ipsec October 2009 archive
Main Archive Page > Month Archives  > ipsec archives
ipsec: Re: [IPsec] [ipsecme] #112: Truncation of SHA-1 ICVs

Re: [IPsec] [ipsecme] #112: Truncation of SHA-1 ICVs

From: Frankel, Sheila E. <sheila.frankel_at_nospam>
Date: Tue Oct 27 2009 - 17:30:45 GMT
To: Scott C Moonen <smoonen@us.ibm.com>, "ipsec@ietf.org" <ipsec@ietf.org>

Thanks, Scott. So is the general consensus that we should just leave HMAC-SHA-1 and HMAC-MD5 as the only algs for which IKEv2 can negotiate either a truncated or non-truncated version?

Your comment also reminded me that RFCs 2404 (HMAC-SHA-1) and 2403 (HMAC-MD5) require truncated ICVs for IPsec. So I guess I should change the new text to only allow IKEv2 to use both versions for its own SAs, but not for IPsec SAs.


From: Scott C Moonen [mailto:smoonen@us.ibm.com] Sent: Tuesday, October 27, 2009 12:17 PM To: Frankel, Sheila E.
Cc: ipsec@ietf.org; ipsec-bounces@ietf.org; Tero Kivinen; Paul Hoffman; suresh.krishnan@ericsson.com Subject: Re: [IPsec] [ipsecme] #112: Truncation of SHA-1 ICVs

Hi Sheila,

  1. I don't think we can expand the registry to include non-truncated versions of HMAC-SHA2-*. RFC 4868 stipulates for IKE and IPsec in general that the authenticator length "is always half the output length of the underlying hash algorithm."
  2. RFCs 3566, 4494 are worded a bit more permissively for AES-XCBC and AES-CMAC, so perhaps there's some wiggle room there.
  3. I'm not sure if HMAC-RIPEMD is defined for use in IKE (there is not even an algorithm identifier for IKEv2), but its use for AH and ESP (RFC 2857) currently only defines a truncated form of the algorithm.

Scott Moonen (smoonen@us.ibm.com)
z/OS Communications Server TCP/IP Development http://www.linkedin.com/in/smoonen


"Frankel, Sheila E." <sheila.frankel@nist.gov>


"ipsec@ietf.org" <ipsec@ietf.org>


Tero Kivinen <kivinen@iki.fi>, Paul Hoffman <paul.hoffman@vpnc.org>, "suresh.krishnan@ericsson.com" <suresh.krishnan@ericsson.com>


10/27/2009 11:46 AM


Re: [IPsec] [ipsecme] #112: Truncation of SHA-1 ICVs

#112: Truncation of SHA-1 ICVs

Proposed change to Roadmap doc:

Add text to Section 5.3 (Integrity-Protection Algorithms)

Current text:
  The integrity-protection algorithm RFCs describe how to use these   algorithms to authenticate IKE and/or IPsec traffic, providing   integrity protection to the traffic. This protection is provided by   computing an Integrity Check Value (ICV), which is sent in the   packet. The RFCs describe any special constraints, requirements, or   changes to packet format appropriate for the specific algorithm. In   general, they do not describe the detailed algorithmic computations;   the reference section of each RFC includes pointers to documents that   define the inner workings of the algorithm. Some of the RFCs include   sample test data, to enable implementors to compare their results   with standardized output.

Additional text:
  Some of these algorithms generate a fixed-length ICV, which is truncated   when it is included in an IPsec-protected packet. For example, standard   HMAC-SHA-1 generates a 160-bit ICV, which is truncated to 96 bits when it   is used to provide integrity-protection to an ESP or AH packet. The   individual RFC descriptions mention those algorithms that are truncated.   When these algorithms are used to protect IKEv1 SAs, they are not   truncated. For HMAC-SHA-1 and HMAC-MD5, the IKEv2 IANA registry contains   values for both the truncated version and the standard non-truncated   version; thus, IKEv2 has the capability to negotiate either version to   protect IKEv2 and/or IPsec-v3 SAs. For the other algorithms (AES-XCBC,   HMAC-SHA-256/384/512, AES-CMAC and HMAC-RIPEMD), only the truncated   version can be used for both IKEv2 and IPsec-v3 SAs.

NOTE to Tero, Paul, Yaron: do we want to expand the IKEv2 IANA registry to include non-truncated AES-XCBC-MAC, HMAC-SHA-256/384/512, AES-CMAC and HMAC-RIPEMD?

From: ipsecme issue tracker [trac@tools.ietf.org] Sent: Friday, October 16, 2009 8:25 PM
To: paul.hoffman@vpnc.org; Frankel, Sheila E. Subject: [ipsecme] #112: Truncation of SHA-1 ICVs

#112: Truncation of SHA-1 ICVs -----------------------------------+---------------------------------------- Reporter: paul.hoffman@... | Owner: sheila.frankel@... Type: defect | Status: new Priority: normal | Milestone: Component: roadmap | Severity: - Keywords: | -----------------------------------+---------------------------------------- In RFC 2404, it mentions that SHA-1 ICVs are truncated to 96 bits for IPsec. We should also mention in Section 5.3 that this truncation is done for IKEv2 as well. Same for RFC 2403. Text is needed. -- Ticket URL: <http://trac.tools.ietf.org/wg/ipsecme/trac/ticket/112> ipsecme <http://tools.ietf.org/ipsecme/> _______________________________________________ IPsec mailing list IPsec@ietf.org https://www.ietf.org/mailman/listinfo/ipsec

_______________________________________________ IPsec mailing list IPsec@ietf.org https://www.ietf.org/mailman/listinfo/ipsec