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

[ih] inter-network communication history

 In reference to Jack's 2nd paragraph, there were two papers published about the early ARPAnet NOC.? They are:
McKenzie,A.A.; Cosell, B.P.; McQuillan, J.M.; Thrope, M.J. 
TheNetwork Control Center for the ARPA Network.
Proceedingsof the First International Conference on Computer Communication,Winkler (Ed.), Washington, DC, pp. 185-191.October1972

TheARPA Network Control Center.
FourthData Communications Symposium, Quebec City, Canada, pp. 5-1 to 5-6.October 1975

    On Thursday, November 7, 2019, 7:37:17 PM EST, Jack Haverty via Internet-history <internet-history at elists.isoc.org> wrote:  
 I think the beginning of the story is captured in IEN 105 - see

David Floodpage wrote that in 1979, as part of our work in operating
SATNET and beginning to think about how to use ARPANET-style tools in
the context of the "Catenet" which Vint had popularized at the time as
the name for what we now call Internet.??? David and Marty Schoffstall
worked in my group on those kinds of things; Marty later was one of the
authors of SNMP.

I don't recall that much was ever written about the internal IMP
operations and management tools and techniques developed during the
ARPANET growth through the 70s.? So that technology may never have
escaped BBN, or is perhaps captured in some obscure QTRs that few ever
read.? But those "gateway guys", including Floodpage and Schoffstall,
and the "Arpanet guys" were intermixed in the BBN building so lots of
technology transfer occurred even if only in the hallways and around coffee.

If you read IEN105, and replace "gateway" with "IMP", and "CMCC" with
"NCC", you'll get a reasonable concept of how some of the ARPANET
internals worked through the 70s.?? That was David's charter - to figure
out and then transfer the ARPANET tools and methods into the "Catenet"

IIRC, this was all before the IETF got started.? I just looked at the
"Internet Timeline" in RFC2235, which claims the IETF was created in
1986.? My personal recollection is that it happened much earlier; I
recall being present at the ICCB meeting when the IETF and IRTF were
formed out of the unwieldy "Internet Group".?? If it was 1986, I
wouldn't have been there.? I'll have to look at my paper notebooks if
the mold in the basement permits.

I don't think there ever was an NMP defined; it may have been just a
vague term used within BBN as a placeholder for some yet-to-be-defined
mechanism.? That's why I said it was more of a concept than a
protocol.?? Concept?? Fantasy?? Yep, but the thinking was still based on
real-world experiences.

Floodpage started the ball rolling in the Internet (aka Catenet) world,
using the ARPANET world as a working example.? More or less concurrent
with that IEN in 1979, his work involved building the CMCC (Catenet
Monitoring and Control Center), which actually got used.?

In that era, documentation like RFCs and IENs often represented what had
already been built and tested out -- even if it was couched as a
"proposal".?? Politics didn't have the chance to delay implementation.

Marty Schoffstall was in the same group, and later an author of the SNMP
RFCs.? SNMP came quite a bit later, and that's when we had the musings
about exactly which component was meant to be "Simple".

In retrospect, I think having the people doing development also in some
way responsible for operations is really important.? When you're
fighting fires with angry users lurking, there's little time for politics.


On 11/7/19 3:46 PM, Dave Crocker wrote:
> On 11/7/2019 3:33 PM, Jack Haverty via Internet-history wrote:
>> SNMP was instantiated as a first step toward a comprehensive NMP
>> (Network Management Protocol).?? As I recall, NMP was more of a concept
>> than a protocol.? It was an umbrella covering a lot of pieces.
> That doesn't match my memory at all.? There were all sorts of
> fantasies being spun around these efforts, so perhaps that was part of
> the mix, but it had nothing to do with the actual work, as I remember it.
> SNMP was the result of a few pragmatic operational people creating a
> really simple protocol, with some of the usual IETF pixy dust tossed
> onto it, except for the political s--t storm it got stuck in with CMIP.
> The CMIP people were very well organized and very political, but had
> produce a very complicated spec and had trouble making it pratical.
> Craig's worthy HEMS proposal was in this mix, of course, but didn't
> have the track record or political support to overcome the... politics.
> It was amusing to those of us at Wollongong, at the time, that I had a
> senior person in multiple, critical positions for both SNMP and CMIP
> (-> CMOT) included the just-lost Keith McGlogrie.
> The (unfortunate) effect of the politics, besides quite a bit of
> delay, was forcing SNMP to use data cast in ASN.1 -- in the hope there
> would at least be that level of interoperability with CMIP/CMOT -- but
> thereby substantially increasing the complexity of using SNMP.
> d/
Internet-history mailing list
Internet-history at elists.isoc.org
Internet-history mailing list
Internet-history at elists.isoc.org