[ih] Robustness Principle and End-to-End Principle

[Jack Haverty]

The End-to-End principle has been an important part of military 
communications -- even back to Greco-Roman times.   The Internet and 
Arpanet were driven by the US DoD.   I suggest looking into military 
history for references.   See also:

https://www.wearethemighty.com/history/cryptology-in-the-military/
and
https://www.weforum.org/agenda/2019/02/a-brief-history-of-cryptography-and-why-it-matters/

Robustness of course also has mattered a lot in military scenarios for 
centuries.  Roman generals would have slaves (aka scribes) replicate an 
important message and then send each copy by alternate routes, e.g,. 
some by sea and some by land, increasing the odds that the message would 
reach its destination.

When one of your switching nodes gets blown up, the command and control 
still has to go through somehow.   When one gets captured by an enemy, 
it's important to limit what that enemy can do.

Such scenarios were certainly considered as TCP/IP details were being 
worked out since the customer was the military.

Jack


On 4/15/24 16:11, Barbara Denny via Internet-history wrote:
> I am looking into the history of the Robustness Principle and the End-to-End Principle.  I have found 2 significant dates.  One is the mention of the Robustness Principle in RFC 761(DoD TCP Spec) dated 1980, and the paper dated 1981 by Jerry Seltzer, David Reed, and Dave Clark on the End-to-End Principle.  I am trying to go earlier than those dates.  These principles guided the work in the 70s so I am looking for information about that time period (or earlier?) and these principles.  Any thoughts, recollections, or references?, including dates/timeframes, greatly appreciated.
> barbara

-------------- next part --------------
A non-text attachment was scrubbed...
Name: OpenPGP_signature.asc
Type: application/pgp-signature
Size: 665 bytes
Desc: OpenPGP digital signature
URL: <http://elists.isoc.org/pipermail/internet-history/attachments/20240415/cdc4b3a9/attachment.asc>