Re: [IPsec] AES-GCM IV length
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [IPsec] AES-GCM IV length



 


From: ipsec-bounces at ietf.org [mailto:ipsec-bounces at ietf.org] On Behalf Of Scott C Moonen
Sent: Thursday, August 13, 2009 2:09 PM
To: ipsec at ietf.org
Subject: [IPsec] AES-GCM IV length


RFC 4106 says:

   The AES-GCM-ESP IV field MUST be eight octets.

NIST publication 800-38D says:

  For IVs, it is recommended that implementations restrict support to
  the length of 96 bits, to promote interoperability, efficiency, and
  simplicity of design. 
See section 4 of RFC 4106: there's also a 4 octet 'salt' which is negotiated (and fixed for a particular SA); the nonce (IV) that is passed to the underlying GCM primitive is made of the of the 4 octet salt concatenated with the 8 byte IV from the packet.  This concatenated nonce is 96 bits in length, matching the above guideline...
 

There are no errata for RFC 4106, so I assume that ESP with ENCR-AES_GCM_nn uses an 8-byte IV.  Unfortunately, this goes against the NIST recommendation and also prevents the use of the RBG-based IV construction method outlined in the NIST document (which requires a minimum IV length of 96 bits).

Does anyone have any observations or comments on this?  Is it correct that existing ESP AES_GCM implementations are using 128-bit IVs? 
If they are, they are not following RFC 4106...
 

Thanks,


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

Note: Messages sent to this list are the opinions of the senders and do not imply endorsement by the IETF.