I have reviewed this document as part of the security directorate's ongoing effort to review all IETF documents being processed by the IESG. These comments were written primarily for the benefit of the security area directors. Document editors and WG chairs should treat these comments just like any other last call comments. This document deprecates 3DES and RC4 in Kerberos. This should have been done several years ago, so the only problem I have with this document is that it has year 2017 not 2012 in header :-) This could go even further as it just marks those weak algorithms as "SHOULD NOT", and especially the RC4 with unsalted MD4 based string2key should be marked as "MUST NOT" at least for my opinion. The document properly explain the reasons why those algorithms are still there to maintaina backward compatibility with Windows XP and similars, but as Windows XP is already end of lifed, so I think those algorithms could also be marked as MUST NOT. This document also gives recommendations both for Kerberos implementations and deployments. Normally we do not give instructions for the adminstrators, we just tell what implementors SHOULD or SHOULD NOT do. If we give recommendations to deployments, then I think those should be MUST NOT instead of SHOULD NOT. I can still see some implementations wanting to implement those algorithms to allow backwards compatibility (i.e., go against SHOULD NOT), but no new deployment should use them ever, and old deployments needs to move away from them ASAP. Anyways as a summary I think this document is Ready. -- kivinen@iki.fi