second pre-image
attacks, so it can be used for certificate creation and verification if the certificate body includes sufficient
entropy (i.e. data bits neither known to nor predictable by the attacker) prior to any data bits controllable
by the attacker.
5.2.2 RIPEMD-160 is no more recommended
RIPEMD-160 is no more recommended due to the short hash value length.
NOTE: RIPEMD-320 is constructed from RIPEMD-160 by initializing the two parallel lines with different initial
values, omitting the combination of the two lines at the end of every application of the compression
function, and exchanging a chaining variable between the 2 parallel lines after each round. The security
level of the 320-bit extension of RIPEMD-160 is the same as that of RIPEMD-160 itself. Similarly the
256-bit extension of RIPEMD-128, i.e. RIPEMD-256 is the same as that of RIPEMD-128.
5.2.3 SHA-224
SHA-224 MAY be used to hash a message, M, having a length of up to 264-1 bits and the output size is 224 bits. The
function is defined in the exact same manner as SHA-256 (clause 5.2.4), except for the initial value and the truncation
of the final hash value. The SHA-224 algorithm is described in FIPS Publication 180-3 [4].
The specification for SHA-224 is identical to SHA-256, except that different initial values are used, and the final hash
value is truncated to 224 bits. Therefore it is NOT RECOMMENDED to use SHA-224, if SHA-256 can be used instead
without truncation.
ETSI
ETSI TS 102 176-1 V2.1.1 (2011-07)17
5.2.4 SHA-256
SHA-256 MAY be used to hash a message, M, having a length of up to 264-1bits.
The final result of SHA-256 is a 256-bit message digest. The SHA-256 algorithm is described in FIPS
Publication 180-3 [4].
5.2.5 WHIRLPOOL
WHIRLPOOL is a hash function defined in ISO/IEC 10118-3 [3] that operates on messages less than 2256-1 bits in
length, and produces a message digest of 512 bits.
Whirlpool MAY be used to compute the imprint of a message placed in a time-stamp token.
Whirlpool MAY be used with the RSA algorithm. The WHIRLPOOL algorithm is described in ISO/IEC 10118-3 [3].
NOTE: The Whirlpool output, i.e. 512 bits, is more than what may be needed for DSA and ECDSA, but there is a
general principle to use the leftmost bits of the output as the hash value, but currently there is no
Whirlpool algorithm variant defined by an OID/URN with a truncated length. Whirlpool has been
included as an alternative to the SHA-2 family and can be used either to compute a hash value (for a
time-stamp token) or with the RSA algorithm.
5.2.6 SHA-384
SHA-384 MAY be used to hash a message, M, having a length of up to 2128-1 bits and the output size is 384 bits. The
function is defined in the exact same manner as SHA-512 (clause 5.2.7), except for the initial value and the truncation
of the final hash value. The SHA-384 algorithm is described in FIPS Publication 180-3 [4].
The specification for SHA-384 is identical to SHA-512, except that different initial values are used, and the final hash
value is truncated to 384 bits. Therefore it is NOT RECOMMENDED to use SHA-384, if SHA-512 can be used instead
without truncation.
5.2.7 SHA-512
SHA-512 MAY be used to hash a message, M, having a length of up to 2128-1bits.
The final result of SHA-512 is a 512-bit message digest. The SHA-512 algorithm is described in FIPS Publication
180-3 [4].
5.2.8 SHA-3
NIST announced a public competition on November 2, 2007 to develop a new cryptographic hash algorithm. Based on
the public discussion of the second-round candidates, NIST selected on December 9, 2010 five SHA-3 finalists BLAKE,
Grøstl, JH, Keccak, and Skein to advance to the third (and final) round of the competition. The final selection
will be announced in 2012.
6 Signature schemes
A signature scheme consists of three algorithms: a key generation algorithm and a signature creation algorithm and a
signature verification algorithm. The latter are identified hereafter as a pair of algorithms. Each pair has its own name.
6.1 Signature algorithms
6.1.1 General
The list of currently recommended signature algorithms is given in table 2. Each signature algorithm has a unique entry
index represented by a string beginning with "2." followed by a two-digit entry number.
ETSI
ETSI TS 102 176-1 V2.1.1 (2011-07)18
Table 2: The list of recommended signature algorithms
Signature algorithm
entry index
Short signature algorithm
entry name
Key and Parameter
generation algorithms
Normative references
2.01 Rsa rsagen1 RFC 3447 [14]
2.02 Dsa dsagen1 FIPS Publication 186-3 [6],
ISO/IEC 14888-3:2006 [8]
2.03 ecdsa-Fp ecgen1 ANSI X9.62 [7]
2.04 ecdsa-F2m ecgen2 ANSI X9.62 [7]
2.05 ecgdsa-Fp ecgen1 ISO/IEC 14888-3 [8]
2.06 ecgdsa-F2m ecgen2 ISO/IEC 14888-3 [8]
The following clauses describe the parameters and key generation algorithms for the signature algorithms listed in
table 2.
6.1.2 Recommended signature algorithms
6.1.2.1 RSA
The RSA algorithm's security is based on the difficulty of factoring large integers. The RSA computations SHALL be
performed as described in RFC 3447 [14]. To generate the key pair two prime numbers, p and q, are generated
randomly and independently, satisfying the following requirements:
• the bit length of the modulus n = p q must be at least MinModLen; its length is also referred to as ModLen;
• p and q should have roughly the same length, e.g. set a range such as 0,1 < | log2p - log2q | < 30;
• the set of primes from which p and q are (randomly and independently) selected SHALL be sufficiently large
and reasonably uniformly distributed.
The private key consists of a positive integer d (the private exponent) and the modulus n.
The public key consists of a positive integer e (the public exponent) and the modulus n.
CRT (Chinese Remainder Theorem) implementations are also allowed, in which case the private key will contain more
values derived from the factorization of the modulus n.
For RSA signatures also a padding method has to be specified.
6.1.2.2 DSA
The DSA algorithm's security is based on the difficulty of computing the discrete logarithm in the multiplicative group
of a prime field Fp. The DSA computations SHALL be performed as described in FIPS Publication 186-3 [6] with the
change notice. The same algorithm is also specified in ISO/IEC 14888-3 [8] which can be used for information and for
selection of larger parameters. The public parameters p, q and g MAY be common to a group of users. The bit length α
of the prime modulus p SHALL be at least pMinLen bits long. The bit length β of q, which is a prime divisor of (p-1),
SHALL be at least qMinLen bits long. g SHALL be computed as indicated in FIPS Publication 186-3 [6] with the
change notice.
According to ISO/IEC 14888-3 [8] only the following choices for α and β are specified:
α = 1024, β = 160;
α = 2048, β = 224;
α = 2048, β = 256;
α = 3072, β = 256.
The value of β determines the defined in FIPS PUBLICATION 180-3 [4] hash function to be used. This requires for
β = 160 the function SHA-1, which should not be used for new applications as recommended in clause 9.2.
ETSI
ETSI TS 102 176-1 V2.1.1 (2011-07)19
NOTE: SHA-224 does not provide security advantages over SHA-256. If it is not required by a signature length
restriction, since a signature with β =224 occupies 448 bits whereas a signature with β =256 needs
512 bits, it is RECOMMENDED to use parameters with β =256.
The private key consists of:
• the public parameters p, q and g;
• a statistically unique and unpredictable integer x, 0 < x < q, which is signatory-specific; and
• a statistically unique and unpredictable integer k, 0 < k < q, which must be regenerated for each signature.
If the distribution of k is significantly different from uniform within the interval then there may be weaknesses.
Bleichenbacher has presented an attack which can be sub-exhaustive depending on the size of the bias and the number
of signatures produced using a single secret key. The value of k must be kept secret as well as the private key, even if k
is only partially known there exists an attack (Nguyen/Shparlinski).
The public key consists of p, q, g and an integer y computed as y = gx mod p.
When computing a signature of a message M, no padding of the hashcode is necessary. However, the hashcode must be
converted to an integer by applying the method described in appendix 2.2 of FIPS Publication 186-3 [6] with the change
notice.
6.1.2.3 Elliptic curve analogue of DSA based on a group E(Fp)
This signature algorithm is referred to as ecdsa-Fp. The algorithm SHOULD be applied as specified in ANSI X9.62 [7].
The same algorithm is also specified in ISO/IEC 14888-3 [8], IEEE P1363 [5] and ISO/IEC 15946-2 (withdrawn) [i.21]
which can be used for information. According to ISO/IEC 14888-3 [8] the elliptic curve analogue of DSA can be used
with non-SHA hash functions. An OID for the combination with RIPEMD-160 is given in clause 11.1.3.
The security of the ecdsa-Fp algorithm is based on the difficulty of computing the elliptic curve discrete logarithm.
The public parameters are as follows:
• p prime;
• q large prime at least qMinLen bits long, p ≠ q;
• E elliptic curve over a finite field Fp whose order n is divisible by q; and
• P point on E(Fp) of order q.
The public parameters MAY be common to a group of users. The quotient h of the group order n divided by q may be
considered as a public parameter too.
The class number of the maximal order of the endomorphism ring of E SHALL be at least MinClass = 200.
The value r0: = min (r: q divides pr-1) SHALL be greater than r0Min=104.
h = n/q SHALL be less or equal 4. All standardized elliptic curves fulfil this requirement (note that
http://www.secg.org/download/aid-780/sec1-v2.pdf [7] requires only h ≤ 2t/8, where 2t is the bit length of p).
In FIPS Publication 186-3 [6] five curves over a prime field are defined. The ECC Brainpool paper on standard curves
and curve generation (RFC 5639 [i.13]) contains an alternative set of curves over prime fields with 160 bits, 192 bits,
224 bits, 256 bits, 320 bits, 384 bits and 512 bits. All these curves fulfil the above requirements.
The private key consists of:
• the public parameters E, m, q and P;
• a statistically unique and unpredictable integer x, 0 < x < q, which is signatory-specific; and
• a statistically unique and unpredictable integer k, 0 < k < q, which must be regenerated for each signature.
The public key consists of E, q, P and Q, a point of E, which is computed as Q = xP.
ETSI
ETSI TS 102 176-1 V2.1.1 (2011-07)20
6.1.2.4 Elliptic curve analogue of DSA based on a group E(F2m)
This signature algorithm is referred to as ecdsa-F2m. The algorithm SHALL be applied as specified in ANSI X9.62 [7].
The same algorithm is also specified in ISO/IEC 14888-3 [8], IEEE P1363 [5], and ISO/IEC 15946-2 [i.21] which can
be used for information. The security of the ecdsa-F2m algorithm is based on the difficulty of computing the elliptic
curve discrete logarithm.
The public parameters are as follows:
• m prime number;
• q large prime at least qMinLen bits long;
• E elliptic curve over a finite field F2m whose order n is divisible by q;
• it must not be possible to define E over F2; and
• P point on E(F2
m) of order q.
h = n/q must be less or equal 4 (note that [7] requires only h ≤ 2t/8, where 2t is the bit length of p).
The class number of the maximal order of the endomorphism ring of E SHALL be at least MinClass = 200. The value
r0:=min(r: q divides 2mr-1) SHALL be greater than r0Min = 104.
In FIPS Publication 186-3 [6] five pseudorandomly generated curves over F2m are defined. All these curves satisfy the
above requirements.
A field representation is required, common to both the signatory and the verifier, so that signatures can be interpreted
correctly. The representations given in IEEE P1363 [5] and FIPS Publication 186-3 [6] are recommended. Thus if a
polynomial basis is required then an irreducible trinomial of the form xm + xa + 1 with minimal a should be used. If
such a polynomial does not exist then an irreducible pentanomial of the form xm + xa + xb + xc + 1 should be used; a
should be minimal, b should be minimal given a and c should be minimal given a and b.
The private key consists of:
• the public parameters E, m, q and P;
• a statistically unique and unpredictable integer x, 0 < x < q, which is signatory-specific; and
• a statistically unique and unpredictable integer k, 0 < k < q, which must be regenerated for each signature.
The public key consists of E, q, P and Q, a point of E which is computed as Q=xP.
6.1.2.5 EC-GDSA based on a group E(Fp)
This signature algorithm is referred to as ecgdsa-Fp. The algorithm SHALL be applied as specified in
ISO/IEC 14888-3 [8]. The security of the ecgdsa-Fp algorithm is based on the difficulty of computing the elliptic curve
discrete logarithm.
The ecgdsa-Fp algorithm is a variant of the ecdsa-Fp algorithm with a modified signature creation equation and
verification method. The parameters are the same as for ecdsa-Fp and therefore should satisfy all the constraints given
in clause 6.1.2.3.
NOTE: The basic difference between ECDSA and EC-GDSA is that during signature creation k does not need to
be inverted for ECGDSA. Under certain circumstances this can be advantageous for the design and
performance of the SCDev.
6.1.2.6 EC-GDSA based on a group E(F2m)
This signature algorithm is referred to as ecgdsa-F2m. The algorithm SHALL be applied as specified in
ISO/IEC 14888-3 [8]. The security of the ecgdsa-F2m algorithm is based on the difficulty of computing the elliptic
curve discrete logarithm.
ETSI
ETSI TS 102 176-1 V2.1.1 (2011-07)21
The ecgdsa-F2m algorithm is a variant of the ecdsa-F2m algorithm with a modified signature creation equation and
verification method. The parameters are the same as for ecdsa-F2m and therefore should satisfy all the constraints given
in clause 6.1.2.4.
NOTE: For the difference between ECDSA and ECGDSA see the note in clause 6.1.2.5.
6.2 Recommended key pair generation methods
6.2.1 General
Key pair generation methods are not part of the definition of a signature suite and may evolve without the need to
change the identifier of the signature suite.
Table 3 summarizes the recommended key pair generation methods for all signature algorithms considered in the
present document. Each key pair generation method has a unique entry index represented by a string beginning with
"3." followed by a two-digit entry number.
Table 3: The list of recommended key pair generation methods
Key
generator
entry index
Short key
generator
entry name
Signature
algorithm
Random
number
generation
method
Random generator
parameters
Adoption
date
Normative references
3.01 Rsagen1 rsa trueran or
pseuran
EntropyBits or
SeedEntropy, resp.
01.01.2001
3.02 Dsagen1 dsa trueran or
pseuran
EntropyBits or
SeedEntropy, resp.
01.01.2001 FIPS Publication 186-3 [6]
3.03 ecgen1 ecdsa-Fp,
ecgdsa-Fp
trueran or
pseuran
EntropyBits or
SeedEntropy, resp.
01.01.2001
3.04 ecgen2 ecdsa-F2m,
ecgdsa-F2m
trueran or
pseuran
EntropyBits or
SeedEntropy, resp.
01.01.2001
6.2.2 Recommended key pair generation methods
6.2.2.1 Key and parameter generation algorithm rsagen1
Generate p and q as indicated in clause 6.1.2.1 by applying a random number generation method satisfying the
requirements trueran (see clause 8.2.1) or using a method satisfying pseuran (see clause 8.2.2) with an appropriate size
seed. Each prime SHALL effectively be influenced by EntropyBits bits of true randomness or a seed of entropy
SeedEntropy bit. Random numbers SHALL be tested for primality until one of them is found to be prime with a
probability of error (i.e. of actually being composite) of at most ErrProb. Details on generating random primes can be
found in ISO/IEC 18032 [i.8], in particular clause 8.2. Examples of algorithms to produce RSA moduli, i.e. pairs of
primes satisfying the condition 0,1 < | log2p - log2q | < 30 are given in annex C.
NOTE 1: Annex A of ISO/IEC 18032 [i.8] contains a table of error probabilities for different probabilistic primality
tests.
EXAMPLE: For a random number of 1 024 bit tested with three successful iterations of the Miller-Rabin test
the probability that this number is not a prime is about 2-93.
