Re: [tcpm] New version of TCP Option for Transparent Middlebox Discovery available
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [tcpm] New version of TCP Option for Transparent Middlebox Discovery available



Melinda Shore wrote:
On Sep 28, 2009, at 11:05 AM, Andrew Knutsen wrote:
The point was made off-list that "responding and forwarding" doesn't work, since the response is on the SYN-AK. However, the action defined for the device-type could be to send a UDP datagram back to the originator, and forward the SYN (with option intact) (unless you're the destination).

That's going to fail if one of the middleboxes is a NAT.


True in many cases... Perhaps to implement the trace with this option, from a non-static source IP, a static destination IP would have to be included in the target data. It seems to me this might be reasonable, since the use case I see for this trace is management, where you'd want the response datagrams to go to a management station which would follow up with SNMP or some such.

Note that this use is not the motivation for this option, even though the name may seem to imply it. If it works that great though.

Andrew


Note: Messages sent to this list are the opinions of the senders and do not imply endorsement by the IETF.