Re: [secdir] Secdir review of draft-ietf-mpls-ldp-gtsm-07

[Brian Weis <[e-mail]>]

Hi Rajiv,

Your proposed text looks good to me. I think it's good to go with this added.

Thanks,
Brian


On Jun 1, 2012, at 1:37 PM, Rajiv Asati (rajiva) wrote:

> Hi Brian,
> 
> Really appreciate your critical review and suggestions. 
> 
> I agree to your both of your suggestions, and would propose the
> following text for us to include in the next revision.
> 
> 
> //
> As discussed in section 3, it is possible that 
> - GTSM for LDP may not always be enforced on a single-hop LDP peering
> session and may still be susceptible to forged/spoofed protocol packets,
> if the single-hop LDP peering session is set up using Extended
> Discovery. 
> - GTSM for LDP may cause LDP peering session to not get established (or
> torn down), if IP routing ever declares that the directly connected peer
> is more than one hop away.
> Suffice to say, use of cryptographic integrity (e.g., RFC 5925) is
> recommended as an alternate solution for detecting forged protocol
> packets (especially for the multi-hop case).
> //
> 
> Cheers,
> Rajiv
> 
> 
>> -----Original Message-----
>> From: Brian Weis (bew)
>> Sent: Friday, June 01, 2012 2:16 PM
>> To: [e-mail]; The IESG
>> Cc: [e-mail]
>> Subject: Secdir review of draft-ietf-mpls-ldp-gtsm-07
>> 
>> 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 applies the Generalized TTL Security Mechanism (GTSM)
>> mechanism defined in RFC 5082. This mechanism is used by routing
>> protocols as a low-cost non-cryptographic method intended to frustrate
> off-
>> path attackers.  It is applicable when the peer is known to be
> connected by a
>> single hop.
>> 
>> The security considerations of this draft mostly point to RFC 5082's
>> extensive security considerations section, which is appropriate.
> However
>> because this I-D discusses multi-hop cases in greater detail it would
> be
>> appropriate for the security considerations section to also discuss
> multi-hop
>> a bit more. Here are some thoughts for that:
>> 
>> 1) Use of cryptographic integrity (e.g., RFC 5925) should be
> recommended as
>> an alternate solution for detecting forged protocol packets in the
> multi-hop
>> case.
>> 
>> 2) GTSM is expected to be enabled by default for Basic Discovery
> because
>> it's usually a single-hop, and disabled for Extended Discovery because
> it's
>> usually multi-hop. But then Section 3 mentions several exceptions,
> which
>> apparently need to be administratively configured away from the
> defaults.
>> Failing to do this when needed results in security risks in either
> case: either
>> GTSM isn't deployed when it should be and the router is inadvertently
> open
>> to spoofing, or GTSM is deployed when it shouldn't be and this results
> in an
>> availability issue because LDP packets will be dropped before reaching
> the
>> LDP peer. This should be stated in the Security Considerations.
>> 
>> Brian
>> 
>> 
>> 
>> 
>> 
> 


-- 
Brian Weis
Security Standards and Technology, SRTG, Cisco Systems
Telephone: [phone]
Email: [e-mail]