The private exponent d and the public exponent e must satisfy ed ≡ 1 (mod lcm (p-1, q-1)) which is automatically the
case if ed ≡ 1 (mod (p-1)(q-1)). The private exponent d must not be too small (Wiener 1990 [i.9], Boneh and Durfee
1999 [i.9], Durfee and Nguyen 1999 [i.11]); it is sufficient to choose d in a range at least n from its minimum and
maximum values.
In practice by randomly choosing the public exponent e (subject to the condition gcd(e,(p-1)(q-1))=1) the corresponding
private exponent d will satisfy that condition with very high probability. If e is chosen small (e.g. less than n0,125) the
condition on d will automatically be satisfied.
ETSI
ETSI TS 102 176-1 V2.1.1 (2011-07)22
NOTE 2: It may also be recommendable to choose e not too small (e ≥ 216+1) as for example results of Boneh and
Venkatesan suggest that for very small e the RSA problem could be easier than factoring. Nevertheless in
contrast to RSA encryption small public exponents generally are not a direct threat for RSA signatures.
A small public exponent (e.g. e = 3) MAY be used if performance is critical, otherwise e ≥ 216+1 is RECOMMENDED.
A new modulus has to be produced for each user of the signature scheme even if different public exponents are used. In
practice if the moduli and public exponents are produced as described above (i.e. random modulus and choosing the
public exponent) the probability of producing the same modulus or secret exponent is negligible.
NOTE 3: It is not recommended to use a prime selection algorithm which prefers a special class of primes.
For example according to Rivest and Silverman [i.12] the use of "strong" primes would not improve
security in practice.
6.2.2.2 Key and parameter generation algorithm dsagen1
The primes p and q SHALL be generated as described in appendix 2.2 of FIPS Publication 186-3 [6] with primality of
an integer regarded as satisfied if the probability that it is composite is at most ErrProb.
Generate x by applying a random number generation method satisfying the requirements trueran (see clause 8.2.1) or
using a method satisfying pseuran (see clause 8.2.2) with an appropriate size seed. Each value of x SHALL effectively
be influenced by EntropyBits bits of true randomness or a seed of entropy SeedEntropy bit. Generate k using one of
these methods; k does not have to be generated using exactly the same method as x. Possible methods for this can be
found in FIPS Publication 186-3 [6] which contains a Change Notice (due to Bleichenbacher's attack).
6.2.2.3 Key and parameter generation algorithm ecgen1 for ecdsa-Fp
The prime numbers p and q, and the point P on E(Fp) SHALL be selected so that the conditions in clause 6.1.2.3 are
satisfied with primality of an integer regarded as satisfied if the probability that it is composite is at most ErrProb.
Possible methods to generate p, q, E and P are specified in [7], in the ECC Brainpool paper on standard curves and
curve generation [i.13] and also in clause D.1.
In situations where an intentional choice of weak public parameters (subject to an unknown "insider" attack) seems to
be possible a countermeasure is to request that these parameters are generated verifiably at random. In such situations it
is recommended to do so at least for the generation of the curve E. In [7], in the ECC-Brainpool paper on standard
curves and curve generation and in clause D.1 possible methods for this are described.
Generate x by applying a random number generation method satisfying the requirements trueran (see clause 8.2.1) or
using a method satisfying pseuran (see clause 8.2.2) with an appropriate size seed. Generate k using one of these
methods; k does not have to be generated using exactly the same method as x. Each value of x and k SHALL effectively
be influenced by EntropyBits bits of true randomness or a seed of entropy SeedEntropy bit.
6.2.2.4 Key and parameter generation algorithm ecgen2 for ecdsa-F2m
The prime numbers m and q, the elliptic curve E over F2
m and the point P on E(F2m) SHALL be selected so that the
conditions in clause 6.1.2.4 are satisfied with primality of an integer regarded as satisfied if the probability that it is
composite is at most Err Prob. Possible methods to generate m, q, E and P are specified in [7] (and also in clause D.2).
In situations where an intentional choice of weak public parameters (subject to an unknown "insider" attack) seems to
be possible a countermeasure is to demand that these parameters are generated verifiably at random. In such situations it
is recommended to do so at least for the generation of the curve E. In [7] a possible method for this is described
(compare also clause D.2).
Generate x by applying a random number generation method satisfying the requirements trueran (see clause 8.2.1) or
using a method satisfying pseuran (see clause 8.2.2) with an appropriate size seed. Generate k using one of these
methods; k does not have to be generated using exactly the same method as x. Each value of x and k SHALL effectively
be influenced by EntropyBits bits of true randomness or a seed of entropy SeedEntropy bit.
6.2.2.5 Key and parameter generation algorithm ecgen1 for ecgdsa-Fp
The parameter and key generation methods should be the same as the ecdsa-F2m methods described in clause 6.2.2.3.
ETSI
ETSI TS 102 176-1 V2.1.1 (2011-07)23
6.2.2.6 Key and parameter generation algorithm ecgen2 for ecgdsa-F2m
The parameter and key generation methods should be the same as the ecdsa-F2m methods described in clause 6.2.2.4.
7 Signature suites
7.1 General
To meet this security requirement and to allow signing of more or less arbitrary long messages, a signature suite
requires a hash function, so that the signing/verification algorithms operate on a fixed-size hash of the message. An
important issue is to tie the hash function to the signature scheme. Without this, the weakest available hash function
could define the overall security level.
Due to possible interactions which may influence security of electronic signatures, algorithms and parameters for secure
electronic signatures SHALL be used only in predefined combinations referred to as the signature suites. A signature
suite consists of the following components:
• a hash function;
• a padding method;
• a signature algorithm and its associated parameters.
If any of the components of a suite is modified, then the suite must be modified accordingly.
The list of recommended hash functions is defined in clause 5.1.
The list of recommended padding methods is defined in clause 7.2.
The list of recommended signature algorithms is defined in clause 6.1.
The list of currently recommended signature suites is given in clause 7.3.
Key generation is not part of the way to identify a signature suite and may change over time. Key generation methods
are addressed in clause 6.2.
Some key generation methods and some signature suites require to generate a (pseudo-) random number. The (pseudo)
random number generation method is not part of the way to identify a signature suite and may change over time.
(Pseudo) random number methods are addressed in clause 8.
7.2 Padding methods
Padding is algorithm dependent and some algorithms need non-trivial padding. This is the case for the RSA algorithm.
Signature algorithms with appendix require methods that encode a message into an integer message representative that
will be the input for the signature primitive. This encoding method can be deterministic, for example a padding of a
fixed string to the hash value computed from the message, but may be also randomized, incorporating a (randomly
generated) salt value, which are converted to and from message representatives. Although these latter encodings are not
true padding schemes, they are listed here.
The list of currently recommended padding methods is given in table 4. Each padding method has a unique entry index
represented by a string beginning with "4." followed by a two-digit entry number.
Table 4: The list of recommended padding methods
Padding
method entry
index
Short padding function
entry name
Random number
generation method
Random generator
parameters
Normative references
4.01 emsa-pkcs1-v1.5 - - RFC 3447 [14]
4.02 emsa-pkcs1-v2.1 - - RFC 3447 [14], section 9.2
4.03 emsa-pss trueran/pseuran MinSaltEntropy RFC 3447 [14], section 9.1
ETSI
ETSI TS 102 176-1 V2.1.1 (2011-07)24
Padding
method entry
index
Short padding function
entry name
Random number
generation method
Random generator
parameters
Normative references
4.04 iso9796ds2 trueran/pseuran MinSaltEntropy ISO/IEC 9796-2 [17]
4.05 iso9796-din-rn trueran/pseuran MinSaltEntropy DIN 66291-1 [41]
4.06 iso9796ds3 - - ISO/IEC 9796-2 [17]
Each salt value SHALL effectively be influenced by at least MinSaltEntropy bits of true randomness or a seed of
entropy at least MinSaltEntropy bits. This rule implies that the salt length is at least MinSaltEntropy bits.
NOTE 1: The above rule of MinSaltEntropy bits salt entropy is not meant in the strict and exclusive manner as the
demand for EntropyBits/SeedEntropy bits of entropy for key generation in clause 6.2, it is a
recommendation. For example Coron [i.24] showed that for emsa-pss already a significantly shorter than
64 bit salt length allows a reduction of the security of the signature scheme to the RSA problem under
realistic assumptions. Nevertheless such a reduction analysis does not take into account every kind of
possible weaknesses e.g. side channels. The default value for the salt length is 20 Bytes. The bit length of
the hash function used in the signature scheme is another appropriate value. Shorter values are not
recommended.
The emsa-pkcs1-v1.5 padding method is included, but it is NOT RECOMMENDED for new implementations, since it
will be phased out.
NOTE 2: Up to May 2011, no real attack on emsa-pkcs1-v1.5 has been published. Some attacks on this padding
scheme were caused by implementation errors or side channels. It should these attacks are impossible in
case the emsa-pss scheme was used.
The emsa-pss method is included as, despite not being widely used, it has been stable for a long time and is a good
improvement to the two emsa-pkcs1 schemes (i.e. -v1.5 and -v2.1 which only differ by the encoding method) and it is
better suited for long term use. The padding method emsa-pss is parameterized by the choice of hash function and a
mask generation function MGF, defined in PKCS#1 (RFC 3447 [14]). In the present document, MGF is based
ALWAYS on the corresponding hash function used, e.g. SHA-1 or SHA-256.
The method iso9796ds2 is "digital signature scheme 2" in ISO/IEC 9796-2 [17].
The method iso9796-din-rn is the variant of a scheme from ISO/IEC 9796-2 [17] called "DSI according to
ISO/IEC 9796-2 [17] with random numbers" in DIN 66291-1 [41]. It is described in annex A of [16].
NOTE 3: This is a variant on Digital Signature Scheme 1 of ISO/IEC 9796-2 [17]. The Digital Signature Scheme 1
has wide deployments and is secure but maybe in the future not be recommended for new systems.
The method iso9796ds3 is "digital signature scheme 3" in ISO/IEC 9796-2 [17].
NOTE 4: iso9796ds1 no longer represents state-of-the-art and in a paper presented by Coron, Naccache, Stern at
Crypto 99 it is shown that the effort to break this padding scheme is about 261 instead of 280.
7.3 Recommended signature suites
A signature suite is defined using three parameters:
• a hash function;
• a padding method;
• a signature algorithm and its associated parameters.
ETSI
ETSI TS 102 176-1 V2.1.1 (2011-07)25
Table 4.a
Entry name of the
signature suite
Entry name for the
hash function
Entry name for the
padding method
Entry name for the
signature algorithm
sha224-with-rsa sha224 (see note) rsa
sha256-with-rsa sha256 (see note) rsa
rsa-pss with mgf1SHA-
1Identifier
mgf1SHA-1 rsa
rsa-pss with
mgf1SHA-224Identifier
mgf1SHA-224 Rsa
rsa-pss with mgf1SHA-
256Identifier
mgf1SHA-256 Rsa
sha224-with-ecdsa sha224 no padding required ecdsa-Fp or ecdsa-F2m
sha256-with-ecdsa sha256 no padding required ecdsa-Fp or ecdsa-F2m
sha384-with-ecdsa sha384 no padding required ecdsa-Fp or ecdsa-F2m
sha512-with-ecdsa sha512 no padding required ecdsa-Fp or ecdsa-F2m
NOTE: The padding scheme for the RSA signature algorithm SHOULD be selected from
the list above.
These are the signature suites based on the currently recommended algorithms.
While SHA-1 is not broken yet, it is NOT RECOMMENDED for signature creation.
RIPEMD-160 is not broken and no attacks are known. Nevertheless it is no more RECOMMENDED due to hash length
of 160 bit.
SHA-512 is RECOMMENDED for very long term signatures, otherwise SHA-256 SHOULD be used.
8 Random number generation methods
8.1 General
The key generation methods and some signature suites require the generation of a random number.
NOTE: For detailed information about random number generation and terminology see ISO/IEC 18031 [i.14].
Some basic information is also given in annex E.
The random number generation methods combined with the key generation methods have to ensure that the expected
effort of guessing a cryptographic key is at least equivalent to guessing a random value that is EntropyBits bit resp.
SeedEntropy bit long. This can be satisfied with respect to different demands like information theoretic vs. just
complexity theoretic security, backward secrecy and/or forward secrecy and so on. Clause 8.2 and annex E in particular
specify by which RNGs these demands can be satisfied.
8.2 Recommended random number generation methods
Table 5 lists the recommended random number generation methods. Each random number generation method has a
unique entry index represented by a string beginning with "5." followed by a two-digit entry number. The terms
"trueran" and "pseuran" denote the requirements for NRNGs and DRNGs respectively (i.e. non-deterministic and
deterministic random number generators).
ETSI
ETSI TS 102 176-1 V2.1.1 (2011-07)26
Table 5: The list of recommended random number generation methods
Random generator
entry index
Short random generator
entry name
Random generator
parameters
Adoption date Normative
references
5.01 trueran EntropyBits 01.01.2001
5.02 pseuran SeedEntropy 01.01.2001
It is strongly recommended to use trueran methods for generating keys that are used more than once. In the case of the
one-time keys k for DSA, ECDSA and ECGDSA there is less urgency for that.
8.2.1 Random generator requirements trueran
A random number generator satisfying trueran has to be a pure or hybrid physical NRNG.
NOTE 1: Non-physical NRNGs are excluded as the designer has no real control of the amount of the produced
entropy.
Thus a random number generator satisfying trueran is based on a physical primary entropy source and possibly a
cryptographic or mathematical post-treatment of the output of the primary entropy source.
The recommended requirements for these components are:
• (TR1): There is a stochastical model for the primary entropy source which is found consistent with thorough
adapted tests of prototypes of the source.
• (TR2): The primary entropy source is subjected to an adapted statistical online test. "Online" means that the
test will detect any non-tolerable loss of quality of the primary entropy source during operation sufficiently
soon after such an event occurs and that there will then at once be suitable countermeasures (e.g. stop of the
generator). "Adapted" means adapted to the statistical model of the primary entropy source. The original
output of the primary entropy source should be tested not the output of the post-treatment instead of that (there
may be justified exceptions to this general rule).
See clause E.2 for some more information about tests for the primary entropy sources.
The stochastical model and the tests should deliver an estimate for the amount of the produced entropy. The primary
entropy source is regarded to be good if it produces nearly one bit entropy per output bit. For a good primary entropy
source no post-treatment is necessary.
• (TR3): If the primary entropy source is not good a post-treatment is employed which by some (necessarily
compressing) techniques delivers an output of nearly one bit entropy per output bit. There must be a
reasonable stochastical model of the post-treatment as well which together with the stochastic model of the
primary entropy source and the tests ensures this property of the output.
Instead of this set of requirements (TR1) - (TR3) the following modified set of requirements is also sufficient although
not recommended:
• (TR1"): There are mathematical models for the primary entropy source and the post-treatment that are
plausible.
• (TR2"): The primary entropy source is subjected to an online test which will detect most defects of the noise
source except for special unlikely events.
• (TR3"): There is a post-treatment (obligatory in this case) that under the assumption of the models (assuming
that the primary entropy source works as expected) delivers an output of nearly one bit entropy per output bit
and that even in the case of a complete breakdown of the primary entropy source (after there has been
accumulated enough entropy at the beginning) satisfies the requirements pseuran including condition (PR3) of
clause 8.2.2.
ETSI
ETSI TS 102 176-1 V2.1.1 (2011-07)27
NOTE 2: This alternative set of requirements is closer to the spirit of ANSI X9.82 [i.15] while the first set is more
similar to AIS 31. In both cases the major target is to achieve forward and backward secrecy. In the latter
case this secrecy can be completely complexity theoretic under certain circumstances and security relies
rather on the post-treatment than on the primary entropy source in contrast to the first case which delivers
information theoretical forward and backward secrecy. With the second set of requirements in the
situation of a readout or manipulation of the internal state also forward secrecy is not ensured.
