User-agent: Mozilla Thunderbird 1.0 (X11/20041208)
Brian E Carpenter wrote:
Well, I find it hard to see how a single-ended shim can help, unless
it actually turns out to be an in-host NAT, and I don't think we want
to go there, do we?
I'm not sure that this is the only possible result.
Even in the two-shim case, we need a function which will trying
different ULIDs pair if there is a failure during initial contact.
One way to do this is to place that module above the transport
protocols. If we have such a module, it could be used in the
single-ended shim case as well.