[e2e] Is a non-TCP solution dead?

[Cannara]

I'll simply add that saying: 'Wireless means there are "no wires".  Not even
"virtual wires"' is not quite correct.  The RF networks all depend on wired
links, because they do not, end-end, have global reach.  Even a satellite
link, as part of a cellphone call to Baghdad is effectively a wired link,
albeit with different error & other properties.  Even wireless PC access to
services will, at some point in a path, be via LAN or other wired segments.

And: "If an ideal TCP yields ALWAYS an improvement of 20 to 150% in
throughput, would it be a useful optimization?" is missing the point that if
this actually happened, usually the reverse does, then there's work to be done
at lower layers that must be done.  Otherwise, non-TCP traffic is improperly
penalized.
--
Alex