NOTE 3: An example of a possible random number generator design based on a noisy diode is given in clause E.2
of ISO/IEC 18031 [i.14] although without the necessary details.
8.2.2 Random generator requirements pseuran
A random number generator satisfying pseuran is a pure or hybrid DRNG satisfying the following conditions:
• (PR1): The DRNG must be initialized by a seed with an entropy of at least SeedEntropy bits.
• (PR2): Even with the knowledge of a partial output bit sequence of the DRNG and having all information
about its initialization (and in the case of a hybrid DRNG also about the output of the additional entropy
source) except for the seed there is no usable method to determine any other m bits of the output with a
probability significantly larger than Max (2-m,2-SeedEntropy).
NOTE 1: The second condition in particular implies that there is no information ascertainable a priori as to the
output bits and that neither the seed nor any internal state of the DRNG can be recovered from a subset of
the output.
(PR1) is meant in the sense (or even implies) that the seed is produced using a NRNG. This NRNG does not need to be
a physical one. Nevertheless to achieve high security it is recommended to use trueran (in particular physical, see
clause 8.2.1) NRNGs for seeding. (PR1) does not exclude constructions in which the DRNG is seeded by a chain of
DRNGs as described in clause 9.3.2 of ISO/IEC 18031 [i.14]. However the first DRNG in this chain must be seeded
with the output of a NRNG and in the output of the last DRNG in the chain enough entropy (i.e. at least EntropyBits
bits) has to be left over. Moreover of course the whole system (chain + DRNG to be seeded) regarded as a DRNG
(including operational freedom like numbers of cycles before the next seeding of links regarded as non-physical
additional entropy source) has to satisfy the second condition. The security of a DRNG is only complexity theoretic.
With a known seed or a known internal state any future output can be calculated. So the seed has to be kept secret and
seeding SHALL follow procedures similar to those for the generation of root keys. No backups of the seed or internal
states of a pseuran generator are permitted. The internal state of the DRNG must be secured against any readout and any
adversarial manipulation.
In situations in which such readout or manipulation of an internal state of the DRNG does not seem to be completely
excluded a re-seeding or a seed-update has to be executed from time to time. If re-seeding is employed the security of
the re-seeding process SHALL be as strong as that of the original seeding. The frequency of this procedure (i.e. the
amount of entropy that is fed in per output bit) depends on the actual risk of such readouts or manipulations.
It is recommended to use DRNGs which in addition to the two above mentioned conditions satisfy the following
additional condition ensuring backward secrecy even in the case of a known internal state:
• (PR3): Even with complete knowledge of an internal state there is no usable method to determine any
previous m output bits with a probability significantly larger than Max(2-m,2-SeedEntropy).
NOTE 2: AIS 20 [i.16] defines the classes K3 and K4 for DRNGs. Roughly said K3 DRNGs satisfy conditions
(PR1) and (PR2), K4 DRNGs also satisfy (PR3).
Depending on the environment it may further be recommendable to use hybrid DRNGs rather than pure ones. In the
case of an hybrid DRNG according to (PR2) even with complete knowledge about the output of the additional entropy
source or with a certain influence on this output it must not be feasible to determine any bits of the output with higher
than the a priori probability.
ETSI
ETSI TS 102 176-1 V2.1.1 (2011-07)28
The following are examples of pseuran generators:
• ANSI X9.17 [i.17] generator. This DRNG was designed to pseudo randomly generate keys and initialization
vectors for use of DES. It uses the triple-DES algorithm with a fixed key to mix a 64-bit seed with the current
date. Iterated encryption enables to generate as many output bits as needed. Condition (PR3) is not satisfied at
least without any further assumptions about the clock input. Instead of triple-DES also other strong block
ciphers could be used as building block of the generator.
• Example E.4 in AIS 20 is another DNRG based on a variable strong block cipher which as well does not
satisfy condition (PR3).
• FIPS 186 generator (see FIPS Publication 186-3 [6]).
• RSA DRNG and Blum-Blum-Shub DRNG [i.18]. Those DRNGs are based on iterated exponentiation modulo
a composite modulus. The advantage is to base the security on the intractability of number theoretic problem
(respectively RSA and the factorization problem) but the main drawback is the poor efficiency in comparison
with the other DRNGs described above, the security of which is only heuristic.
9 Recommended hash functions and key sizes versus
time
In this clause recommendations are provided regarding the use of hash functions given in clause 5 and the key sizes to
be used with the algorithms mentioned in clause 6.
This clause is structured as follows:
• Clause 9.1 explains the considerations on which the recommendations are based.
• In clause 9.2, hash functions versus time are recommended.
• In clause 9.3, key sizes versus time are recommended.
9.1 Basis for the recommendations
The aim of the recommendations given in clauses 9.2 and 9.3 is to achieve the necessary security level for advanced and
qualified signatures according to Directive 1999/93/EC [1] and to complement the "generally recognized standards"
within the meaning of Article 3(5), in particular the CWA 14167-2 [38] and CWA 14169 [34]. These standards are
intended as protection profiles for the evaluation of electronic signature products, in particular trustworthy systems
operated by certification service providers and secure signature creation devices. This implies that this update of this
multi-part deliverable would primarily affect new electronic signature products, as long as no emergency procedure in
the case of new attacks requires the removal of a recommended algorithm (see clause 4).
The recommendations for algorithm and parameter strengths are characterized by taking a reasonable margin above
minimum key lengths based on both extrapolation of current trends as well as estimations based on the necessary
computing power needed to break a given algorithm. Such extrapolations can be found in the literature,
e.g. (LenstraVerheul) or the ECRYPT doc base (http://www.ecrypt.eu.org/documents.html). The recommendations are
not based on recently achieved records.
Unfortunately there are no rigorous security proofs for the components of signature schemes (hash function, signature
algorithm, RNG), basically all security statements rely on results about the most effective attacks known today. The
possibility of a complete break of such a component (like, e.g. a fast universal factorization algorithm against RSA) that
renders it useless can theoretically not completely be excluded but "breakthroughs" of that kind are regarded as
improbable. In contrast to that certain unforeseen advances of moderate degree in analyzing cryptographic algorithms
are regarded as a realistic threat: A recent example is given by the collision attacks on SHA-1 which demonstrated that
this hash function is actually much weaker against collision attacks than predicted. The security margin for the
recommendations below are chosen so that advances of this level should be compensated without changing the
parameters.
ETSI
ETSI TS 102 176-1 V2.1.1 (2011-07)29
This is justified by the very severe consequences in case that a whole generation of chip cards or, even worse, a root key
has to be regarded as no longer secure. Moreover, stability of the requirements in the present document is also highly
desirable for reasons of planning reliability. This means that if in e.g. 2007 a key length y is declared as sufficient for
resistance during 6 years, i.e. at least until the end of 2012, an updated version in e.g. 2010 should normally still declare
this key length y as sufficient at least until the end of 2012. The following tables contain recommendations for the
lifetime of keys and were chosen accordingly. The recommendations for a whole decade ("during 10 years") are
explicitly declared "speculative" because of the uncertainty of predictions over such long periods. That means that the
described principle of stability may not apply to these recommendations.
The concept of a "liberal view" appearing in the previous versions of the document is removed for the near time frames,
because significant differences did not appear therein.
An attempt was made to achieve roughly similar security for all the components. For example the security level
demanded by the tables is very roughly equivalent to 80 bits symmetric keys for an intended use in the short term and
100 bit for an intended use in the medium and longer term.
9.2 Recommended hash functions versus time
Table 7 provides indication about recommended hash functions during X years.
Definitions:
Usable: The algorithm with the given security parameters can be considered secure at the given time.
Unknown: The security of the algorithm is unknown; it may become secure if additional measures are applied.
Unusable: The algorithm cannot be considered secure for any kind of use in the context of electronic signatures.
"3 years" in table 7 means "until the end of 2014" and so on.
Table 6: recommended hash functions for a resistance during X years
entry name of the hash function 1 year 3 years 6 years 10 years
(speculative)
sha1 unusable unusable unusable unusable
ripemd160 unusable unusable unusable unusable
sha224 usable usable usable unknown
sha256 usable usable usable unknown
sha384 usable usable usable usable
sha512 usable usable usable usable
Whirlpool usable usable usable usable
NOTE: The listed hash functions are expected to be 2nd pre-image resistant and pre-image resistant for a
longer period of time.
NOTE: The listed hash functions are expected to be 2nd pre-image resistant and pre-image resistant for a longer
period of time.
9.3 Recommended key sizes versus time
In the sequel a recommendation for RSA of the form, MinModLen=y for a resistance during 3 years" means,
MinModLen should be y for RSA keys with an intended life time of 3 years from now (i.e. until end of 2004)". A
recommendation for EC(G)DSA of the form "qMinLen=y for a resistance during 6 years" means "qMinLen should be y
for keys (and curve parameters) with an intended life time of six years from now (i.e. until the end of 2017)". The
meaning of all the listed parameters is explained in the respective clauses in clause 6.
ETSI
ETSI TS 102 176-1 V2.1.1 (2011-07)30
Table 7: Recommended parameters for RSA and rsagen1 for a resistance during X years
Parameter 1 year 3 years 6 years 10 years
(speculative)
MinModLen 1 536 2 048 2 048 ?
ErrProb 2-80 2-100 2-100 2-100
SeedEntropy/EntropyBits 80 100 100 ?
NOTE 1: There exist implementations in hardware for which a few of the 2 048 base number bits may be reserved
for some other information such that the maximum modulus length is in fact 1 976 bit for these
implementations. Since the loss of security seems to be negligible a MinModLen=1 976 can be
considered as recommended at least until the end of 2012 too.
NOTE 2: Up to May 2011, no real attacks on RSA-1024 are reported. The factorization of RSA-768 dated on
December 2010 was done using several hundred PCs over two years of calendar time. Therefore a freshly
generated 1 024 bit RSA key certified for an end user and valid e.g. for one year may be considered today
as secure too. Nevertheless it may be necessary to apply additional cryptographic measures urgently
during the validity time of the end user certificates, including their revocation, if the algorithm RSA-1024
becomes really weak, e.g. if the next RSA challenges RSA-896 (cf. http://www.rsa.com/rsalabs) is
factored.
NOTE 3: In the gap left between the recommendations for one and three years it is on the discretion of the issuing
CA to use either the first or the second recommendation. Depending on the measures foreseen by the CA
in case of cryptographic weakness it may even be still appropriate to use RSA-1024 keys for end user
certificates for more than two years, if the CA is prepared to revoke all certificates with that key length
and to re-issue new certificates with a greater key length in a timely manner.
Table 8: Recommended padding schemes and values
for MinSaltEntropy for a resistance during X years
Entry name of the
padding scheme
1 year 3 years 6 years 10 years
(speculative)
PKCS#1-v1.5 usable/n.a usable/n.a. usable/n.a. unusable/n.a.
PKCS#1-v2.1 usable/n.a usable/n.a usable/n.a. unusable/n.a.
PKCS#1-PSS usable/64 usable/64 usable/64 usable/64
ISO-DS 2 usable/64 usable/64 usable/64 usable/64
ISO-DS 3 usable usable usable usable
ISO-DIN-RN usable/64 usable/64 usable/64 usable/64
Table 9: Recommended parameters for DSA and dsagen1 for a resistance during X years
Parameter 1 year 3 years 6 years 10 years
(speculative)
pMinLen 1 024 2 048 3 072 3 072
qMinLen 160 224 256 256
ErrProb 2-80 2-80 2-100 2-100
SeedEntropy/EntropyBits 80 80 100 100
ETSI
ETSI TS 102 176-1 V2.1.1 (2011-07)31
If it is not required by a signature length restriction it is RECOMMENDED for future interoperability to use
qMinLen=256, such that DSA can be used with SHA-256 and without truncation.
Table 10: Recommended parameters for ecdsa-Fp and ecgen1 for a resistance during X years
Parameter 1 year 3 years 6 years 10 years
(speculative)
pMinLen - - qMinLen
160 224 256 256
r0Min 104 104 104 ?
MinClass 200 200 200 ?
ErrProb 2-80 2-100 2-100 2-100
SeedEntropy/EntropyBits 80 100 100 256
Table 11: Recommended parameters for ecdsa -F2m and ecgen2 for a resistance during X years
Parameter 1 year 3 years 6 years 10 years
(speculative)
pMinLen - - qMinLen
160 224 256 256
r0Min 104 104 104 ?
MinClass 200 200 200 ?
ErrProb 2-80 2-100 2-100 2-100
SeedEntropy/EntropyBits 80 100 100 256
Table 12: Void
Table 13: Void
Table 14 summarizes the recommendations form the tables above.
Table 14: Recommended signature suites for a resistance during X years
Entry name of the signature suite 1 years 3 years 6 years 10 years
sha1-with-rsa not recommended
sha256-with-rsa 1 536 2 048 2 048 not recommended
RSASSA-PSS with mgf1SHA-1Identifier 1 536 not recommended
RSASSA-PSS with mgf1SHA-224Identifier 1 536 2 048 2 048 not recommended
RSASSA-PSS with mgf1SHA-256Identifier 1 536 2 048 2 048 3 072
sha1-with-dsa not recommended
sha1-with-ecdsa not recommended
sha224-with-ecdsa 224 224 224 not recommended
sha256-with-ecdsa 256 256 256 256
NOTE 4: Because sha224-with-rsa has no security or performance advantages or disadvantages compared with
sha256-with-rsa it is not listed here for interoperability reasons only.
NOTE 5: The generation of pseudo random masks used in PSS formatting requires the mixing properties of the
MGF (mask generation function), which is not affected by the collision attacks on SHA-1.
NOTE 6: In the gap left between the recommendations for one and three years it is on the discretion of the issuing
CA to use either the first or the second recommendation. Depending on the measures foreseen by the CA
and accepted by the certificate holder in case of algorithmic weaknesses it may be appropriate to use for
example RSA-1024 keys in a one year valid end user certificates.
ETSI
ETSI TS 102 176-1 V2.1.1 (2011-07)32
10 Time period resistance of hash functions and keys
The hash functions and signature algorithms defined in the present document are suitable to be used in the context of
advanced electronic signatures as defined by ETSI Technical Standards and CWAs.
As a general rule, a private key SHALL resist during the validity period of certificates, (defined by the "notBefore" and
"notAfter" elements of the validity period field) which contain the corresponding public key.
NOTE: The validity period is defined by the "notBefore" and "notAfter" elements of the validity period field from
the certificate.
Since key sizes are directly dependent upon the usage of the certificate, no single key size value may be given.
The time period during which a given key SHALL or SHOULD resist depends on the usage of the key. To this respect
different use cases will be explored. Once the time period is known, then the figures provided in clause 9 can be used to
know the appropriate key size.
10.1 Time period resistance for hash functions
As a general rule, hash functions SHOULD resist as long as a signature verification still needs to be done. If not, a
specific signature maintenance process SHALL be performed (see annex H for more information).
A hash function used to compute the hash of a certificate, which is not a self-signed certificate, SHOULD resist during
the validity period of that certificate. However, a hash function used to compute the hash of a self-signed certificate
SHALL resist during the validity period of that self-signed certificate.
A hash function used to compute the imprint of a message placed in a time-stamp token is not used in combination of a
signature scheme. The length of its output is not dependent upon the size of the parameters of the signature scheme. It
may be advisable, in order to reduce the signature maintenance process, to use a hash function that is presumed to be
resistant over a very long time period. If the signature suite that has been used by the signer is also presumed to be
resistant over a very long time period, then the signature maintenance process can be minimized.
10.2 Time period resistance for signer's key
The focus is very often placed on the resistance of signer's keys.
Signer's keys SHOULD resist during the validity period (from notBefore to notAfter) of the associated
certificate. If they do not, revocation will be necessary, and there would be a large burden to re-issue new keys and
certificates. However, there is no security breach.
