Re: Protocol Definition

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 



Your are correct. In fact, from the beginning the ARPANET and CYCLADES groups saw this as a distributed computing problem. We have often said that much of the reason that early effort was a success is that we were operating systems guys not telecom guys.

The early applications were all aimed at replicating OS functionality in the network. My favorite "proof" of this is that contrary to what many textbooks say, Telnet is not a remote login protocol; but a terminal device driver protocol. But there was much work prior to 1975 on the network as distributed computation. The problem was our ideas were ahead of what the hardware could do "cheaply." Even so, by 1975 we had fielded an application that implemented a land use planning system that used databases on both coasts invisibly to the user sitting at a plasma screen with touch. The keyboard was only for data entry. (There were other equally ambitious efforts) Somehow this direction was lost and we took a step backwards to a focus on "endpoints" as it has been characterized.)

You are also correct that strictly speaking the words "protocol" and "algorithm" are probably the same. Other fields, such as biology, use protocol to mean the list of steps to produce something.

There has always been a debate about the abstract definition of "process." (To punt the definition I have thought of it as "a locus of execution" or "the instantiation of a program" but of course those beg other questions.) ;-) Which is why OSI gave up and adopted "entity" as about the only word they could find that didn't have implications about how it was implemented.

Take care,
John

At 9:03 +0100 2012/01/08, t.petch wrote:
----- Original Message -----
From: "Dave CROCKER" <dhc@xxxxxxxxxxxx>
To: "Joel M. Halpern" <jmh@xxxxxxxxxxxxxxx>
Cc: "IETF-Discussion" <ietf@xxxxxxxx>
Sent: Friday, January 06, 2012 4:17 AM

 On 1/5/2012 7:10 PM, Joel M. Halpern wrote:
 > I suspect that the "correct" choices depends upon how you look at the
analogy.
 > What seemed to me the closest analog to "process" would be the actual
messages
 > on the wires.

Nah. A message on the wire is a single unit in an activity. And taken on its
 own, in the host or on the wire, it's actually static.

It isn't the activity. A process is an activity. The challenge is a term for
 the /flow/ of messages.

 It would be nice if it were a single word.

I agree that a message is not the right word, but I think that protocol is:-)
'Protocol' started as the draft treaty that formed part of diplomatic exchanges,
ie it was the physical manifestation, not the abstract concept, so I would use
it in that sense for networking.

For the abstract side of networking, I would use the same terminology as I would
use for a 'program'.  After all, a network is just a single, multi-tasking
system in which the 'links' that tie together the multiple tasks have been
stretched a little and made manifest so I use the same constructs, the same
tools - eg state machines - for both. In a multi-tasking operating system, you
will have post and wait and some such, in a network you have send and receive
and some such, same difference.

Tom Petch


 d/
 --

    Dave Crocker
    Brandenburg InternetWorking
    bbiw.net
 _______________________________________________
 Ietf mailing list
 Ietf@xxxxxxxx
 https://www.ietf.org/mailman/listinfo/ietf



_______________________________________________
Ietf mailing list
Ietf@xxxxxxxx
https://www.ietf.org/mailman/listinfo/ietf

_______________________________________________
Ietf mailing list
Ietf@xxxxxxxx
https://www.ietf.org/mailman/listinfo/ietf


[Index of Archives]     [IETF Annoucements]     [IETF]     [IP Storage]     [Yosemite News]     [Linux SCTP]     [Linux Newbies]     [Fedora Users]