NO. Name the single application which couldn't be run over lossy protocol.

Thats an easy one....

Any generic system that receives time sensitive real-time information from numerous distributed nodes that must be processsed and relayed to other processing stations *and* run the typical IP e-mail, ftp, httpd, etc. packet services concurrently over the same pipes, as well as network management traffic ad. infinitum.....

Real application, not theory, please? I've had some real experience with telemetry networks (did some harware design for my dad's Constructor Bureau For Tele-Mechanics And Automated Systems) and can say for sure that anybody who needs lossless telemetry with millisecond delay tolerance must be insane to do it over any public networks. 99.99% of real-life telemetry can live with losses just fine. In fact, telemetry stuff often designed to receive several measurements and do sanity check on them before anything actually happens. Practically all real-life sensors are prone to fluctuations (splashes for liquid level sensors, falling leaves for light brigtness sensors, ad infinitum).

Running real time sensory data, for example that must arrive at remote locations within a few milliseconds on a packet switched network, for example....

Real application, please? I never seen one, not in large controlling systems, not in military crap, not in scientific experiments (i've got more than few friends who did lots of data collection & networking stuff for high-energy physics research; in fact, my old e-mail address at Kurchatov Institute for Atomic Energy is still valid :) It is always either delay-tolerant in range of seconds, or it is "collect data now, analyze later" or it is dedicated lines because you can't afford delays in switches.

I'll attach a few references for those who don't wish nor have the time to pull down the 12 page postscript draft mentioned earlier...

It seems to me that a lot of people are very busy solving a non-existant problem. Anything new? --vadim