[ih] Quantifying OSI

John Day jeanjour at comcast.net
Tue May 12 09:09:11 PDT 2026


Alex,
I should have pointed out that there was at least one exception to your last comment: A paper that tried to go the other way.

David D. Clark and David L. Tennenhouse titled "Architectural Considerations for a New Generation of Protocols,

This paper recasts many OSI concepts into the Internet. Tennenhouse had been a member of the OSI Naming and Addressing group I chaired. Many of the phrases common in those discussions are found in this paper. This is the paper where Application Layer Structure (ALS) is proposed, which of course is nothing more than an OSI Service-Data-Unit (SDU), a property of any layer.

Now a major feature of QUIC which dispenses with the stream-orientation of TCP.  Of course, SDU is what CYCLADES called a ‘letter’ in the CYCLADES Transport Protocol in 1972. It was clear then that that was what users of the protocol wanted, but not what Multics and later Unix provided.

I often wondered whether or not, Clark knew the paper was about OSI ideas. But it was a pretty amusing read.

Take care,
John

> On May 11, 2026, at 11:54, Alexander McKenzie via Internet-history <internet-history at elists.isoc.org> wrote:
> 
> Others from BBN who worked on the NIST contract to represent "the US
> Government's view" in OSI committees included John Burress (Transport
> Layer), e (Presentation Layer) and Kathy Huber (terminal support in the
> Application Layer, IIRC).  However the majority of our work was done in
> committees of ANSI, where the "US position" was developed.  ANSI meetings
> were roughly every 3 months. Between meetings we developed position papers
> based on input from NIST and from ARPAnet/Internet experience. As part of
> this project Tom Blumer developed a protocol compiler which took in a
> "formal" description of a protocol in a c-like language describing a
> protocol state machine and output a procedure implementing the protocol.
> However, any progress we made in getting the OSI system to resemble the
> good parts of ARPAnet/Internet experiments was summarily dismissed by
> people like Jon Postel, Dave Clark, and Mike Padlisky as "Not Invented
> Here" when we attempted to report back to the ARPAnet/Internet community.
> Ultimately it didn't much matter; the superior funding of implementations
> by ARPA, and the "give it away freely" attitude toward those
> implementations, carried the day.
> 
> Cheers,
> Alex
> 
> *From:* Craig Partridge via Internet-history <
> internet-history at elists.isoc.org>
> *To:* Carl Malamud <carl at media.org>
> *Cc:* Carl Malamud via Internet-history <internet-history at elists.isoc.org>
> *Sent:* Monday, May 11, 2026 at 10:48:43 AM EDT
> *Subject:* Re: [ih] Quantifying OSI
> 
> For a chunk of time, NIST funded a bunch of BBN folks to attend including
> Ross Callon (whom, I understand, did a big chunk of the heavy lifting on
> CLNP) and Debbie Deutsch (ASN.1 and I think parts of X.400).  They used to
> meet in the cafeteria periodically to brush up on their (I think French?)
> language skills.
> 
> Craig
> 
> On Mon, May 11, 2026 at 8:29 AM Carl Malamud via Internet-history <
> internet-history at elists.isoc.org> wrote:
> 
>> Guessing NIST had a ton of employees going to ISO meetings, perhaps a FOIA
>> request (or several) is in order. Imagine they will have copies of meeting
>> minutes, enough to start to estimate the number of attendees and number of
>> meetings.
>> 
>> 
> 
> -- 
> *****
> Craig Partridge's email account for professional society activities and
> mailing lists.
> -- 
> Internet-history mailing list
> Internet-history at elists.isoc.org
> https://elists.isoc.org/mailman/listinfo/internet-history
> -
> Unsubscribe:
> https://app.smartsheet.com/b/form/9b6ef0621638436ab0a9b23cb0668b0b?The%20list%20to%20be%20unsubscribed%20from=Internet-history
> -- 
> Internet-history mailing list
> Internet-history at elists.isoc.org
> https://elists.isoc.org/mailman/listinfo/internet-history
> -
> Unsubscribe: https://app.smartsheet.com/b/form/9b6ef0621638436ab0a9b23cb0668b0b?The%20list%20to%20be%20unsubscribed%20from=Internet-history



More information about the Internet-history mailing list