If a signer's key does not resist during the validity period of its associated certificate, then the protection provided
through the use of time-stamping is sufficient to provide an adequate protection.
10.3 Time period resistance for trust anchors
A trust anchor SHOULD remain secure during the whole time period during which advanced electronic signature needs
to be verified. If it does not, it cannot be used anymore for immediate verifications. It can be used for subsequent
verifications, if a specific maintenance process is performed before the trust anchor becomes insecure.
This is an important difference to the estimation of the time period resistance for signer's key, therefore a stronger
margin SHOULD be chosen to avoid additional maintenance procedures.
10.4 Time period resistance for other keys
All other keys (TSU keys, CA keys, CRL issuer keys, OCSP responder keys) SHOULD resist during the validity period
of the associated certificate and the certificates that rely on its validity. This implies that their security parameters
SHOULD be chosen stronger than the corresponding parameters of the certified keys.
ETSI
ETSI TS 102 176-1 V2.1.1 (2011-07)33
If they do not resist the for seen time period, a maintenance process SHOULD be applied before the algorithm is
broken.
For these keys the same rule as for trust anchors applies.
11 Practical ways to identify hash functions and
signature algorithms
In order to be able to use a function or an algorithm with ETSI TS and CWA documents, it is mandatory to be able to
reference it, and when the algorithm has parameters to be able to define these parameters. An "object" needs to be
defined to support these parameters. That object MUST be referenced using an OID and/or a URN. Only the owner of
the OID or the URN is allowed to define its meaning and thus the meaning of the algorithm, usual referencing another
document. It may be observed that ISO standards are not referenced in RFCs documents. The primary reason is that
these documents are sold and the IETF always gives its preference to documents that can be obtained for free.
As a general rule the "OID/URN criterion" may be applied: An algorithm to be included must be defined
unambiguously by an OID/URN. If such an OID/URN is not available the algorithm unusable.
11.1 Hash functions and signature algorithms objects identified
using OIDs
11.1.1 Hash functions
The hash functions are defined using the following OIDs.
Table 15
Short object
name
OID Normative
references
id-sha1 { iso(1) identifiedOrganization(3) oIW(14) oIWSecSig(3) oIWSecAlgorithm(2) 26 } RFC 3279 [12]
ripemd160 { iso(1) identifiedOrganization(3) teletrust(36) algorithm(3) hashAlgorithm(2)
ripemd160(1) }
ISIS-MTT
Part 6 [26]
id-sha224 { joint-iso-itu-t(2)country(16) us(840) organization(1) gov(101) csor(3)
nistalgorithm(4) hashalgs(2) sha224(4) }
RFC 4055 [15]
id-sha256 { joint-iso-itu-t(2) country(16) us(840) organization(1) gov(101) csor(3)
nistalgorithm(4) hashalgs(2) 1 }
RFC 4055 [15]
whirlpool { iso(1) standard(0) encryption-algorithms(10118) part3(3) algorithm(0) whirlpool(55)
}
ISO/IEC 10118-3 [3]
id-sha384 { joint-iso-itu-t(2) country(16) us(840) organization(1) gov(101) csor(3)
nistalgorithm(4) hashalgs(2) 2 }
RFC 4055 [15]
id-sha512 { joint-iso-itu-t(2) country(16) us(840) organization(1) gov(101) csor(3)
nistalgorithm(4) hashalgs(2) 3 }
RFC 4055 [15]
11.1.2 Signature algorithms
Table 16
Short object name OID Normative
references
rsaEncryption { iso(1) member-body(2) us(840) rsadsi(113549) pkcs(1) pkcs-1(1) 1 } RFC 3279 [12]
id-dsa { iso(1) member-body(2) us(840) x9-57(10040) x9cm(4) 1 } RFC 3279 [12]
id-ecPublicKey { iso(1) member-body(2) us(840) 10045 2 1 } RFC 3278 [43]
ecgPublicKey { iso(1) identified-organization(3) teletrust(36) algorithm(3)
signatureAlgorithm(3) ecSign(2) ecgDsaStd(5) ecgKeyType(2) 1 }
ETSI
ETSI TS 102 176-1 V2.1.1 (2011-07)34
11.1.3 Signature suites
Table 17
Short object name OID Normative
references
sha1WithRSAEncryption { iso(1) member-body(2) us(840) rsadsi(113549) pkcs(1) pkcs-1(1) 5 } RFC 3279 [12]
sha256WithRSAEncryption { iso(1) member-body(2) us(840) rsadsi(113549) pkcs(1) pkcs-1(1) 11 } RFC 4055 [15]
sha512WithRSAEncryption { iso(1) member-body(2) us(840) rsadsi(113549) pkcs(1) pkcs-1(1) 13 } RFC 4055 [15]
id-RSASSA-PSS with
mgf1SHA-1Identifier
{ iso(1) member-body(2) us(840) rsadsi(113549) pkcs(1) pkcs-1(1) 10 } RFC 4055 [15]
id-RSASSA-PSS
with mgf1SHA-224Identifier
{ iso(1) member-body(2) us(840) rsadsi(113549) pkcs(1) pkcs-1(1) 10 } RFC 4055 [15]
id-RSASSA-PSS
with mgf1SHA-256Identifier
{ iso(1) member-body(2) us(840) rsadsi(113549) pkcs(1) pkcs-1(1) 10 } RFC 4055 [15]
rsaSignatureWithripemd160 {iso(1) identified-organization(3) teletrust(36) algorithm(3)
signatureAlgorithm(3) rsaSignature(1) rsaSignatureWithripemd160(2)}
ISIS-MTT [26]
id-dsa-with-sha1 { iso(1) member-body(2) us(840) x9-57 (10040) x9cm(4) 3 } RFC 3279 [12]
id-dsa-with-sha224 { joint-iso-ccitt(2) country(16) us(840) organization(1) gov(101) csor(3)
algorithms(4) id-dsa-with-sha2(3) 1 }
id-dsa-with-sha256 { joint-iso-ccitt(2) country(16) us(840) organization(1) gov(101) csor(3)
algorithms(4) id-dsa-with-sha2(3) 2 }
ecdsa-with-SHA-1 { iso(1) member-body(2) us(840) ansi-X9-62(10045) signatures(4)
sha1(1) }
RFC 3279 [12]
ecdsa-with-Recommended { iso(1) member-body(2) us(840) ansi-X9-62(10045) signatures(4)
recommended(2) }
ANSI X9.62 [7]
ecdsa-with-Sha224 { iso(1) member-body(2) us(840) ansi-X9-62(10045) signatures(4)
specified(3) 1 }
ANSI X9.62 [7]
ecdsa-with-Sha256 { iso(1) member-body(2) us(840) ansi-X9-62(10045) signatures(4)
specified(3) 2 }
ANSI X9.62 [7]
ecdsa-with-Sha384 { iso(1) member-body(2) us(840) ansi-X9-62(10045) signatures(4)
specified(3) 3 }
ANSI X9.62 [7]
ecdsa-with-Sha512 { iso(1) member-body(2) us(840) ansi-X9-62(10045) signatures(4)
specified(3) 4 }
ANSI X9.62 [7]
ecdsa-with-RIPEMD160 {itu-t(0) identified-organization(4) etsi(0) reserved(127) etsi-identifiedorganization(0)
bsi-de(7) algorithms (1) id-ecc(1) signatures(4) ecdsasignatures(1)
6 }
ecgSignatureWithripemd160 { iso(1) identified-organization(3) teletrust(36) algorithm(3)
signatureAlgorithm(3) ecSign(2) ecgDsaStd(5) ecgSignature(4) 1 }
ecgSignatureWithsha1 { iso(1) identified-organization(3) teletrust(36) algorithm(3)
signatureAlgorithm(3) ecSign(2) ecgDsaStd(5) ecgSignature(4) 2 }
ecgSignatureWithsha224 { iso(1) identified-organization(3) teletrust(36) algorithm(3)
signatureAlgorithm(3) ecSign(2) ecgDsaStd(5) ecgSignature(4) 3 }
ecgSignatureWithsha256 { iso(1) identified-organization(3) teletrust(36) algorithm(3)
signatureAlgorithm(3) ecSign(2) ecgDsaStd(5) ecgSignature(4) 4 }
ecgSignatureWithsha384 { iso(1) identified-organization(3) teletrust(36) algorithm(3)
signatureAlgorithm(3) ecSign(2) ecgDsaStd(5) ecgSignature(4) 5 }
ecgSignatureWithsha512 { iso(1) identified-organization(3) teletrust(36) algorithm(3)
signatureAlgorithm(3) ecSign(2) ecgDsaStd(5) ecgSignature(4) 6 }
NOTE: Because sha224WithRSAEncryption and sha384WithRSAEncryption have no security or performance
advantages or disadvantages compared with sha256WithRSAEncryption and sha512WithRSAEncryption they
are not listed here for interoperability reasons only.
ETSI
ETSI TS 102 176-1 V2.1.1 (2011-07)35
11.2 Hash functions and signature algorithms identified objects
using URNs
11.2.1 Hash functions
The hash functions are defined using the following URNs.
Table 18
Short object
name
URN Normative references
sha1 http://www.w3c.org/2000/09/xmldsig#sha1 W3C Recommendation XML-Signature
Syntax and Processing [30].
ripemd160 http://www.w3.org/2001/04/xmlenc#ripemd160 W3C Recommendation XML Encryption
Syntax and Processing.
10 December 2002 [31]
sha224 http://www.w3.org/2001/04/xmldsig-more#sha224 RFC 4051 [29]
sha256 http://www.w3.org/2001/04/xmlenc#sha256 W3C Recommendation XML Encryption
Syntax and Processing.
10 December 2002 [31]
sha384 http://www.w3.org/2001/04/xmldsig-more#sha384 RFC 4051 [29]
sha512 http://www.w3.org/2001/04/xmlenc#sha512 W3C Recommendation XML Encryption
Syntax and Processing.
10 December 2002 [31]
11.2.2 Signature algorithms
There is no need to define such URNs since XAdES uses the signature algorithms contained in X.509 certificates which
are referenced using OIDs.
11.2.3 Signature suites
The signature suites are defined using the following URNs.
Table 19
Short object name URN Normative references
dsa-sha1 http://www.w3.org/2000/09/xmldsig#dsa-sha1 XML-Signature Syntax and Processing.
W3C Recommendation [30]
rsa-sha1 http://www.w3.org/2000/09/xmldsig#rsa-sha1 XML-Signature Syntax and Processing.
W3C Recommendation [30]
ecdsa-sha1 http://www.w3.org/2001/04/xmldsig-more#ecdsa-sha1 RFC 4051 [29]
rsa-ripemd160 http://www.w3.org/2001/04/xmldsig-more/rsa-ripemd160 RFC 4051 [29]
rsa-sha256 http://www.w3.org/2001/04/xmldsig-more#rsa-sha256 RFC 4051 [29]
rsa-sha384 http://www.w3.org/2001/04/xmldsig-more#rsa-sha384 RFC 4051 [29]
rsa-sha512 http://www.w3.org/2001/04/xmldsig-more#rsa-sha512 RFC 4051 [29]
ecdsa-sha224 http://www.w3.org/2001/04/xmldsig-more#ecdsa-sha224 RFC 4051 [29]
ecdsa-sha256 http://www.w3.org/2001/04/xmldsig-more#ecdsa-sha256 RFC 4051 [29]
ecdsa-sha384 http://www.w3.org/2001/04/xmldsig-more#ecdsa-sha384 RFC 4051 [29]
ecdsa-sha512 http://www.w3.org/2001/04/xmldsig-more#ecdsa-sha512 RFC 4051 [29]
ETSI
ETSI TS 102 176-1 V2.1.1 (2011-07)36
11.3 Recommended hash functions and signature algorithms
objects that do not yet have an OID or a description
Whirlpool
Signature suites based on a combination of a hash algorithm based on Whirlpool and a signature scheme do not yet have
an OID. Nevertheless this hash function can be used with any of the mentioned elliptic curves. It is a general principle
for ECDSA, that if the output of a hash function generated values longer than the order of the group of points, the hash
values are truncated to the leftmost bits. I.e. if the order q has n bits, then the n leftmost bits of the output of the hash
function are used as the hash value h. Note that this hash value may be greater than the order q but not longer.
ECDSA with other hash functions
The ecdsa signature scheme and signature suite based on ecdsa with RIPEMD160 have an OID. The normative
description is given in "Elliptic Curve Cryptography (ECC)", Bundesamt für Sicherheit in der Informationstechnik,
TR-03111(2006) available at http://www.bsi.bund.de.
ecdsa-with-RIPEMD160
"Elliptic Curve DSA with RIPEMD-160"
• {itu-t(0) identified-organization(4) etsi(0) reserved(127) etsi-identified-organization(0) bsi-de(7) algorithms
(1) id-ecc(1) signatures(4) ecdsa-signatures(1) 6 }
11.4 Recommended hash functions and signature algorithms
objects that do not yet have a URN or a description
Whirlpool
Whirlpool has currently no URN. In addition, it would be desirable to have variants of Whirlpool with an output less
than 512 bits in order to match the requirements of Elliptic Curves algorithms.
Signature suites based on a combination of a hash algorithm based on Whirlpool and a signature scheme do not yet have
a URN.
ECGDSA
The ecgdsa signature scheme and signature suites based on ecgdsa do not have a URN, and thus no normative reference
for the data structures.
ETSI
ETSI TS 102 176-1 V2.1.1 (2011-07)37
Annex A (normative):
Algorithms for various data structures
TS 101 733 [18] and TS 101 903 [19] define the formats of advanced electronic signatures. These two documents
reference other documents defining various standardized data structures.
These other documents or companion documents define the algorithms which SHOULD be supported by the issuers of
the data structures and the algorithms which SHALL (for interoperability purposes) and SHOULD be supported by the
users of the data structures.
• Signer Certificates (RFC 5280 [2] and RFC 3279 [12]).
• Certificate Revocation Lists (RFC 5280 [2] and RFC 3279 [12]).
• OCSP responses (RFC 2560 [22]).
• Certification Authority Certificates (RFC 5280 [2] and RFC 3279 [12]).
• Self-signed certificates for CA certificates (RFC 5280 [2] and RFC 3279 [12]).
• Time-Stamping Tokens (TSTs) (RFC 3161 [9] and TS 101 861 [20]).
• Time-Stamping Unit certificates (RFC 3161 [9] and TS 101 861 [20]).
• Self-signed certificates for TSU Certificates (RFC 5280 [2] and RFC 3279 [12]).
• Attribute Certificates (ACs) (RFC 5280 [2] and RFC 3279 [12]).
• Attribute Authority Certificates (RFC 3281 [21]).
For each data structure, the set of algorithms to be used is specified.
Since many of these documents have been published some years ago, they cannot be all up to date with the latest
cryptographic advancements. In particular, some of the algorithms specified in the above documents exhibit weaknesses
or, worse, are now broken.
For that reason, when it is the case, algorithms that were initially recommended and that shall or should not be used
anymore will be indicated.
In the same way, more recent algorithms do not appear in these documents. This does not mean that they should not be
used, but that at this time they do not yet fall into the SHALL or SHOULD categories.
Each set is identified by an identifier which is either an OID (Object IDentifier) or a URI/URN. The use of such
identifiers is necessary so that interoperability can be achieved. In order to allow for data interchange, the document
references algorithms in terms of OIDs and URIs/URNs together with algorithm parameters.
The algorithms which MAY be supported by issuers or users are NOT indicated.
A.1 Advanced Electronic Signatures based on
TS 101 733
An advanced electronic signature contains an identifier of the hash function that has been used (contained in the
digestAlgorithm element from the SignerInfo data structure) and an identifier of the signature algorithm that has
been used (contained in the signatureAlgorithm element from the SignerInfo data structure) which must be
consistent with the identifier of the signature algorithm contained in the signer's certificate.
