[ih] mssage vs. packet (was: Re: Early Internet history)

[Vint Cerf]

The basic methods of queueing theory apply to both message and packet
switching.
In the packet case, segmentation allows the message to be in flight
concurrently over multiple hops so that reduces the total delay since there
is, as you say, overlapping transmission along multiple hops of the path.Of
course a one hop path produces no advantage since all the segments have to
transit the one hop. TCP segmented the entire transmission partly based on
flow control signals. For a time, TCP Segments might have been broken into
smaller pieces in, e.g., the ARPANET. Later transit over Frame Relay and
ATM also potentially broke things up and permitted similar overlaps. I was
only making the point that the same mathematical models work for both cases.

v


On Sat, Jul 7, 2018 at 7:15 PM, Dave Crocker <dhc2 at dcrocker.net> wrote:

>
> > Kleinrock's analysis was for message switching but the mathematics of
> > message switching and packet switching are essentially comparable
> > especially when you consider variable length messages.
>
>
> Howdy.
>
> A question to the group...
>
> That distinction coincidentally surfaced in a discussion a couple of
> months ago, after some decades of my not hearing it.
>
> I know what it meant on the Arpanet.  And I know what wikipedia and some
> other entries say about it.  But while the ability to handle smaller
> chunks independently -- and even in an overlapping manner -- encourages
> some useful performance improvements, I find myself generally thinking
> of them as the same category of communications technology.
>
> Namely:  Discrete segments of data being handled through a network.
> Certainly for some form of multiplexing and possibly with dynamic
> routing.  (These days, we'd take stat mux and dynamic routing as
> inherent, but my recollection is that 45 years ago, those were
> variations being played with.)
>
> I'm not looking to re-start the religious wars on the distinction but am
> curious whether, from the perspective of those 45 years and global
> scaling, it is fair to have most discussions -- I emphasize most, not
> all -- treat them as the same construct?
>
> If not, why not?
>
>
> d/
> --
> Dave Crocker
> Brandenburg InternetWorking
> bbiw.net
> _______
> internet-history mailing list
> internet-history at postel.org
> http://mailman.postel.org/mailman/listinfo/internet-history
> Contact list-owner at postel.org for assistance.
>



-- 
New postal address:
Google
1875 Explorer Street, 10th Floor
Reston, VA 20190
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://elists.isoc.org/pipermail/internet-history/attachments/20180707/75158450/attachment.htm>