Hi all,
Please see my comments between [Danping----->>]...[<<-----Danping].

-----Original Message-----
From: Sam Hartman [mailto:[email protected]] 
Sent: Wednesday, August 27, 2014 4:26 AM
To: Warren Kumari
Cc: [email protected]; [email protected]; 
[email protected]
Subject: Re: Call for Adoption: draft-hmac-sha-2-usm-snmp


Hi.

I've reviewed both draft-hartman-snmp-sha2 and draft-hmac-sha-2-usm-snmp.

In general, I believe that draft-hartman-snmp-sha2 provides a better starting 
point for a SHA2 authentication algorithm for USM.

draft-hmac-sha-2-usm-snmp provides differences between the MD5 algorithm 
described in RFC 3414 and the SHa2 algorithm.

In general I've found that  describing a specification in terms of differences 
in cases where there is not a clear abstraction present is problematic.  It 
tends to lead to implementation errors by focusing on the steps that are 
different rather than on all the steps that need to be performed.

In this case, I think that we'll tend to see key derivation and to a lesser 
extent messaging processing errors because of the brief text pointing at these 
issues.

In our draft (draft-hartman), we create an abstraction for an HMAC-based hash 
authentication in USM and plug sha-2 into that abstraction. I think that's a 
cleaner approach that will lead to higher implementation quality than the 
draft-hmac-sha-2-usm-snmp approach.

[Danping----->>] 
Agree. The complexity of implementing the 'draft-hmac' would tend to result in 
confusions and missing steps related to the key derivation etc. 'draft-hartman' 
has a better explanation of how the protocol fits into USM and SNMP, as well as 
a clearer explanation of how the overall system works. It seems that this 
document is more likely to provide good long-term implementation results.
[<<-----Danping]

In addition, I'm not convinced that truncating the HMAC is a good idea in this 
instance.  If the WG decides that truncation of the HMAC is desirable, we 
should add a description of why that's the case and a security discussion.  (I 
don't think the truncation proposed has significant security problems)

[Danping----->>] 
Agree. Although truncation is fairly well analyzed, there are security 
disadvantages to do it, and I don't really see significant advantages to 
truncation. I've searched the mailing list about comments on 'draft-hmac', 
several members who contributed to improve it, still had doubts and confusions 
on the truncation to various output lengths. 
[<<-----Danping]

My recommendation would be that the WG start with
draft-hartman-snmp-sha2 as a basis for this work, but that the authors of 
draft-hmac and draft-hartman work together to make sure that all the best ideas 
from both proposals make their way into the final product.

[Danping----->>]
I support 'draft-hartman' as a basis and a proper direction towards enhancing 
USM of SNMP.
[<<-----Danping]

_______________________________________________
OPSAWG mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/opsawg

Reply via email to