Requirements apply both to the hash function and the signature algorithm.
ETSI
ETSI TS 102 176-1 V2.1.1 (2011-07)38
Since TS 101 733 [18] is built upon RFC 3852 [23], algorithms may be selected from RFC 3370 [13] or elsewhere.
RFC 3370 [13] recommended the MD5 and SHA-1 hash functions. Since MD5 was broken in August 2004, it is no
longer mentioned. SHA-1 should be phased out as its security level has significantly decreased since February 2005.
Table A.1
AdES based on TS 101 733 [18] Issuers of AdES Users of AdES
Hash functions SHOULD support sha256 SHALL support sha1
SHOULD support sha256
Signature algorithms SHOULD support RSA
or SHOULD support DSA
or SHOULD support ECDSA
SHALL support RSA
SHOULD support DSA
SHOULD support ECDSA
NOTE: Because the usage of SHA-224 as hash functions gives no advantage compared with SHA-256 neither in
security nor in performance it is not required that SHA-224 should be supported as a hash function.
A.2 Advanced Electronic Signatures based on
TS 101 903
TS 101 903 [19] uses a URN to reference the hash function in the ds:DigestMethod element. Since TS 101 903 [19] is
built upon XML DigSig, the algorithm requirements from XML DigSig apply.
Table A.2
AdES based on TS 101 903 [19] Issuers of AdES Users of AdES
Hash functions SHOULD support sha256 SHALL support sha1
SHOULD support sha224
SHOULD support sha256
Signature algorithms SHOULD support DSAwithSHA-1 SHALL support DSAwithSHA-1
SHOULD support RSAwithSHA-1
or ECDSA
NOTE: For canonicalization:
1. Required Canonical XML (omits comments):
http://www.w3.org/TR/2001/REC-xml-c14n-20010315
2. Recommended Canonical XML with Comments:
http://www.w3.org/TR/2001/REC-xml-c14n-20010315#WithComments
A.3 Signer's certificates
A signer certificate contains a subject public key and is signed by a CA issuing key. According to RFC 5280 [2],
implementations of that specification are not required to use any particular cryptographic algorithms. However,
conforming implementations which use the algorithms identified in RFC 3279 [12] MUST identify and encode the
public key materials and digital signatures as described in that specification. These requirements apply to signer public
keys and CA issuing keys.
RSA is actually the most supported algorithm and therefore it is the only algorithm bearing a SHOULD for issuers of
signer certificates. But this requirement does not inhibit a CA to issue DSA- or ECDSA-signed certificates, which may
be necessary if a DSA or ECDSA user key is to be certified.
ETSI
ETSI TS 102 176-1 V2.1.1 (2011-07)39
Table A.3
Signer certificates Issuers of signer certificates Users of signer certificates
Signer public keys SHOULD support RSA SHALL support RSA
SHOULD support DSA
SHOULD support ECDSA
CA issuing keys SHOULD support RSA with SHA-1
SHALL support RSA with SHA-256
SHALL support RSA with SHA-1
SHOULD support DSA with SHA-1
SHOULD support ECDSA with SHA-1
SHOULD support ECDSA with SHA-224
SHALL support RSA with SHA-256
SHOULD support DSA with SHA-256
SHOULD support ECDSA with SHA-256
NOTE: Because the usage of SHA-224 with RSA and DSA gives no advantage compared with SHA-256 neither
in security nor in performance it is not required that SHA-224 should be supported with these algorithms.
It is recommended that with RSA and DSA the hash functions SHA-256 and SHA-512 are used instead of
SHA-224 or SHA-384.
A.4 CRLs
A CRL is signed by a CRL Issuer. According to RFC 5280 [2], implementations of that specification are not required to
use any particular cryptographic algorithms. However, conforming implementations which use the algorithms identified
in RFC 3279 [12] MUST identify and encode the public key materials and digital signatures as described in that
specification. These requirements apply to CRL Issuer public keys.
RSA is actually the most supported algorithm and therefore it is the only algorithm bearing a SHOULD for then issuers
of CRLs.
Table A.4
CRLs Issuers of CRLs Users of CRLs
CRL issuer keys SHOULD support RSA with SHA-1
SHALL support RSA with SHA-256
SHALL support RSA with SHA-1
SHOULD support DSA with SHA-1
SHOULD support ECDSA with SHA-1
SHOULD support ECDSA with SHA-224
SHALL support RSA with SHA-256
SHOULD support DSA with SHA-256
SHOULD support ECDSA with SHA-256
NOTE: Because the usage of SHA-224 with RSA and DSA gives no advantage compared with SHA-256 neither
in security nor in performance it is not required that SHA-224 should be supported with these algorithms.
It is recommended that with RSA and DSA the hash functions SHA-256 and SHA-512 are used instead of
SHA-224 or SHA-384.
A.5 OCSP responses
A CRL is signed by an OCSP responder. The algorithm requirements from RFC 2560 [22] apply, i.e. "Clients that
request OCSP services SHALL be capable of processing responses signed used DSA keys identified by the DSA sigalg-oid
specified in clause 7.2.2 of RFC 5280 [2]. Clients SHOULD also be capable of processing RSA signatures as
specified in clause 7.2.1 of RFC 5280 [2]. OCSP responders SHALL support the SHA-1 hashing algorithm." These
requirements apply to OCSP the hash algorithm and the signature algorithm used by OCSP responders.
NOTE: The reference in RFC 2560 [22] to RFC 2459 [i.26] should be replaced by its successor RFC 5280 [2].
ETSI
ETSI TS 102 176-1 V2.1.1 (2011-07)40
Table A.5
OCSP response Issuers of OCSP responses Users of OCSP response
OCSP responder keys SHOULD support sha1 with dsa
SHOULD support sha1 with rsa
SHOULD support sha256 with rsa
SHALL support sha1 with dsa
SHALL support sha1 with rsa
SHOULD support sha25 with rsa
A.6 CA certificates
A CA certificate contains a CA public key and is signed by a CA private key. The algorithm requirements from
RFC 3279 [12] apply. These requirements apply to CA public keys (as subject) and CA public keys (as issuer).
Table A.6
CA certificates Issuers of CA certificates Users of CA certificates
Subject CA public key SHOULD support RSA with SHA-256 SHALL support RSA with SHA-1
SHOULD support DSA with SHA-1
SHOULD support ECDSA with SHA-1
SHOULD support ECDSA with SHA-224
SHALL support RSA with SHA-256
SHOULD support DSA with SHA-256
SHOULD support ECDSA with SHA-256
Issuer CA public keys SHOULD support RSA with SHA-256 SHALL support RSA with SHA-1
SHOULD support DSA with SHA-1
SHOULD support ECDSA with SHA-1
SHOULD support ECDSA with SHA-224
SHALL support RSA with SHA-256
SHOULD support DSA with SHA-256
SHOULD support ECDSA with SHA-256
NOTE: Because the usage of SHA-224 with RSA and DSA gives no advantage compared with SHA-256 neither
in security nor in performance it is not required that SHA-224 should be supported with these algorithms.
It is recommended that with RSA and DSA SHA-256 and SHA-512 are used instead of SHA-224 or
SHA-384.
A.7 Self-signed certificates for CA issuing CA certificates
A self-signed certificate contains a single root CA public key. The algorithm requirements from RFC 3279 [12] apply.
Self-signed certificates need to resist quite long (e.g. more than 10 years). For that reason, at least the hash function
SHA-256, used in combination with RSA, is recommended. These requirements apply to root CA public keys.
Table A.7
Self-signed certificates Issuers of self-signed certificates Users of self-signed certificates
Root CA public keys SHOULD support RSA with SHA-256 SHALL support RSA with SHA-1
SHOULD support DSA with SHA-1
SHOULD support ECDSA with SHA-1
SHOULD support ECDSA with SHA-224
SHALL support RSA with SHA-256
SHOULD support DSA with SHA-256
SHOULD support ECDSA with SHA-256
ETSI
ETSI TS 102 176-1 V2.1.1 (2011-07)41
A.8 TSTs based on RFC 3161 and TS 101 861
The following requirements apply to hash functions and TST signature algorithms. The algorithm requirements from
TS 101 861 [20] apply.
• for the requests: The following hash algorithms MAY be used to hash the information to be time-stamped:
SHA-1, RIPEMD-160;
• for the responses, the following signature algorithm must be supported: SHA-1 with RSA.
However, for time-stamp tokens that need to resist quite long (e.g. more than 10 years), the SHA-2 family (SHA-256,
SHA-384 and SHA-512) is recommended.
Table A.8
Time-Stamping Tokens TST requesters TST issuers TST verifiers
Hash function SHOULD support sha1
SHOULD support
ripemd160
SHOULD support sha256
SHALL support sha1
SHOULD support
ripemd160
SHOULD support sha256
SHALL support sha1
SHOULD support
ripemd160
SHOULD support sha256
TST signature algorithms SHALL support sha1 with
rsa
SHOULD support sha256
with rsa
SHALL support sha1 with
rsa
SHALL support sha256
with rsa
SHALL support sha1 with
rsa
SHOULD support sha256
with rsa
A.9 TSU certificates
A TSU certificate contains a TSU public key and is signed by a CA private key. The algorithm requirements from
RFC 3279 [12] apply. These requirements apply to TSU public keys (as subject) and CA public keys (as issuer).
Since TST last for longer time the SHA-2 family (SHA-256 and SHA-512) is recommended.
Table A.9
TSU certificates Issuers of TSU certificates Users of TSU certificates
TSU public key SHOULD support RSA with SHA-256 SHALL support RSA with SHA-1
SHOULD support DSA with SHA-1
SHALL support RSA with SHA-256
Issuer CA public keys SHOULD support RSA with SHA-256 SHALL support RSA with SHA-1
SHOULD support DSA with SHA-1
SHALL support RSA with SHA-256
A.10 Self-signed certificates for CAs issuing TSU
certificates
A self-signed certificate contains a single root CA public key. The algorithm requirements from RFC 3279 [12] apply.
Self-signed certificates need to resist quite long (e.g. more than 10 years). For that reason, the SHA-2 family (SHA-256
and SHA-512), used in combination with RSA, is recommended.
These requirements apply to root CA public keys.
ETSI
ETSI TS 102 176-1 V2.1.1 (2011-07)42
Table A.10
Self-signed certificates Issuers of self-signed certificates Users of self-signed certificates
root CA public keys SHOULD support RSA with SHA-256 SHALL support RSA with SHA-1
SHOULD support DSA with SHA-1
SHALL support RSA with SHA-256
A.11 Attribute certificates
An Attribute Certificate is signed by an Attribute Authority. The algorithm requirements from RFC 3279 [12] apply.
These requirements apply to Attribute Authority public keys.
Table A.11
Attribute Certificates Issuers of OCSP Attribute Certificates Users of OCSP Attribute Certificates
Attribute Authority public keys SHOULD support RSA with SHA-256 SHALL support RSA with SHA-1
SHOULD support DSA with SHA-1
SHALL support RSA with SHA-256
A.12 AA certificates
An AA certificate contains an Attribute Authority public key and is signed by a CA private key. The algorithm
requirements from RFC 3279 [12] apply. These requirements apply to Attribute Authority public keys (as subject) and
CA public keys (as issuer).
Table A.12
AA certificates Issuers of AA certificates Users of AA certificates
Attribute Authority public key SHOULD support RSA with SHA-256 SHALL support RSA with SHA-1
SHOULD support DSA with SHA1
SHALL support RSA with SHA-256
Issuer CA public keys SHOULD support RSA with SHA-256 SHALL support RSA with SHA-1
SHOULD support DSA with SHA-1
SHALL support RSA with SHA-256
ETSI
ETSI TS 102 176-1 V2.1.1 (2011-07)43
Annex B (informative):
Recommended key sizes (historical)
This annex will later on contain the outdated tables provided in clause 9 so that a history about previously
recommended hash functions and key sizes can be easily done at a given time and for a given time period.
B.1 Changes in 2005
2005-04: Taking in consideration the recently published attacks on hash functions, the clause 9.3 in the main
body is updated. The former text is provided here and the values that have been changed are printed in bold and
underlined.
10.3 (2005-04) Recommended hash functions versus time
Tables 6 and 7 provides indication about recommended hash functions during X years. With respect to the above
distinction between conservative and liberal views, the conservative view is first provided and then the liberal.
Definitions:
Usable: The algorithm with the given security parameters can be considered secure at the given time.
Unknown: The security of the algorithm is unknown; it may become secure if additional measures are applied.
Table 6/clause 9.3: Conservative view of recommended hash functions
for a resistance during X years
Entry name of the hash function 3 years
(2008)
5 years
(2010)
10 years
(2015)
sha1 usable unknown unknown
ripemd160 usable usable unknown
sha224 usable usable usable
sha256 usable usable usable
Whirlpool usable usable unknown
Table 7/clause 9.3: Liberal view for recommended hash functions
for a resistance during X years
Entry name of the hash function 3 years
(2008)
5 years
(2010)
10 years
(2015)
sha1 usable usable unknown
ripemd160 usable usable unknown
sha224 usable usable usable
sha256 usable usable usable
whirlpool usable usable unknown
Table B.3 predicts hash function resistance over 20 years. Due to the inherent unpredictability of such predictions, these
predictions are largely speculative, and no distinction is made be between conservative and liberal.
Additional definition:
Unusable: The algorithm cannot be considered secure for any kind of use in the context of electronic signatures.
ETSI
ETSI TS 102 176-1 V2.1.1 (2011-07)44
Table 8/clause 9.3: Speculated hash function resistance over 20 years,
based on current trends and estimated computational power
Entry name of the hash function Speculated Usability in 20 years
(2025)
sha1 unusable
ripemd160 unusable
sha224 usable
sha256 usable
Whirlpool usable
2005-04: Taking in consideration the recently published attacks on hash functions, the clause 9.4 in the main
body is updated. The former text is provided here and the values that have been changed are printed in bold and
underlined.
10.4 (2005-04) Recommended key sizes versus time
Tables B.4 and B.5 provides indication about recommended key lengths for a resistance of the algorithm or the
signature suite during X years. A conservative view and a liberal view are provided. For explanations of these terms
please refer to clauses 2 and 3 in this clause. These tables are provided for 3 years, 5 years and 10 years.
Definitions:
Unusable: The algorithm cannot be considered secure for any kind of use in the context of electronic signatures.
Unknown: The security of the algorithm is unknown at this time; it may become secure if additional measures are
applied.
Table 9/clause 9.4: Conservative view of recommended key lengths
for a resistance during X years
entry name of the signature suite 3 years
(2008)
5 years
(2010)
10 years
(2015)
sha-1-with-rsa ≥ 768 unknown unknown
sha224-with-rsa ≥ 768 1 024 2 048
sha256-with-rsa ≥ 768 1 024 2 048
RSASSA-PSS with mgf1SHA-1Identifier ≥ 768 1 024 unknown
RSASSA-PSS with mgf1SHA-224Identifier ≥ 768 1 024 2 048
RSASSA-PSS with mgf1SHA-256Identifier ≥ 768 1 024 2 048
sha1-with-dsa ≥ 768 unknown unknown
sha1-with-ecdsa 163 unknown unknown
sha224-with-ecdsa 224 224 224
sha256-with-ecdsa 256 256 256
Table 10/clause 9.4: Liberal view of recommended key lengths
for a resistance during X years
entry name of the signature suite 3 years
(2008)
5 years
(2010)
10 years
(2015)
sha-1-with-rsa ≥ 768 1 024 unknown
sha224-with-rsa ≥ 768 1 024 2 048
sha256-with-rsa ≥ 768 1 024 2 048
RSASSA-PSS with mgf1SHA-1Identifier ≥ 768 1 024 unknown
RSASSA-PSS with mgf1SHA-224Identifier ≥ 768 1 024 2 048
RSASSA-PSS with mgf1SHA-256Identifier ≥ 768 1 024 2 048
sha1-with-dsa ≥ 768 1 024 unknown
sha1-with-ecdsa 163 190 unknown
sha224-with-ecdsa 224 224 224
sha256-with-ecdsa 256 256 256
ETSI
ETSI TS 102 176-1 V2.1.1 (2011-07)45
An additional table with speculations regarding security over 20 years is provided. Due to the unpredictability in such
long term statements, there is no distinction made between liberal and conservative for the 20 year prediction, and such
information should be considered as largely speculative.
