W3C home > Mailing lists > Public > ietf-http-wg-old@w3.org > September to December 1995

Re: Persistent connections: aborting requests in progress.

From: Jeffrey Mogul <mogul@pa.dec.com>
Date: Fri, 22 Sep 95 13:31:56 MDT
Message-Id: <9509222031.AA20205@acetes.pa.dec.com>
To: Shel Kaphan <sjk@amazon.com>
Cc: http-wg%cuckoo.hpl.hp.com@hplb.hpl.hp.com
    This is beginning to remind me of something else -- interleaved MPEG
    streams.  (and other related technology).  I detect a wheel beginning
    to be reinvented.  Soon (a year or so) I bet we'll be worrying about
    bandwidth requirements on the different interleaved media. Maybe it
    would be worth looking into work that has already been done.
    
I have already argued that we should be using protocols besides HTTP
to carry real-time, continuous-media streams.

    It is also not obvious to the casual observer that TCP will remain the
    protocol of choice when all is said and done with this sort of thing.
    The question is, how much can and should be wedged into it, and when
    should we start thinking about fundamental reengineering. 

TCP has its flaws.  T/TCP (for example) might be better.  But unless you
can come up with a plausible scenario about how a majority of the users
in the world will have working, reliable, interoperable T/TCP
implementations on their PCs in next few years, I think this is completely
pointless line of debate.

And I do not believe that a plausible scenario exists.

TCP is what we have; we need to work with it.

-Jeff
Received on Friday, 22 September 1995 14:08:23 EDT

This archive was generated by hypermail pre-2.1.9 : Wednesday, 24 September 2003 06:31:32 EDT