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

[ih] The history of "This" network?

On 2019-02-12 16:18, Dave Crocker wrote:
> On 2/12/2019 3:58 PM, Michael Greenwald wrote:
>> The Arpanet Protocol Handbook was available in every office I sat in
>> at MIT
> I long-ago lost my copy of that.  Plus it came a bit late4
> The Computer History Museum shows two versions in its archive -- 1976
> and 1978 -- but apparently neither has been digitized. sigh.
>>    so the relevant NICs and RFCs seemed equally
> I'm not sure what you mean by NICs as you are using the term.  I'm
> used to its meaning Network Information Center, which is a group
> operation, not a document series.  My bad memory?

Could be *my* bad memory. I thought I recollected that the Arpanet docs
were numbered NIC xxxx (where xxxx seemed like a very high number) and
the IP/TCP documents were RFC yyy (where yyy seemed much smaller than
the NIC numbers).

I think maybe several documents had both numbers (but I am really unsure
about this): maybe if a NIC (or whatever the real abbreviation was, if I
got that wrong) document referred to an RFC it got a NIC number?

>> available. So I didn't feel that it not existing in an RFC form was a
>> lack in any way. NIC vs. RFC didn't make much difference to me,
> From 1969, work produced by that community got circulated by RFC.  It
> was easy and, I thought, automatic.

I don't know about 1969, but to the extent that my memory is correct,
by 1978 I think there were at least two different series of documents.
Someone with more reliable memories than me should respond. But I doubt
that I am hallucinating this *entirely*, so I am fairly confident that
there's at least a kernel of truth in my recollection. They might not
have been called NIC, or perhaps NIC xxxx meant some other series of
documents produced by the NIC? I am fairly confident that there were
documents with NIC/<whatever> numbers that were not RFCs, but were
available to people.

>> That said, although NCP was easily available, I believe the 1822
> Well, yeah, the BBN specs were essentially independent of documents
> produced by the 'network working group'.
> 0
>> spec (Host-Imp protocol) was *not* as easily available (it was also a
>> much bigger specification [if I am remembering correctly] than any
>> other protocol I came across back then --- maybe that's the reason it
>> wasn't included). The Host-Host protocol was, I think, in the
> NCP and Host-Host were the same thing.  (Well, NCP referred to the
> implementation.)
> Uh oh.  facepalm.  That's why 'NCP' doesn't up in the early RFCs.

My bad. You are almost certainly right --- Host-Host protocol and NCP
were essentially the same. My brain conflating things -- I was probably
confusing NCP with ICP (initial connection protocol) or some other 

>    HOST-HOST Communication Protocol in the ARPA Network*
>    https://www.rfc-editor.org/rfc/rfc33.txt
>> Totally a guess here: I don't think anything surprising or
>> interesting occurred with this. I think the NICs were just older and
>> more established than the RFCs, and because they were older and not
>> directly relevant to the IP/TCP effort
> Older than 1969?

By older, I just meant that as a group Arpanet documents were older than
Internet documents, and that they referred to older technology. 
facing, as opposed to forward-facing. So there could have been reason to
avoid adding every single Arpanet doc into the Internet series.

I just looked at RFC6529 (should have looked before) and it says, about
Host-Host Protocol Document No. 1" that:
    "It was one of a set
    of such standards maintained as a separate set of documentation by
    the Network Information Center (NIC) at Stanford Research Institute
    (SRI).  The January 1972 version (NIC 8246) reproduced here also
    followed that approach. "
So NIC 8246 was the numbering I was referring to.

> d/
> --
> Dave Crocker
> Brandenburg InternetWorking
> bbiw.net