On Oct 9, 2008, at 5:25 PM, Paul Kyzivat wrote:
Ben,
Miki is hunting up something that I guess accidentally didn't get
sent to you.
Meanwhle:
Ben Campbell wrote:
Section 1.1, paragraph 2:
You say that these use cases are not specific to the transport
protocol. Should I infer from that that all of these cases are the
same for TCP as for UDP? I am skeptical that this is true for
cases that indicate a dropped packet.
Certainly *reordered* messages can occur with TCP if there are any
intermediate nodes.
I think perhaps a single message could also be *lost* with TCP on
every hop and an intermediate proxy if the proxy crashed and
restarted.
Certainly a message can be lost if the UAs have TCP connections to
proxies, and there is at least one UDP connection between the proxies.
Of course proxies are typically not shown, so that is hard to show
in detail, and would needlessly complicate things.
But I certainly agree that there might be a better way to to explain
this. Would it be sufficient to beef up the above paragraph to
explain that we don't show proxies, but assume they might be
present, and that in the presence of proxies there is always the
possibility of rearranging or losing messages for the above reason?
Maybe that general principle is a good one to point out, since its
another one of those things that is periodically brought up again.
I agree that this would help, and would address my comment.
Thanks!
Ben.
_______________________________________________
Gen-art mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/gen-art