Table 11/clause 9.4: Speculated resistance of recommended algorithm parameters
for the next 20 years, based on current trends and estimated computational power
entry name of the signature suite 20 years
(2025)
sha-1-with-rsa unusable
sha224-with-rsa 2 048
sha256-with-rsa 2 048
RSASSA-PSS with mgf1SHA-1Identifier unusable
RSASSA-PSS with mgf1SHA-224Identifier 2 048
RSASSA-PSS with mgf1SHA-256Identifier 2 048
sha1-with-dsa 2 048
ecPublicKey unknown
sha1-with-ecdsa unknown
B.2 Changes in 2007
The document was revised and updated by an STF. Because the changes concern not only the tables, they are
documented in annex I and not here.
B.3 Changes in 2011
The document was updated by ESI taking in consideration the actual progress in cryptography. Based on publications
on cryptographic algorithms and their parameters from ECRYPT [i.6] and NIST [i.7] the usage of SHA-1 is no more
recommended. RIPEMD-160 is not broken and no attacks are known. Nevertheless it is no more recommended due to
hash length of 160 bit.
The recommended parameters for RSA are updated as it was foreseen by the previous version of the present document.
The 1 024 bit length is no more recommended.
ETSI
ETSI TS 102 176-1 V2.1.1 (2011-07)46
Annex C (informative):
Generation of RSA modulus
An RSA modulus is obtained by multiplying two prime numbers of roughly the same size. Furthermore, the two factors
must not be too close in order to be far enough from the square root of the modulus.
If we let p and q be the two prime factors of the modulus n, we can require that, for example:
0,1 < |log2(p) - log2(q)| < 30 (1)
which means that none of the factors is small or close to the square root of the modulus. This condition implies that:
log2(n)/2 - 15 < log2(p), log2(q) < log2(n) / 2 + 15 (2)
The generation of an RSA modulus of exactly k bits could be done with the following algorithm:
• Choose a random prime number p in the range ]2k/2-15, 2k/2+15[.
• Choose a random prime number q in the range [2k-1/p, 2k/p[.
• If the condition 0,1 < |log2(p)-log2(q)| < 30 is not satisfied, go back to the first step.
• Let n be the product of p and q.
A more complicated method that avoids the third step altogether but produces differently distributed primes is:
• Choose a random prime number p in the range [2k/2-9/20, 2k/2+15[.
• Choose a random prime number q in the range]a,b[ where a=max(ceil(2k-1/p)-1, p.2-30) and
b=min(2k/p, p.2-1/10).
• Let n be the product of p and q.
ETSI
ETSI TS 102 176-1 V2.1.1 (2011-07)47
Annex D (informative):
Generation of elliptic curve domain parameters
NOTE 1: This annex refers to ANSI X9.62 [7] (1998) which has been replaced by ANSI X9.62 [7] (2005). The
latter specifies ways to generate elliptic curve parameters except for the check of the class number
condition which is described in clause D.3. By the new version of the ANSI X9.62 [7] clauses D.1 and
D.2 have become obsolete. The information contained in clause D.3 can now also be found in the ECC
Brainpool paper on standard curves and curve generation [i.13].
This annex describes possible ways to generate elliptic curve domain parameters for ECDSA and ECGDSA satisfying
the conditions given in clauses 6.1.2.3 and 6.1.2.4 and also ways to select curves verifiably at random.
For this, basically the algorithms described in ANSI X9.62 [7] (1998) annex A.3 can be used. The only necessary
modifications are due to the fact that two of the conditions imposed on elliptic curves in the present document were not
included in ANSI X9.62 [7] (1998): In step 4. of algorithm A.3.2 only the condition about r0 (which is equivalent to the
"MOV condition" of ANSI X9.62 [7] (1998)) and the condition p ≠ q of clause 6.1.2.3 (which follows from the
"anomalous condition" of ANSI X9.62 [7] (1998)) are ensured while the condition on the class number of the maximal
order of the endomorphism ring of the curve and (in the F2
m case) the condition that the curve is not be definable over
F2 are not respected by the algorithm. These latter two conditions should also be checked during the generation of the
curve. Clauses D.1 and D.2 describe in more detail how the algorithm A.3.2 of ANSI X9.62 [7] (1998) can be modified
accordingly. The algorithms A.3.3.1 and A.3.3.2 of ANSI X9.62 [7] (1998) for selecting curves verifiably at random
which basically produce random j-invariants from a seed by means of a hash function only need to be modified in the
case that the security of SHA-1 is no longer regarded to be sufficient.
Clause D.3 gives more information about the class number condition and in particular describes how it can be checked.
For checking the class number condition an integer M has to be factorized and a certain triple (α,β,γ) of integers has to
be found. To ease the validation of the domain parameters the prime factorization of M and the triple (α,β,γ) should
always be made public together with the domain parameters. Otherwise in particular the factorization of M would
consume much time.
NOTE 2: Additional standard curves and further information about the class number condition can be found in the
ECC Brainpool paper "ECC Brainpool Standard Curves and Curve Generation" [i.13]; OID:
{1(iso) 3(identified organization) 36(teletrust) 3(algorithm) 3(signature algorithm) 2(ecSign)
8(ecStdCurvesAndGeneration)}.
D.1 ECDSA and ECGDSA based on a group E(Fp)
The prime p can be generated by one of the algorithms described in ISO/IEC 18032 [i.8] in a way that the probability of
being composite is at most 2-100.
The generation of an appropriate curve E, the point P and the prime q can be done with the algorithm in annex A.3.2 of
ANSI X9.62 [7] (1998) with lower bound rmin>2qMinLen, the MOV threshold B= r0Min and with Step 4. of algorithm
A.3.2 substituted by the following:
"4. (a) Check the MOV condition (see annex A.1.1) with inputs B, q and n. If the result is "false" go to Step 1.
(b) Check the Anomalous condition (see annex A.1.2). If the result is "false" go to Step 1.
(c) Find an element in the ideal class group of the number field K:=Q( d− ) of order at least MinClass where
d is the squarefree factor of 2
))(#1(4: pFEppM −+−= . If such an element of the class group cannot
be found go to Step 1."
The number M is always a positive integer and by the prime factorization of M one determines uniquely defined
positive integers ld, with 2
dlM = . Then d is called the squarefree factor of M .
ETSI
ETSI TS 102 176-1 V2.1.1 (2011-07)48
If an element of the ideal class group of order at least MinClass can be found then the class number condition of
clause 6.1.2.3 is satisfied.
Elements of the ideal class group of K can effectively be represented by certain triples (α,β,γ) of integers. The details in
particular about the group operation and the neutral element in this set of triples are given in clause D.3.
Here it is neither specified how to choose the elements in the ideal class group which are checked for sufficiently high
order nor after how many unsuccessful selections of elements to decide that the demanded element "cannot be found".
This is another reason to attach the triple (α,β,γ) (which has an order at least MinClass) to the domain parameters for
validation.
The algorithm to generate E, P and q can be successful only if p is chosen large enough, i.e. at least about as large as q
which itself is greater than 2qMinLen-1.
To select a curve verifiably at random one can use the algorithm given in annex A.3.3.1 of ANSI X9.62 [7]. The hash
function SHA-1 has to be substituted by a more secure hash function after the recommended use date for SHA-1.
D.2 ECDSA and ECGDSA based on a group E(F2
m)
The selection of an appropriate curve E, the point P and the prime q can be done with the algorithm in annex A.3.2 of
ANSI X9.62 [7] (1998) with lower bound rmin > 2qMinLen, with MOV threshold B=r0Min and with Step 4. of algorithm
A.3.2 substituted by the following:
"4. (a) Verify that b ≠ 1. If this is not the case go to Step 1.
(b) Check the MOV condition (see annex A.1.1) with inputs B, q and n. If the result is "false" go to Step 1.
(c) Find an element in the ideal class group of the number field K:=Q( d− ) of order at least MinClass where
d is the squarefree factor of 2
2
2
))(#12(2: mFEM mm
−+−= +
. If such an element in the ideal class group
cannot be found go to Step 1."
The parameter b is the constant term in the representation y2+xy=x3+ax2+b of the curve E used in the algorithm. b is the
j-invariant of E and as in a former step of the algorithm b ≠ 0 was already checked b ≠ 1 means that the j-invariant is not
contained in F2 and in particular that E cannot be defined over F2.
The number M is always a positive integer and by the prime factorization of M one determines the uniquely defined
positive integers ld, with 2
dlM = . Then d is called the squarefree factor of M .
If an element of the ideal class group of order at least MinClass can be found then the class number condition of
clause 6.1.2.4 is satisfied.
Elements of the ideal class group of K can effectively be represented by certain triples (α,β,γ) of integers. The details in
particular about the group operation and the neutral element in this set of triples are given in clause D.3.
Here it is neither specified how to choose elements in the ideal class group which are checked for sufficiently high order
nor after how many unsuccessful selections of elements to decide that the demanded element "cannot be found". This is
another reason to attach the triple (α,β,γ) (which has an order at least MinClass) to the domain parameters for validation.
The algorithm to generate E and P can be successful only if 2m is chosen large enough, i.e. at least about as large as q
which itself is greater than 2qMinLen-1.
To select a curve verifiably at random one can use the algorithm given in annex A.3.3.1 of ANSI X9.62 [7] (1998). The
hash function SHA-1 has to be substituted by a more secure hash function after the recommended use date for SHA-1.
ETSI
ETSI TS 102 176-1 V2.1.1 (2011-07)49
D.3 The class number condition
The class number condition was introduced because of the following reason: A hypothetical lift of the curve to an
elliptic curve over a number field cannot exist if the degree of the number field is less than the class number of the
endomorphism ring End(E) of E (regarded as order in the imaginary quadratic number field K:=Q( d− ) defined in
clauses D.1 and D.2 respectively) and if the degree of the number field is large then a solution of the corresponding lift
of the elliptic curve discrete logarithm problem is not feasible. The class number of End(E) is always a multiple of the
class number of K so what is actually demanded is a sufficiently large class number of K. Also the recent results of
Huang and Raskind (see bibliography) can be regarded as arguments for demanding a sufficiently large class number of
K.
Because the complexity of the best known algorithms for explicitly determining the class number of K is too high in
practice one just tries to find elements of the ideal class group of K with a large order as the class number is not smaller
than the order of an element.
Randomly selected curves will violate the class number condition with very low probability. But the best known
rigorously proven upper bounds do not exclude the possibility that the actual probability is significantly higher than 2-80
and heuristic arguments show that this probability should be at least of about the same magnitude as 2-80 (for a key
length of approximately 160 bits). So the class number condition should be checked also for randomly selected curves.
We now briefly describe how the elements of the ideal class group of the number field K:=Q( d− ) for a positive
squarefree integer d can be represented, how to determine the product of two elements and what the representation of
the neutral element with respect to this product looks like. Further information can be found in the text book of
Cohen [i.25].
First define
⎩
⎨
⎧
≡−≡−−
≡−−
=
)4(3)4(24
)4(1
:
dordifd
difd
D
A triple (α,β,γ) of integers satisfying =:
• gcd(α,β,γ)=1;
• α>0 and |β| ≤ α ≤ γ and if α=γ or |β|=α then also β ≥ 0;
• β2-4αγ= D ;
is called a primitive reduced triple of discriminant D .
NOTE 1: (α,β,γ) is a primitive reduced triple of discriminant D if and only if the quadratic form αx2+ βxy+γy2 is
primitive reduced and has discriminant D (which in particular implies that it is positive definite).
The elements of the ideal class group of the number field K:=Q( d− ) correspond one-to-one to the primitive reduced
triples of discriminant D . The group operation in this set of triples can be calculated as follows.
Given two primitive reduced triples (α1,β1,γ1) and (α2,β2,γ2) the so called composition (α´,β´,γ´) of (α1,β1,γ1) and
(α2,β2,γ2) can be determined by algorithm 5.4.7 of Cohen's book (see bibliography). This triple (α´,β´,γ´) is primitive
and has discriminant D but is not necessarily reduced. Applying the reduction algorithm 5.4.2 of the same book to this
triple (α´,β´,γ´) delivers a primitive reduced triple (α,β,γ) with determinant D . This triple represents the product of the
two elements representing (α1,β1,γ1) and (α2,β2,γ2).
(α1,β1,γ1) o (α2,β2,γ2) := (α,β,γ) (3)
The neutral element is represented by the triple (1,0,- D /4) if )4(0≡D and it is represented by the triple (1,1,(1- D )/4)
if )4(1≡D . In either case the triple corresponding to the neutral element is denoted by I.
The following is an algorithm that determines whether the order of an element of the ideal class group of the number
field K:=Q( d− ) has an order of at least MinClass:
Input: A primitive reduced triple (α,β,γ) of discriminant D .
ETSI
ETSI TS 102 176-1 V2.1.1 (2011-07)50
Output: The message "true" if the order of the corresponding element of the ideal class group is at least
MinClass; the message "false" otherwise.
1) Set t=I.
2) For i from 1 to MinClass-1 do:
- Set t:=t o (α,β,γ).
- If t=I then output "false" and stop.
3) Output "true".
In 2) the triple t o (α,β,γ) is calculated by the procedure described above.
NOTE 2: Most of the common computer algebra packages contain implementations for the described manipulations
in the class group of K or in the set of primitive reduced quadratic forms respectively.
ETSI
ETSI TS 102 176-1 V2.1.1 (2011-07)51
Annex E (informative):
On the generation of random data
E.1 Classes of random number generators
Figure E.1 shows a schematic classification of random number generators according to ISO/IEC FCD 18031 [i.14]
where more detailed information can be found. That document uses the term "random bit generator" while here the term
"random number generator" is used.
RNG
NRNG DRNG
physical
pure hybridd pure hybridd pure hybridd
non-phys.
Figure E.1: Schematic classification of random number generators
according to ISO/IEC FCD 18031 [i.14]
Every random number generator RNG has a primary entropy source. If this entropy source is non-deterministic which
means unrepeatable and unpredictable the RNG is called non-deterministic or a NRNG. If this entropy source consists
just of seed values it is called deterministic and also the RNG is called deterministic or a DRNG.
The primary entropy source of a NRNG can either be physical or non-physical. A physical primary entropy source (also
called physical primary noise source) uses dedicated hardware to measure the physical characteristics of a sequence of
events in the physical world, e.g. radioactive emissions of atoms or the noise of diodes. Typical non-physical primary
entropy sources are based for example on RAM contents, system clocks or "random user inputs" via PC-keyboard or
PC-mouse.
If the only entropy source for a RNG is the primary entropy source it is called pure RNG. A RNG can also have an
additional entropy source. A NRNG with an additional deterministic entropy source (i.e. seed values) is called hybrid
NRNG. A DRNG with an additional non-deterministic entropy source is called hybrid DRNG.
A well constructed NRNG is information theoretically secure while (pure) DRNGs can only be complexity theoretically
secure that means there is no feasible way to break its security. The advantage of the former is obviously that there is no
(even theoretical) possibility to calculate future or previous outputs from known ones. The security of DRNGs depends
on assumptions about the algorithmic complexity of certain problems which may turn out to be wrong sooner or later.
So NRNGs are better suited for long term security.
The following terminology for DRNGs is used in clause 8.2.2:
• A re-seeding of a DRNG is a complete new initialization of the DRNG with a newly produced seed.
