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

[no subject]

"During this quarter, responsibility for? gateway maintenance and
development was transferred from the Information Sciences Division to
the Computer Systems Division (now Communications Systems Division).?
The motivation for this transfer was the need to emphasize the treatment
of the gateways as an operational communications system, rather than a
research tool to support the growing user community.? In this approach,
we plan increasingly to treat the gateway system much as we do the
ARPANET and SATNET systems in terms of monitoring and maintenance.? This
will require increased emphasis on the development and enhancement of
tools for the remote operation of the gateways."

I remember writing that.? Vint had talked to me earlier that year to see
if I was willing to take over the gateway work and fold it in to the
"operations and maintenance" we had already been doing on the ARPANET
for the previous 10 years as well as more recently SATNET.?? I think
Vint saw the need for the Internet to be up all the time, not just for
experiments and demos, and for someone to be called to report problems.

To me now, this was an inflection point in the history of the Internet
-- when it went from being a research tool to being an operational 7x24
service.? To accomplish that, we plagiarized eagerly from the ARPANET,
introducing the same kinds of tools and processes that had evolved and
been proven over the previous decade. ? It also involved rewriting the
gateways into assembly language from the earlier research implementation
in BCPL.??? Our Division had been running the ARPANET for a decade, and
the NOC was just down the hall from the "Gateway Guys" offices, so
technology transfer was straightforward.

At some point in that process, the gateways were added to the repertoire
of things that the ARPANET NOC operated on a 24x7 basis, and a gateway
control terminal appeared inside the ARPANET/SATNET operations room, and
the operator(s) on duty were responsible for also keeping the gateways
running, just as they had been doing for the ARPANET and SATNET IMPs.

Prior to that, of course we had to build and debug the appropriate
software.? The "control console" was simply a terminal connected to the
BBN PDP-10 where the management software ran.?? Sorry, I can't remember
the name of the software, or which BBN-xxx machine it was on.? You could
"control the Internet" simply by connecting a terminal to that software,
and your terminal became the "control console".

So, as that quote says, at some point before it went to the NOC I'm sure
I tried it out by connecting from the terminal on my desk.? I had a
reputation for being able to find bugs within minutes after somebody
declared something ready.

However, it was much more likely that the control console was in use by
somebody else, either working in my group or one of the ARPANET-related
ones.? At the time, Bob Hinden, Mike Brescia, and Alan Sheltzer were
working on gateway development, and writing the code.? David Floodpage
had been developing the CMCC - Catenet Monitoring and Control Center,
which was used to operate SATNET.? Marty Schoffstall was working on
other pieces - e.g., what later became SNMP.? Lots of other people who I
have probably missed.

We pushed very hard on getting mechanisms into place in the IP/gateway
world that reflected the tools that had proven useful in the ARPANET -
things like Traps, Software Download (see XNET), traffic statistics,
controls, patching, etc.? Lots of that stuff eventually made its way
into RFCs et al, and also made the Internet into a 24x7 service.

Hope this helps,

/Jack Haverty

Internet-history mailing list
Internet-history at elists.isoc.org