Label Switched Path (LSP) and Pseudowire (PW) Ping/Trace over MPLS Networks Using Entropy Labels (ELs)
draft-ietf-mpls-entropy-lsp-ping-05

Just a nit:  It would be nice to move Section 10 towards the front of the document.  I know there are a couple of references to it, but knowing upfront the scope would be helpful to any reader.

As noted by Scott Bradner in his OPS DIR review, 2 issues worth addressing IMO.

I did an OPS-DIR review of Label Switched Path (LSP) and Pseudowire (PW) Ping/Trace over
 MPLS Network using Entropy Labels (EL) (draft-ietf-mpls-entropy-lsp-ping-04)

The draft extends the existing MPLS LSP Ping and Traceroute multipath mechanisms to
 support LSPs that use an Entropy Label.

The primary operational impact of this technology is to provide an additional tool for network
 operators to debug their networks - a good thing.

I found the draft a bit hard to follow, it seems to be more a collection of data points than a
 clear narrative but I do not think it is worth a rewrite to make it easier to understand.

I found one thing that raises an operational concern - the next to last paragraph in 
section 2 says: “All LSRs along the LSP need to be able to understand the new flags 
and the new multipath information type.” But I do not see a mechanism discussed to check to 
see if that is the case  (like the high order two bits of IPv6 options).  If there is a 
mechanism it might be good to describe it, if there is not, a statement that 
verifying this condition is outside of the scope of the draft would be helpful

The same paragraph goes on to say  “It is also required that the initiating 
LSR can select both the IP destination address and label to use when 
transmitting MPLS echo request packets.” It might be helpful to say under 
what conditions this is or is not the case.

Otherwise, the draft seems ready for publication.

Editorial observations from Peter (in his Gen-ART review) are worth noting before final sending of this draft to the RFC Editor.

Thank you for the updated text in your draft version to address my prior discuss questions.

As Scott mentioned in his review, I also think that the following sentence needs more discussion:

"All LSRs along the LSP need to be able to understand the new flags
   and the new multipath information type."

How can you ever know if this is true? Isn't the whole problem that you don't know what different LSRs implement?

Just to note that there's an editorial change that was
agreed [1] as a result of the secdir review that has still
to be made. (No big deal, just noting it in case it
gets forgotten.)

  [1] https://www.ietf.org/mail-archive/web/secdir/current/msg06754.html