[e2e] Open the floodgate

[Cannara]

With all due respect to VJ (Vernon Jordan? :), the issue is not how well TCP
backs off, but (among other things) how poorly it functions in moving data
under a variety of loss causes, some of which require the opposite of backoff.

Of course, the fundamental issue is why any Transport protocol should be at
all concerned with Network-layer loading and congestion points.  That, after
all, is opposite to the concept of Transports providing reliable end-end
service, regardless of what the next layers down are troubled with.

The history is very relevant here, as folks have reminded me offlist.  ARPA's
scared over-reaction to continued, open protocol development in the '70s, for
example.  So the bureaucracy of TCP/IP mediocrity began at least there.  And
now seems maintained by a zombie frame of mind.  At least the Internet
protocols' development history provides some good examples for students of how
not to do things well.  In any case, most any Internet would be 'successful'
-- why?  Because, anything that supports satistfaction of the very human
appetites for sex and money always 'succeeds'.  Just ask any ISP manager what
traffic they know they can depend on.  :]

Alex

Saverio Mascolo wrote:
> 
> I think that we are still not completely aware of how well VJ designed TCP
> congestion control.
> 
> An indication of this is the number of new TCP congestion control proposals
> that never take off because they simply do not work. Reasons: they violate
> the self-cloking principle, they omit to consider time delays etc...
> 
> Saverio Mascolo
>