[ih] Flow Control in IP

[John Day]

First of all, great story!!!  ;-)  Sounds like a number of things outside the rules that went on. It was a great time!!

> On Apr 30, 2024, at 14:50, Karl Auerbach via Internet-history <internet-history at elists.isoc.org> wrote:
> 
> 
> In our early security work at SDC (circa 1972-4) we were trying to interpose an end-to-end encryption into ARPAnet protocols.  One aspect kept getting our technical goats - RFNM - Request For Next Message.  The reason was that this message was generated by elements (destination IMPS) below our encryption but needed to be delivered to elements above our encryption.
> 
> (I know that RFNM was supposed to from destination IMP to source IMP, but for reasons that I no longer remember, it ended up crossing our security/cryptography barrier.)

I don’t know, but let me suggest what it was and you can tell me if that is it. About this time there was something called ‘receipt confirmation.’  The idea was that receiving acks should notify the sender.  CCITT was especially big on it.  It was a bone of contention, some thought it was a good idea, some didn’t. Those who liked more determinism liked it. The argument for Transport Protocols was that it was IPC and IPC didn’t give the user an ack, so the Transport protocol shouldn’t. The acks were entirely inside the layer.

So was this receipt confirmation?

Take care,
John

> 
> RFNM became a pejorative, or rather, was the subject of many pejorative outbursts by our group (Dave Kaufman, Frank Heinrich, Marv Schaeffer, Jerry Cole, Carl Switzky, Jerry Simon, Josie Althous, Val Schorre, John Scheid, Hillary X <can never remember her last name>, Jay Egglestun, Dave Golber, and myself.)
> 
> We had to engineer a trusted (mathematical specifications of security, formal verification of code against that security spec) hardware/software bypass around our cryptographic layer (much of which was in very expensive Tempest grade hardware) to deal this this.
> 
> (As our work progressed, TCP came along and we moved our work over to that approach, using the evolving split of an IP-like layer from the bottom of TCP, as a wedge into which to insert our security protocols. This was a much better design for our purposes when measured by Wirth's definitions of modularity (minimal information flow between modules). Our designs got easier and less Rube Goldberg - except that along the way we had begun to use much more complex modes of encryption (we'd call it blockchain today) and key management.
> 
> Apropos Hamlet's line that “There are more things in Heaven and Earth, Horatio, than are dreamt of in your philosophy.” the Internet grew from a soil rich with pranks, strange events, and  not a little romance. Such as this:
> 
> As that happened our network security work got moved behind multiple layers of guards and locked doors (and special RF containment rooms). That had the effect of isolating us from company management who lacked the clearances to come into our area.  It was at that time what some of us wanted to make our offices nicer - in violation of SDCs rather strict organizational hierarchy.  So one evening we (Carl Switzky and myself) found a large spool of rather nice, essentially new, white wool carpet that was being discarded by a super high end shop on San Vicente and I had an International Harvester truck large enough to carry that spool. It was also at that time we discovered that while the SDC guards had instructions not to allow things to be carried out from the buildings that they had no instructions about carrying things in.  And through a strange coincidence of the dark forces of the universe one of our group was working late and also had a carpet knife attached to his belt.  So the next day we all had really nice white wool carpet in our offices, inside the high security zone (we did all the offices in order to create plausible deniability about our role.)
> 
> 	--karl--
> -- 
> Internet-history mailing list
> Internet-history at elists.isoc.org
> https://elists.isoc.org/mailman/listinfo/internet-history