Re: secdir review of draft-ietf-mpls-lsp-ping-relay-reply

["lizho.jin@xxxxxxxxx" <lizho.jin@xxxxxxxxx>]

Hi Sean Turner

Thank you for the review. I am so sorry to reply after a long time. We need to update the document according to George's comments.

Now we have post v7, URL: http://www.ietf.org/internet-drafts/draft-ietf-mpls-lsp-ping-relay-reply-07.txt

The change lists are as below:

1. add a destination point in stack TLV to avoid potential loop.

2. add a source address in stack TLV, then the initiator could get the replying node address.

I also reply inline to your questions as below. Thank you.

---

Regards

Lizhong

 

From: Sean Turner

Date: 2014-12-14 03:29

To: draft-ietf-mpls-lsp-ping-relay-reply; secdir; The IESG; ietf

Subject: secdir review of draft-ietf-mpls-lsp-ping-relay-reply

Do not be alarmed.  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 with the intent

of improving security requirements and considerations in IETF drafts.

Comments not addressed in last call may be included in AD reviews

during the IESG review.  Document editors and WG chairs should treat

these comments just like any other last call comments.

 

Version: 06

Summary: Ready with some non-security/non-privacy nits.

 

Ping mechanisms always give me the heebie-jeebies [1]

because of the security concerns associated with them (i.e., DoS,

spoofing/hijacking/etc., and unauthorized disclosure).  This document

specifies an extension to the existing ECHO mechanism in RFC 4379

and it does nothing to address these concerns in fact it increases the

concerns wrt DoS. *BUT* it rightly points this increase exposure out in

the security considerations section.  It provides remediation techniques

similar to those specified in RFC 4379: rate limit and validate source

against access list.  This draft, unlike RFC 4379, does recommend that

operators wishing to not disclose their nodes blank the address out in

the TLV.  This draft also refers to RFC 4379 for additional security

considerations.

 

WARNING - questions and nits follow:

 

s3 - 1st paragraph: Relayed Echo Reply “replaces” Echo Reply - does this

mean you’re deprecating the use of “Echo Reply”?

[Lizhong] no. The reply message from any node to initiator will still be Echo Reply.

But the reply message from replying LSR to a relay node or from a relay node to 

another relay node will be Relayed Echo Reply.

 

s4.1: Is the outermost label allowed to be set to 255 to support the

“ping” mode or must it always be set to 1, 2, etc. to support “traceroute"

mode - as described in RFC 4379 s4.3?   I know s5 is just an example

but it really looks like this extension is just supposed to be for fault

isolation.

[Lizhong] It is possible to set 255. One implementation in my mind is to manually

generate the stack TLV list, and the replying LSR could still relay the message back

to the initiator according to the received stack TLV list. The case in the document is

for stack TLV list auto learning.

 

s4.1 - last paragraph: Does the next initiator put it’s address in the stack

before or after the previous initiator?  I assume it’s after, but I maybe

missed that part?  Would be good to state that explicitly.

[Lizhong] are you referring to section 4.2. There is some change in v7. Please help

to check again, to see if that is OK for you now.

 

Cheers,

spt

 

[1] http://en.wikipedia.org/wiki/Heebie-jeebies_(idiom)