RE: pppcn

> Exactly what modern application is usable on a link with effective RTTs
> of 0.75 seconds?

A properly engineered TCP should be usable for bulk data transport even at
these RTTs.  Streaming media can also work fine with appropriate
buffering.  Interactive applications will be intolerable.

> Exactly what application can tolerate packet loss rates so
> high that 3 RTTs turn into 7 on average?

Agree with you here.  TCP won't function at these loss rates, and neither
will interactive applications.  I'd claim that any link layer that
produces these loss rates is improperly engineered -- it requires link
layer reliability improvement via FEC, etc.

>It is for both peers to anticipate each others' sequence
>and magjic numbers and options, and to send a single burst of LCP,
>authentication, and IPCP packets.  This solution requires no changes
>in PPP state machines except for what might be called "pre-loading.

How is it possible to "anticipate" authentication protocols?  If this were
true, it would imply that the authentication algorithms were broken since
they are required to demonstrate pseudo-randomness.