• A seed-update of a DRNG is an external modification of the internal state (not by the regular updating
function of the DRNG) in a way that: (i) After the modification the modifier has no more information about
the internal state than before. (ii) Anybody else than the modifier having some information about the previous
internal state has less information about the internal state after the modification.
NOTE: The difference between these two possible ways to add new entropy to a DRNG is that if the new seed is
known then the future output is known after re-seeding while this is not the case for a seed-update.
ETSI
ETSI TS 102 176-1 V2.1.1 (2011-07)52
Of course it is desirable that after a seed-update the loss of information about the internal state in condition (ii) should
be as large as possible. In an ideal case there remains no information. A typical example for that is an XOR of the
internal state with a new seed produced with a strong NRNG.
E.2 On tests for NRNGs
Examples of generic test suites for the statistical properties of the primary entropy source can be found in
Ruhkin et al [i.18]. But usually tests specifically adapted to the mathematical model of the source are more suitable.
Online tests should be specific to the primary entropy source. An example for such an online test can be found in
example E7 of AIS 31 [i.19].
To avoid a misunderstanding about tests and test suites it should be pointed out that:
There is no test or test suite which can show that the output of a generator has a certain minimum entropy
without certain additional statistical assumptions about the source.
EXAMPLE: The term "universal" for Maurer's test (Maurer 1991, [i.20]) could cause some confusion about this
fact. Actually in Maurer's article it is assumed that the source is a binary, stationary, ergodic
source with finite memory. These assumptions are explicitly mentioned in that article.
It should also be observed that an NRNG has to be evaluated as an entire system i.e. taking into account the interaction
of the components. Thus it is not enough to regard the mathematical model, the online tests and the evaluation of the
post-treatment separately.
ETSI
ETSI TS 102 176-1 V2.1.1 (2011-07)53
Annex F (informative):
Algorithm identifiers defined in various documents
F.1 Algorithm identifiers defined in RFC 3278
The title of the document is: "Use of Elliptic Curve Cryptography (ECC) Algorithms in Cryptographic Message Syntax
(CMS)" [11].
Signature suite
ECDSA with SHA-1
ecdsa-with-SHA-1 OBJECT IDENTIFIER ::=
{ iso(1) member-body(2) us(840) 10045 signatures(4) 1 }
F.2 Algorithm identifiers defined in RFC 3279
The title of the document is: "Algorithms and Identifiers for the Internet X.509 Public Key Infrastructure Certificate and
Certificate Revocation List (CRL) Profile" [12].
Signature suites for CA issuing keys and CRL issuing keys
RSA with SHA-1
sha-1WithRSAEncryption OBJECT IDENTIFIER ::=
{ iso(1) member-body(2) us(840) rsadsi(113549) pkcs(1) pkcs-1(1) 5 }
DSA with SHA-1
id-dsa-with-sha1 OBJECT IDENTIFIER ::=
{ iso(1) member-body(2) us(840) x9-57 (10040) x9cm(4) 3 }
ECDSA with SHA-1
ecdsa-with-SHA-1 OBJECT IDENTIFIER ::=
{ iso(1) member-body(2) us(840) 10045 signatures(4) 1 }
Preferred signature algorithms for subject public keys (any is allowed)
RSA
rsaEncryption OBJECT IDENTIFIER ::=
{ iso(1) member-body(2) us(840) rsadsi(113549) pkcs(1) 1 1}
DSA
id-dsa OBJECT IDENTIFIER ::=
{ iso(1) member-body(2) us(840) x9-57(10040) x9cm(4) 1 }
ECDSA
id-ecPublicKey OBJECT IDENTIFIER ::=
{ iso(1) member-body(2) us(840) 10045 2 1 }
ETSI
ETSI TS 102 176-1 V2.1.1 (2011-07)54
F.3 Algorithm identifiers defined in RFC 3370
The title of the document is: "Cryptographic Message Syntax (CMS) Algorithms" [13].
Hash-functions
sha-1 OBJECT IDENTIFIER ::=
{ iso(1) identified-organization(3) oiw(14) secsig(3) algorithm(2) 26 }
Signature suite
DSA defined in RFC 3370 [13] is always used with the SHA-1 message digest algorithm. The algorithm identifier for
DSA is:
id-dsa-with-sha1 OBJECT IDENTIFIER ::=
{ iso(1) member-body(2) us(840) x9-57 (10040) x9cm(4) 3 }
An alternative OID for DSA with SHA-1 is defined in ISO/IEC 14888-3 [8]
id-dswa-dl-DSA OBJECT IDENTIFIER ::=
{ iso(1) standard(0) digital-signature-with-appendix(14888) part3(3) algorithm(0) 1}
F.4 Algorithm identifiers defined in RFC 3447
The title of the document is: "PKCS #1: RSA Cryptography Specifications" [14].
Signature algorithm
The algorithm identifier for RSA is:
rsaEncryption OBJECT IDENTIFIER ::=
{ iso(1) member-body(2) us(840) rsadsi(113549) pkcs(1) pkcs-1(1) 1 }
F.5 Algorithm identifier defined in RFC 3874
The title of the document is: "A 224-bit One-way Hash Function: SHA-224" [27].
id-sha224
id-sha224 OBJECT IDENTIFIER ::= { joint-iso-itu-t(2) country(16) us(840) organization(1) gov(101)
csor(3) nistalgorithm(4) hashalgs(2) sha224(4) }
F.6 Algorithm identifiers defined in XML-Signature
Syntax and Processing W3C Recommendation
This recommendation from February 12, 2002 is about "XML-Signature Syntax and Processing" [30].
Hash-function
SHA-1: http://www.w3.org/2000/09/xmldsig#sha1
Signature suite
DSAwithSHA-1 (DSS): http://www.w3.org/2000/09/xmldsig#dsa-sha1 (Required)
RSAwithSHA-1 (RSA): http://www.w3.org/2000/09/xmldsig#rsa-sha1 (Recommended)
ETSI
ETSI TS 102 176-1 V2.1.1 (2011-07)55
F.7 Algorithm identifiers defined in XML Encryption
Syntax and Processing. W3C Recommendation
This recommendation from December 10, 2002 is about "in XML Encryption Syntax and Processing" [31].
Hash-functions
SHA-1: http://www.w3.org/2000/09/xmldsig#sha1
SHA-256: http://www.w3.org/2001/04/xmlenc#sha256
RIPEMD-160: http://www.w3.org/2001/04/xmlenc#ripemd160
F.8 Algorithm identifiers defined in RFC 4050
The title of the document is: "Using the Elliptic Curve Signature Algorithm (ECDSA) for XML Digital
Signatures" [28].
Signature suite
ECDSA: http://www.w3.org/2001/04/xmldsig-more#ecdsa-sha1
NOTE: Like DSA, ECDSA incorporates the use of a hash function. The only hash function defined for use with
ECDSA defined in this RFC is the SHA-1 message digest algorithm.
F.9 Algorithm identifiers defined in RFC 4051
The title of the document is: "Additional XML Security Uniform Resource Identifiers (URIs)" [29].
SHA-224
http://www.w3.org/2001/04/xmldsig-more#sha224
RSA-SHA-256
http://www.w3.org/2001/04/xmldsig-more#rsa-sha256
RSA-RIPEMD160
http://www.w3.org/2001/04/xmldsig-more/rsa-ripemd160
ECDSA-SHA
http://www.w3.org/2001/04/xmldsig-more#ecdsa-sha1
http://www.w3.org/2001/04/xmldsig-more#ecdsa-sha224
http://www.w3.org/2001/04/xmldsig-more#ecdsa-sha256
ETSI
ETSI TS 102 176-1 V2.1.1 (2011-07)56
F.10 Algorithm identifiers defined in RFC 4055
The title of the document is: "Additional Algorithms and Identifiers for RSA Cryptography for use in the Internet X.509
Public Key Infrastructure. Certificate and Certificate Revocation List (CRL) Profile".
RFC 4055 [15] supplements RFC 3279 [12] to describe how to use some newer cryptographic algorithms.
Hash-functions
id-sha224 OBJECT IDENTIFIER ::= {{ joint-iso-itu-t(2) country(16) us(840) organization(1)
gov(101) csor(3) nistalgorithm(4) hashalgs(2) 4 }
id-sha256 OBJECT IDENTIFIER ::= { joint-iso-itu-t(2) country(16) us(840) organization(1) gov(101)
csor(3) nistalgorithm(4) hashalgs(2) 1 }
id-sha384 OBJECT IDENTIFIER ::= { joint-iso-itu-t(2) country(16) us(840) organization(1) gov(101)
csor(3) nistalgorithm(4) hashalgs(2) 2 }
id-sha512 OBJECT IDENTIFIER ::= { joint-iso-itu-t(2) country(16) us(840) organization(1) gov(101)
csor(3) nistalgorithm(4) hashalgs(2) 3 }
Mask Generation functions
mgf1SHA-1Identifier AlgorithmIdentifier ::= { id-mgf1, sha1Identifier }
mgf1SHA-224Identifier AlgorithmIdentifier ::= { id-mgf1, sha224Identifier }
mgf1SHA-256Identifier AlgorithmIdentifier ::= { id-mgf1, sha256Identifier }
mgf1SHA-384Identifier AlgorithmIdentifier ::= { id-mgf1, sha384Identifier }
mgf1SHA-512Identifier AlgorithmIdentifier ::= { id-mgf1, sha512Identifier }
Signature algorithms
id-RSASSA-PSS OBJECT IDENTIFIER ::= { pkcs-1 10 }
Signature suites
sha224WithRSAEncryption OBJECT IDENTIFIER ::= { pkcs-1 14 }
sha256WithRSAEncryption OBJECT IDENTIFIER ::= { pkcs-1 11 }
sha384WithRSAEncryption OBJECT IDENTIFIER ::= { pkcs-1 12 }
sha512WithRSAEncryption OBJECT IDENTIFIER ::= { pkcs-1 13 }
ETSI
ETSI TS 102 176-1 V2.1.1 (2011-07)57
Annex G (informative):
Abstracts of ISO/IEC 10118-3 and ISO/IEC 9796-2
Abstract of ISO/IEC 10118-3 [3]
ISO/IEC 10118-3 [3] specifies the following seven dedicated hash-functions, i.e. specially-designed hash-functions:
1) RIPEMD-160 in clause 7 provides hash-codes of lengths up to 160 bits.
2) RIPEMD-128 in clause 8 provides hash-codes of lengths up to 128 bits.
3) SHA-1 in clause 9 provides hash-codes of lengths up to 160 bits.
4) SHA-256 in clause 10 provides hash-codes of lengths up to 256 bits.
5) SHA-512 in clause 11 provides hash-codes of lengths up to 512 bits.
6) SHA-384 in clause 12 provides hash-codes of a fixed length, 384 bits.
7) WHIRLPOOL in clause 13 provides hash-codes of lengths up to 512 bits.
For each of these dedicated hash-functions, ISO/IEC 10118-3 [3] specifies a round-function that consists of a sequence
of sub-functions, a padding method, initializing values, parameters, constants, and an object identifier as normative
information, and also specifies several computation examples as informative information.
Abstract of ISO/IEC 9796-2 [17]
ISO/IEC 9796-2 [17] specifies three digital signature schemes giving message recovery, two of which are deterministic
(non-randomized) and one of which is randomized. The security of all three schemes is based on the difficulty of
factorizing large numbers. All three schemes can provide either total or partial message recovery.
The method for key production for the three signature schemes is specified in ISO/IEC 9796-2 [17]. However,
techniques for key management and for random number generation (as required for the randomized signature scheme),
are outside the scope of ISO/IEC 9796-2 [17].
Wherever possible, the second mechanism (Digital signature scheme 2) is RECOMMENDED. However, in
environments where generation of random variables by the signer is deemed infeasible, then Digital signature scheme 3
is RECOMMENDED. Digital signature scheme 1 SHALL only be used in environments where compatibility is
required with systems implementing the first edition of this International Standard.
ETSI
ETSI TS 102 176-1 V2.1.1 (2011-07)58
Annex H (informative):
Signature maintenance
An advanced electronic signatures SHOULD be verified according to a signature policy that meets the business needs.
There may exist valid reasons under particular circumstances to use a signature policy different from the one which
should normally be used. In such a case, the full implications must be understood and carefully weighted by the verifier.
A signature policy MAY include constraints about which algorithms and key lengths are deemed appropriate under that
policy and/or define a time beyond which the algorithms/keys related to an advanced electronic signature should not be
trusted anymore, unless additional security measures are taken.
It may be needed to re-verify advanced electronic signatures (this is called a subsequent verification) well beyond the
time they were initially verified. At the time of re-verification, trust anchors and algorithms that were initially defined in
the signature policy may not be secure anymore. Additional security measures need to be taken so that this can be done.
It may also happen that some keys were secure at the time the initial verification of an advanced electronic signature
was performed, but due to some "accident" this is no more the case later on (e.g. due to a key compromise).
In both cases, it is possible to maintain the security of an advanced electronic signature which has already been
successfully verified. This may be done with security measures such as:
• the secure archival of both the definition of the signature policy (or an unambiguous reference to it) and all the
data initially used to verify the advanced electronic signature according to that signature policy; or
• the secure archival of both the definition of the signature policy and the addition to the advanced electronic
signature of other data (e.g. time-stamps) that will allow subsequent verifications.
These measures may be defined in the signature policy itself or "elsewhere" in a set of rules called a "signature
maintenance policy" which will allow to maintain the validity of advanced electronic signatures.
At a time where it is possible or likely that the algorithms and key lengths originally used will not be secure anymore,
an application of a signature maintenance process allows nevertheless to re-verify advanced electronic signatures under
a given signature policy. The sooner the process is applied, the better.
ETSI
ETSI TS 102 176-1 V2.1.1 (2011-07)59
Annex I (informative):
Major changes from previous versions
The version 2.0.0 of the present document was a revision and update of clause 9 of version 1.2.1. Beside error
correction and update of some references the following changes were made.
4. Maintenance of the document
The maintenance activity was described in more details.
5. Hash functions
In the table of recommended hash functions SHA-384 and SHA-512 were included, and their description is given in
new clauses. This leads also to new signature suites in clause 7.
6. Signature schemes
The description was extended for clarification.
7. Signature suites
New signature suites with ECDSA and new hash function were added. This implies also additions in the corresponding
tables of OIDs in clause 11.
9. Recommended hash functions and key sizes versus time
The almost undistinguishable liberal and conservative was removed and all the tables were rebuilt.
Annex A (normative): Algorithms for various data structures
The phasing out of SHA-1 was reflected in the reworked tables of the sections. The recommendation or support of
SHA-256 and the based on its signature suites were added.
Annex I (informative): Major changes from previous versions
The version 2 of this multi-part deliverable was a revision and update of clause 9. Therefore this former empty annex
was used for the description of the major changes.
Annex J (informative): National Bodies
New clause was added indicating the National Bodies and their contacts at the time of publication of the present
document.
ETSI
ETSI TS 102 176-1 V2.1.1 (2011-07)60
Annex J (informative):
National Bodies
This annex contains references to national bodies known at issuing time. It is based on the information of FESA
(http://www.fesa.eu) and lists Bodies responsible for supervision (according Art. 11 of the Directive 1999/93/EC [1]).
