Hello,

The UNH-IOL would like to ask the Working Group for feedback regarding an issue 
we've observed.

This issue concerns how a Security Gateway handles IPv6 MTU restrictions and 
fragmentations.  Specifically, how should a SGW handle a received Packet Too 
Big message, for an ESP packet which it transmitted?

>From RFC 4301, Section 6.1.1, there are two options:

 "If an ICMP PMTU message passes the checks above and the system is
 configured to accept it, then there are two possibilities.  If the
 implementation applies fragmentation on the ciphertext side of the
 boundary, then the accepted PMTU information is passed to the
 forwarding module (outside of the IPsec implementation), which uses
 it to manage outbound packet fragmentation.  If the implementation is
 configured to effect plaintext side fragmentation, then the PMTU
 information is passed to the plaintext side and processed as
 described in Section 8.2."

The first option, applying fragmentation on the ciphertext side of the boundary 
seems to be optional, although it's not clear to us if it only applies to IPv4, 
according to RFC 4303, Section 3.3.4:
 "Thus, an ESP implementation MAY choose to not support fragmentation
 and may mark transmitted packets with the DF bit, to facilitate Path
 MTU (PMTU) discovery."

The second option is describe in RFC 4301, Section 8.2.1, which is to propagate 
the PMTU information via a synthesized Packet Too Big message.

So, there are two questions we would like to raise.

First, if ciphertext side fragmentation is indeed optional, and an IPv6 SGW 
implementation should choose to not support it, MUST it support generating the 
synthesized PTB message?

Second, the SGW can set the MTU to 1280 bytes, or less, in the synthesized 
Packet Too Big message, however, the originator is not required to reduce the 
size of fragments to less than 1280 bytes, but by adding the ESP header the 
resulting packets will be larger than 1280 bytes.  So, if the upstream MTU is 
1280 bytes, and an SGW implementation chooses to not to support ciphertext side 
fragmentation, what is the correct behavior?

Regards,
Timothy Carlin

----
Timothy Carlin
UNH-IOL
_______________________________________________
IPsec mailing list
IPsec@ietf.org
https://www.ietf.org/mailman/listinfo/ipsec

Reply via email to