Key Management Considerations for the TCP MD5 Signature Option
draft-ietf-idr-md5-keys-00

In author's 48 hours:

From: RFC Editor
Subject: authors 48 hours: RFC 3563
        NOW AVAILABLE
Date: Mon, 30 Jun 2003 15:02:30 -0700
To: mleech@nortelnetworks.com
Cc: RFC Editor , Bill Fenner
    , Alex Zinin ,
    skh@nexthop.com, yakov@juniper.net

RFC EDITOR:

Please change the title to "Key Management Considerations for the
TCP MD5 Signature Option".

Please change the following:

In section 3, the first bullet:
OLD:
      o Key lengths SHOULD be between 12 and 24 bytes, with larger keys
        having effectively zero cost when compared to shorter keys.

NEW:
      o Key lengths SHOULD be between 12 and 24 bytes, with larger keys
        having effectively zero additional computational cost when
                            ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
        compared to shorter keys.

In section 5, first paragraph:

OLD:
  this option may have lifetimes on the order of months.  It would seem
  prudent, then, to choose a *minimum* key length that guarantees that
  key-guessing runtimes are some reasonable [3-5??] multiple of the
  key-change interval under best-case (for the attacker) practical

NEW:
  this option may have lifetimes on the order of months.  It would seem
  prudent, then, to choose a minimum key length that guarantees that
                              ^^^^^^^ (remove emphasis)
  key-guessing runtimes are some small multiple of the key-change
                                  ^^^^^^^^^^^^^^
  interval under best-case (for the attacker) practical

In section 6, first paragraph:

OLD:
  that the reasonable upper-bound for software-based attack performance
  is 1.0e13 MD5 operations per second, then the *minimum* required key
  entropy is approximately 68 bits.  It is reasonable to round this

NEW:
  that the reasonable upper-bound for software-based attack performance
  is 1.0e13 MD5 operations per second, then the minimum required key
                                                ^^^^^^^ (remove emphasis)
  entropy is approximately 68 bits.  It is reasonable to round this

Proposed RFC-Editor note follows.  I've sent this to the author for a quick check.

RFC EDITOR:

Please change the title to "Key Management Considerations for the
TCP MD5 Signature Option".

Please change the following:

In section 3, the first bullet:
OLD:
      o Key lengths SHOULD be between 12 and 24 bytes, with larger keys
        having effectively zero cost when compared to shorter keys.

NEW:
      o Key lengths SHOULD be between 12 and 24 bytes, with larger keys
        having effectively zero additional computational cost when
                            ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
        compared to shorter keys.

In section 5, first paragraph:

OLD:
  this option may have lifetimes on the order of months.  It would seem
  prudent, then, to choose a *minimum* key length that guarantees that
  key-guessing runtimes are some reasonable [3-5??] multiple of the
  key-change interval under best-case (for the attacker) practical

NEW:
  this option may have lifetimes on the order of months.  It would seem
  prudent, then, to choose a minimum key length that guarantees that
                              ^^^^^^^ (remove emphasis)
  key-guessing runtimes are some small multiple of the key-change
                                  ^^^^^^^^^^^^^^
  interval under best-case (for the attacker) practical

In section 6, first paragraph:

OLD:
  that the reasonable upper-bound for software-based attack performance
  is 1.0e13 MD5 operations per second, then the *minimum* required key
  entropy is approximately 68 bits.  It is reasonable to round this

NEW:
  that the reasonable upper-bound for software-based attack performance
  is 1.0e13 MD5 operations per second, then the minimum required key
                                                ^^^^^^^ (remove emphasis)
  entropy is approximately 68 bits.  It is reasonable to round this

State Changes to Reading List                                      from Pre AD Evaluation                                by fenner