Contact details of notified supervisory authorities are available at:
http://ec.europa.eu/information_society/policy/esignature/index_en.htm
1. Austria
Telekom-Control-Kommission
Telekom-Control Commission
Mariahilfer Str. 77-79, A-1060 Wien/Vienna
Tel: +43/1/58058-0
Fax: +43/1/58058-9191
Email: signatur@signatur.rtr.at
Website: http://www.signatur.rtr.at/
2. Belgium
SPF Economie, PME, Classes moyennes et Energie - Qualité et Sécurité - Information Management / FOD
Economie, KMO, Middenstand en Energie - Kwaliteit en Veiligheid - Information Management / FÖD
Wirtschaft, KMB, Mittelstand und Energie - Qualität und Sicherheit - Informationsmanagement
FPS Economy, SMEs, Self-employed and Energy - Quality and Security - Information Management
NG III - 4 verdiep, Koning Albert II-laan 16, 1000 BRUSSEL
Tel: +32 2 277 74 30
Fax: +32 2 277 54 01
Email: be.sign@economie.fgov.be
Website: http://www.mineco.fgov.be
3. Bulgaria
Communications Regulation Commission (CRC)
Комисия за регулиране на съобщенията
гр. София 1000, ул. Гурко 6
Email: info@crc.bg
Website: http://www.crc.bg
4. Czech Republic
Ministerstvo vnitra
Ministry of the Interior
Nad Štolou 3, poštovní schránka 21, 170 34 Praha 7
Email: posta@mvcr.cz
Website: http://www.mvcr.cz
5. Cyprus
Υπουργείο Εμπορίου, Βιομηχανίας και Τουρισμού
Ministry of Communications and Works
286 Strovolos Avenue, 2048 Nicosia
E-Mail: mailto:info.dec@mcw.gov.cy
Website: http://www.mcw.gov.cy/dec
6. Denmark
IT- og Telestyrelsen
National IT and Telecom Agency
Holsteinsgade 63, 2100 København Ø
Tel: 4535450000
Email: digsig@itst.dk
Website: http://www.itst.dk
ETSI
ETSI TS 102 176-1 V2.1.1 (2011-07)61
7. Estonia
Sideamet
Estonian National Communications Board
Ädala 2, Tallinn 10614
Tel: (0) 693 1154
Fax: (0) 693 1155
Email: sideamet@sa.ee
Website: http://www.sa.ee/
8. Finland
Viestintävirasto
Finnish Communications Regulatory Authority (FICORA)
Itämerenkatu 3 A, PO BOX 313, FI-00181 Helsinki
Tel: +358 9 69661
Fax: +358 9 6966 873
Website: http://www.ficora.fi
9. France
Agence nationale de la sécurité des systèmes d’information (ANSSI)
Secrétariat général de la défense et de la sécurité nationale
51, boulevard de La Tour-Maubourg, F-75700 Paris 07 SP
Tel: 0033171758405
Fax: 0033171758400
Email: secretariat.anssi@ssi.gouv.fr
Website: http://www.ssi.gouv.fr/
10. Germany
Bundesnetzagentur für Elektrizität, Gas, Telekommunikation, Post und Eisenbahnen
Federal Network Agency for Electricity, Gas, Telecommunications, Post and Railway
Referat Elektronische Signatur, Canisiusstr. 21, D 55122 Mainz
Tel: +49 6131 18-0
Fax: +49 6131 18-5618
Email: ElektronischeSignatur@BNetzA.de
Website: http://www.bundesnetzagentur.de/
11. Greece
Εθνικη Επιτροπη Τηλεπικοινωνιων και Ταχυδρομειων (EETT)
National Telecommunications and Post Commission (EETT)
60, Kifissias Av. , 151 25 Maroussi Athens GREECE
Tel: +30 10 6151000
Fax: +30 10 6105049
Email: info@eett.gr
Website: http://www.eett.gr
12. Hungary
Nemzeti Média- és Hírközlési Hatóság
National Media and Infocommunications Authority
H- 1015 Budapest, Ostrom u. 23-25.
Tel: +36-1-4577100
Fax: +36-1-356-5520
Email: info@nmhh.hu
Website: http://www.nmhh.hu/
13. Iceland
Neytendastofa, The Consumer Agency
Borgartúni 21, 105 Reykjavík
Tel +354 510 1126
Fax +354 510 1101
Email: postur@neytendastofa.is
Website: http://www.neytendastofa.is
14. Ireland
Department of Communications, Marine and Natural Resources
29-31 Adelaide Road, Dublin 2
Tel +353 1 6782000
Fax +353 1 6782449
Website: http://www.dcmnr.gov.ie
ETSI
ETSI TS 102 176-1 V2.1.1 (2011-07)62
15. Italy
DigitPA - Ente nazionale per la digitalizzazione della pubblica amministrazione
National Center for IT in the Public Administration
Viale Marx 31/49, 00137 Roma
Tel: +39 06 852641
Fax: +39 06 85264 137
Email: arbia@cert.cnipa.it
Website: http://www.digitpa.gov.it
16. Latvia
Datu valsts inspekcija, Data State Inspectorate
Kr.Barona iela 5-4, Rīga, LV-1050
Website: http://www.dvi.gov.lv
Email:info@dvi.gov.lv
Tel +371 2 7223131
Fax +371 2 7223556
17. Liechtenstein
Amt für Kommunikation, Office for Communications
Kirchstrasse 10, Postfach 681, 9490 Vaduz
Tel +423 236 64 88
Fax +423 236 64 89
Email: info@ak.llv.li
Website: http://www.llv.li/amtsstellen/llv-ak-home.htm
18. Lithuania
Lietuvos Respublikos ryšiu reguliavimo tarnyba
Communications Regulatory Authority of the Republic of Lithuania
Algirdo str. 27A, LT-03219 Vilnius, Lithuania
Tel: +370 5 210 5633
Fax: +370 5 216 1564
Email: rrt@rrt.lt
Website: http://www.rrt.lt/
19. Luxembourg
Institut Luxembourgeois de la Normalisation, de l'Accréditation, de la Sécurité et qualité des produits et
services (ILNAS)
34-40 Avenue de la Porte-Neuve; L-2227 Luxembourg
Tel: +352 46 97 46 1
Fax: +352 22 25 24
Email: info@ilnas.public.lu
Website: http://www.ilnas.public.lu
20. Malta
Awtorita' ta' Malta dwar Il-Komunikazzjoni
Malta Communications Authority
Il-Piazzetta, Suite 43/44, Tower Road, Sliema SLM 16, Malta, Europe
Tel: +356 21 336 840
Fax: +356 21 336 846
Email: mca@mca.org.mt
Website: http://www.mca.org.mt/
21. Netherlands
Onafhankelijke Post en Telecommunicatie Autoriteit (OPTA)
Independent Post and Telecommunications Authority (OPTA)
Zurichtoren, Muzenstraat 41, Postbus 90420, 2509 LK, The Hague
Tel: +31703153500
Fax: +31703153501
Email: ttp@opta.nl
Website: http://www.opta.nl/
22. Norway
Post- og teletilsynet; Norwegian Post and Telecommunications Authority (NPT)
postal address: Postboks 93, 4791 Lillesand
Tel +47 22 82 46 00
Fax +47 22 82 46 40
Email: firmapost@npt.no
Website: http://www.npt.no/
ETSI
ETSI TS 102 176-1 V2.1.1 (2011-07)63
23. Poland
Ministerstwo Gospodarki
Ministry of Economy
Pl. Trzech Krzyzy 3/5, 00-507 Warsaw, Poland
Email: Marcin.Fijalkowski@mg.gov.pl
Website: http://www.mg.gov.pl
24. Portugal
Autoridade Nacional de Segurança
National Security Authority
Av. Ilha da Madeira, 1, 8º
1449-004 Lisboa
Tel +351 213 031 710
Fax +351 213 031 711
Email: gns@netcabo.pt
25. Romania
Ministry of Communications and Information Technology
Bd.Libertăţii nr.14, 050706, Bucureşti
Email: arsadmin@mcti.ro
Website: http://ars.mcti.ro
26. Slovak Republic
Národný bezpecnostný úrad
National Security Authority
Budatinska 30, P.O. BOX 16 850 07 Bratislava
Tel: +421-2-68699503
Fax: +421-2-63824005
Email: sep@nbusr.sk
Website: http://www.nbusr.sk/en/electronic-signature
27. Slovenia
Ministry of the Economy
Inspectorate for electronic communications, electronic signature and post (Ministrstvo za gospodarstvo;
Inšpektorat Republike Slovenije za elektronske komunikacije, elektronsko podpisovanje in pošto),
Stegne 7, SI-1000 Ljubljana,
Website: http://www.mg.gov.si/
28. Sweden
Post- och telestyrelsen
National Post and Telecom Agency
P.O. Box 5398, SE-102 49 Stockholm, Sweden
Tel: +46 8 678 55 00
Fax: +46 8 678 55 05
Email: pts@pts.se
Website: http://www.pts.se
29. Spain
Secretaría de Estado de Telecomunicaciones y para la Sociedad de la Información (SETSI). Ministerio de
Industria, Turismo y Comercio (MITYC)
State Secretariat for Telecommunications and for the Information Society (SETSI). Ministry of Industry,
Tourism and Trade (MITYC).
C/ Capitán Haya, 41. 28020 Madrid. España-Spain
Tel: +34 91 346 1597
Fax: +34 91 346 1577
Website: http://www.mityc.es
30. United Kingdom
Department of Business, Innovation & Skills
151 Buckingham Palace Road, London, SW1 9SS
Tel: +44 207 215 1961
Fax: +44 207 931 7194
ETSI
ETSI TS 102 176-1 V2.1.1 (2011-07)64
Annex K (informative):
Bibliography
• ECRYPT Yearly Report on Algorithms and Keysizes (2005). Revision 1.1, 29 January 2007.
• ECRYPT Position Paper: Recent Collision Attacks on Hash Functions. Revision 1.1 17 February 2005.
• IETF RFC 5462 (2009): "Multi-Protocol Label Switching (MPLS) Support of Differentiated Services".
• [LenstraVerheul] Selecting Cryptographic Key-sizes, A.K. Lenstra E.R. Verheul, Journal of Cryptology 14
(2001) pp 255-293.
• Dobbertin, H., Bosselaers, A. and Preneel, B., "RIPEMD-160: A strengthened version of RIPEMD", in Fast
software encryption, Proceedings of the Third International Workshop, Cambridge, UK, February 21-23,
1996, pp. 71-82, D. Gollmann (ed.), LNCS 1039, Springer-Verlag, 1996.
• Rivest, R.L., "Finding four million random primes", in Advances in Cryptology - Crypto '90, pp. 625-626, A.J.
Menezes (ed.), LNCS 537, Springer-Verlag, 1991.
• Blum, M. and Micali, S., "How to generate cryptographically strong sequences of pseudo-random bits", SIAM
Journal on Computing, Vol. 4, No. 13, pp. 850-863, 1984.
• Rivest, R., Shamir, A. and Adleman, L., "A method for obtaining digital signatures and public key
cryptosystems", Communications of the ACM, Vol. 21, No. 2, pp. 120-126, 1978.
• BSI. Technical Guideline: Elliptic Curve Cryptography (ECC). Technical Guideline TR-03111, version 1.11,
2009, https://www.bsi.bund.de/ContentBSI/Publikationen/TechnischeRichtlinien/tr03111/index_htm.html
• Bundesnetzagentur für Elektrizität, Gas, Telekommunikation, Post und Eisenbahnen, "Bekanntmachung zur
elektronischen Signatur nach dem Signaturgesetz und der Signaturverordnung (Übersicht über geeignete
Algorithmen)", German Federal Gazette No. 85 (2011), pp. 2034ff.,
http://www.bundesnetzagentur.de/SharedDocs/Downloads/DE/BNetzA/Sachgebiete/QES/Veroeffentlichungen
/Algorithmen/2011_2_AlgoKatpdf.pdf.
• "Verordnung über elektronische Signaturen (Signaturverordnung – SigV)", Austrian Federal Law Gazette II
No. 30/2000 as amended by Austrian Federal Law Gazette II No. 527/2004.
• RSA Security, Inc., "RSA Laboratories' Frequently Asked Questions About Today's Cryptography",
Version 4.1, July 2000.
• GMD, "SECUDE: Algorithms", SECUDE-5.1, November 1997.
• Galbraith, S.D., and Smart, N.P., "A Cryptographic Application of Weil Descent", in Cryptography and
Coding, M. Walker (ed.), LNCS 1746, Springer-Verlag, 1999.
• FESA - Forum of European Supervisory Authorities, http://www.fesa.rtr.at
• Johnson, D.B., and Menezes, A.J., "Elliptic Curve DSA (ECDSA).
• TeleTrust OID tree structure, Version 0.3.
NOTE: Available at: http://www.teletrust.de/projekte/oid/
• Damgard, I., Landrock P.and Pomerance C., "Average case error estimates for the strong probable prime test",
Math. Comp., 61:177-194, 1993.
• Knuth, D., "The art of computer programming - Volume 2 / Seminumerical algorithms (chapter 3)",
Addison-Wesley, 1981.
• Kelsey, J., Schneier, B., Wagner, D. and Hall, C., "Cryptanalytic Attacks on Pseudorandom Number
Generators", Fast Software Encryption '98, LNCS 1372, pp. 168-188, 1998.
ETSI
ETSI TS 102 176-1 V2.1.1 (2011-07)65
• M.D. Huang, W. Raskind "Global Methods for Discrete Logarithm Problems".
See: http://www.cacr.math.uwaterloo.ca/conferences/2004/ecc2004/huang.pdf
• "A Unified Approach to the Discrete Logarithm Problem for the Multiplicative Group and for Elliptic Curves
over finite Fields". September 20, 2004. See:
http://www.cacr.math.uwaterloo.ca/conferences/2004/ecc2004/raskind.pdf
• Common ISIS-MTT Specification for interoperable PKI applications. Part 6. Cryptographic Algorithms.
Version 1.1. 16 March 2004
• FIPS Draft Publication 186-3 (2006): "Digital Signature Standard (DSS)".
• IETF RFC 5758: "Internet X.509 Public Key Infrastructure: Additional Algorithms and Identifiers for DSA
and ECDSA".
NOTE: Available at: http://www.ietf.org/rfc/rfc5758.txt.
• TeleTrusT Deutschland e. V., "OID-Liste".
NOTE: Available at: http://www.teletrust.de/index.php?id=171.
• The Digital Signature Scheme ECGDSA.
NOTE: Available at: http://www.teletrust.de/fileadmin/files/oid/ecgdsa_final.pdf.
• ANSSI, Référentiel Général de Sécurité Version 1.0, Annexe B1 Mécanismes cryptographiques, Règles et
recommandations concernant le choix et le dimensionnement des mécanismes cryptographiques Version 1.20,
2010-01-26.
NOTE: See http://www.ssi.gouv.fr/IMG/pdf/RGS_B_1.pdf.
• Rundfunk und Telekom Regulierungs-GmbH (RTR-GmbH), Empfohlene Algorithmen und Parameter für
elektronische Signaturen, 2006-04-03.
NOTE: See http://www.signatur.rtr.at/repository/rtr-algorithms-20070601-de.pdf.
• Nemzeti Hírközlési Hatóság, A hitelesítés-szolgáltatás területén alkalmazható kriptográfiai algoritmusokról és
paraméterekről szóló NHH határozat melléklete, 2006-04-05.
NOTE: See http://www.nhh.hu/dokumentum.php?cid=9201
• Bundesnetzagentur für Elektrizität, Gas, Telekommunikation, Post und Eisenbahnen, Bekanntmachung zur
elektronischen Signatur nach dem Signaturgesetz und der Signaturverordnung (Übersicht über geeignete
Algorithmen), 2011-05-20.
NOTE: See
http://www.bundesnetzagentur.de/SharedDocs/Downloads/DE/BNetzA/Sachgebiete/QES/Veroeffentlichu
ngen/Algorithmen/2011_2_AlgoKatpdf.pdf?__blob=publicationFile
ETSI
ETSI TS 102 176-1 V2.1.1 (2011-07)66
History
Document history
V1.1.1 March 2003 Publication as SR 002 176
V1.2.1 July 2005 Publication
V2.0.0 November 2007 Publication
V2.1.1 July 2011 Publication