[ih] Dan Lynch has passed away

[Karl Auerbach]

Dan was a great person.

My wife worked for him at ACE and I worked with him for the better part 
of a decade and a half via the Interop show network.

We have lots of great stories about Dan - many of which are 
non-technical and some, perhaps, even a tad colored (Dan was always 
surrounded by a cloud of beautiful women) - but those stories can wait 
for another time.

In our stalled History of the Internet (1965-1995) project we had (and 
still have) plans to do several short episodes on Dan, ACE, and the 
impact of the Interop shownet on the growth of the Internet [it was an 
invaluable metronome that drove us to get things working with one 
another], on people [such as my wife and I meeting via Dan, and getting 
married], and just a plain lot of fun.

One of the earliest major successes of the show and Dan's efforts was 
the maturation of twisted-pair ethernet.  The show net was also often 
the first place where many router or switch devices ever met a complex 
topology with lots of alternative patch choices.  (And it is little 
known, but convention centers are not cooled during setup - so we ended 
up heat stressing a lot of equipment, especially at the Las Vegas shows, 
where the convention center temperatures got far above the 120F level 
until the morning of the show.)

Most people saw the glitzy show net as a result but few saw the almost 
continuous efforts, done under Dan's watch, between shows to design 
(often at my house), pre-build (in ever larger warehouses), ship (I know 
that once our gear, which does not include vendor gear, took 46 full 
sized trucks to haul from the warehouse to the venue - we seriously 
considered buying a C130 aircraft), deploy (we often covered much of a 
city, such as Las Vegas, in addition to filling a convention center - 
under very tight deadlines), operated, and then removed.)

Dan was always ready to provide whatever was needed to make it happen.  
Sometimes that was a warehouse, often it was tools. Sometimes it was 
just a moment to relax:  One of the first things we would do is figure 
out Dan's hotel room number - we put rather massive bar bills on his tab 
- and he never once mentioned it, much less complained.

Dan covered for us - because we were always operating on the edge, often 
physically - we managed to get permission (or at least avoid being 
arrested) when we did things like running fibers through railroad 
tunnels in Atlanta or setting up rooftop-to-rooftop laser links between 
locations across Las Vegas.

Dan's shownet trained hundreds of electricians in the arts of network 
wiring (starting with the lesson that it is not nice to cut and attempt 
to splice yellow-hose Ethernet.)

There is no way to overstate the contribution of Dan and his Interop 
vision on the expansion of the Internet.  We used to hold TCP/IP 
bakeoffs, but those were small potatoes to the fire that Dan, ACE, and 
Interop put under actual, demonstrable interoperability of different 
implementations, different products, and different technologies.  
Today's Internet world where one can generally expect things to plug-in, 
turn-on, and work happened largely because of Dan.

Unlike a typical trade show, the Interop show net floor one could 
generally find the people who wrote relevant RFCs and we could figure 
out problems and solutions on the spot.  (I remember once when we were 
deploying FDDI  [at a show in San Francisco] and a bug was found in the 
specifications [by one of our show team crew, Merike K.] and solutions 
were developed among the specifiers and developers, and deployed within 
a few hours.

Dan encouraged my wife to start a company to test the resilience of 
network code and products under real world conditions. (The company is 
still going fine and we are still finding lots of reasons to validate 
Dan's concerns.)

In Monterey, at one of the early TCP/IP Interoperability gatherings - 
around 1987 - I remember overhearing Dan and Craig P. having lunch and 
discussing network control and management.  From that came the push to 
develop protocols like SNMP (and HEMS and CMOT) and instrumentation MIBs.

My wife had been planning to go up to do another video interview with 
him.  Sigh.

At least we got a few hours talking with him on video maybe seven to ten 
years ago.

I don't know how many have seen the Linda Fefferman film of Interop in 
1993 - Dan has a cameo role (as do I):

https://youtu.be/SMkKIaHee4c

         --karl--