One-way Active Measurement Protocol (OWAMP) Requirements
draft-ietf-ippm-owdp-reqs-06

[Note]: 'Returning document.  Important as guides main protocol of WG.  Now has strong mandatory to implement security and supports timestamp encryption - on agenda  for SMB to check against his pseudo-discuss.
' has been cleared by Allison Mankin

[Note]: 'Returning document.  Important as guides main protocol of WG.  Now has strong mandatory to implement security and supports timestamp encryption - on agenda  for SMB to check against his pseudo-discuss.
' added by Allison Mankin

Pseudo-Discuss by SMB from 2002-Oct-03 (since Secretariat did not enter):

  What are the effects on the measurement if an unencrypted (and,
  presumably, unauthenticated) timestamp is modified?  They seem to
  have a strong preference for leaving them in the clear.  It may not
  be a solvable problem, given their other constraints, but they should
  document the tradeoff.

  Still looks vulnerable to replays.  They mention the word "replay",
  but say nothing about what to do.  Timestamps are one traditional
  mechanism -- but not if they're unprotected...

  For the replay issue, what I'm asking for is a section analogous
  to 6.3, 6.4, etc., explaining the problem and giving the constraints
  on possible solutions (such as not using unprotected timestamps).