130 likes | 275 Vues
Workshop on algorithms and parameters for Electronic Signatures. November 25, 2004. Brussels. Algorithms and parameters for Electronic Signatures. Two parts: Part 1: Hash functions and asymmetric algorithms Part 2: Symmetric algorithms and protocols for secure channels.
E N D
Workshop on algorithms and parameters for Electronic Signatures November 25, 2004. Brussels
Algorithms and parameters for Electronic Signatures • Two parts: • Part 1: Hash functions and asymmetric algorithms • Part 2: Symmetric algorithms and protocols for secure channels
Part 1: Hash functions and asymmetric algorithms • 4. Maintenance of the document • 5. Hash functions • 5.1. General • 5.2. Recommended one way hash functions • 6. Signature algorithms • 6.1. General • 6.2. Recommended signature algorithms • 7. Signature suites • 7.1.General • 7.2. Padding methods • 7.3. Recommended signature suites
Part 1: Hash functions and asymmetric algorithms • 8. Recommended key pair generation methods • 8.1. General • 8.2. Recommended key pair generation methods • 9. Random number generation methods • 9.1. General • 9.2. Recommended random number generation methods • 10. Recommended hash functions and key sizes versus time • 10.1. Liberal view • 10.2. Conservative view • 10.3. Recommended hash functions versus time • 10.4. Recommended key sizes versus time
Part 1: Hash functions and asymmetric algorithms • 11. Practical ways to identify hash functions and algorithms • 11.1 Functions and algorithms identified using OIDs • 11.2 Functions and algorithms identified using URNs • 11.3 Functions and algorithms with no OID • 11.4 Functions and algorithms with no URN • 12. Algorithms in the context of Advanced Electronic Signatures • 12.1 Time period resistance of hash functions and keys • 12.1.1 Time period resistance for hash functions • 12.1.2 Time period resistance for signer’s key • 12.1.3 Time period resistance for root keys • 12.1.4 Time period resistance for other keys • 12.2 Algorithms for the various data structures
Part 1: Hash functions and asymmetric algorithms • Annex A (normative): Updating algorithms and parameters • A.1 Introduction • A.2 Maintenance Process • Annex B (informative): Recommended key sizes (historical) • Annex C (informative): Generation of RSA keys for signatures • C.1 Generation of random prime numbers • C.2 Generation of RSA modulus • C.3 Generation of RSA keys • Annex D (informative): Generation of elliptic curve domain parameters • D.1 ECDSA and ECGDSA based on a group E(Fp) • D.2 ECDSA and ECGDSA based on a group E(F2m)
Part 1: Hash functions and asymmetric algorithms • Annex E (informative): On the generation of random data • E.1 Classes of random number generators • E.2 On tests for NRNGs • Annex F (informative) Algorithms identifiers defined in various documents • Annex G (informative): Explanatory text about the “liberal view” and the “conservative view • G.1 Estimates based on past experience • G.2 Estimates based on power of computation • G.3 Recommended key sizes and use dates drawn from past estimates • G.4 Recommended key sizes and use dates drawn from Lenstra and Verheul’s table
Part 2: Symmetric algorithms and protocols for secure channels • Secure messaging for smart cards • 5.1 General • 5.2 Channel keys establishment • 5.2.1 Authentication steps • 5.2.2 Session Key creation • 5.2.3 Compute channel key • 5.2.4. Compute send sequence counter SSC • 5.3 Secure Messaging Mode • 5.3.1. CLA byte • 5.3.2. TLV coding of command and response message • 5.3.3. Treatment of SM-Errors • 5.3.4. Padding for checksum calculation
Main points of discussion from SAGE meeting (1/2) • The criteria we mention (secure/commonly used/easily referenced) was generally agreed to be OK. • There was some concern about the Whirlpool algorithm not being well studied enough. • There was a question about why we have not included SHA-512. • There were some concerns raised about whether the German elliptic curve variants were well enough studied. They had been “approved” by the BSI on some level. • There were several suggestions regarding the key lengths. The most significant is that 640 be changed to 768 for the 3 year predictions in Table 8, and the inclusion of 163 for ecdsa in the bottom two rows of the same table.
Main points of discussion from SAGE meeting (2/2) • The insertion of a statement as to why we haven’t distinguished between the use of algorithms in various contexts. • About Annex G: some editing out of the original text has been done in this annex, taking out some of the argumentation as to why we prefer the Lenstra/Verheul arguments to the extrapolation methods. The point was raised that this original argumentation should either be included in its entirety, completely left out or replaced with a much shorter explanatory text.
Question 1 from Helmut Biely • The title of chapter 11 of -1 is "Practical ways to identify...". Is this the only purpose of this chapter of the ensuing chapters on OID's ? • The real question is whether such OID's are there : • for interoperability purposes only, or/and • to identify the algos, that are considered as sufficiently secure by this TS.
Question 2 from Helmut Biely • The OID for ECDSA in 11.1.2 : • RFC 3278 is used (and not 3279 !) as reference. • However, RFC 3278 (applying to CMS only) sets the parameters part of the ASN.1 SEQUENCE explicitly to NULL, i.e no curve is defined, as is necessary for a complete OID e.g. in an X509 certificate. • Now, does this mean, that all curves defined in X9.62 are covered by the TS, whereas other curves are outside ? • Are there any intentions to differentiate between curves (e.g. for interoperability reasons) or do no such plans exist ?
Questions from Franco Ruggieri • The questions are addressed in the original document.