An engineer battles network problems while a grumpy new bride is getting impatient
My story is set in the
late 1980s when I was network manager for an airline telecomms
Harare in Zimbabwe. My job involved
ensuring that all our customers (airline offices) had their terminal
connected to their mainframes on the other side of the world, via our
In those days the now-universal
TCP-IP protocols were still in their infancy, and we used a number of
proprietary airline protocols, notably IPARS, a (now) weird 6-bits per
polling protocol designed to make the best use of the 2400 BPS lines
commonly used in those days. Anyway – to my
Qantas (the Australian airline)
came to me one day and asked me to provide a working terminal for them
travel show that was being held at a conference center on the outskirts
town. Not a big ask, except that this
show was in two weeks time and getting lines put in within that
at that time was well nigh impossible. But being a good
service-oriented manager, I said I would do my
to expect too much, and could I have a box of giveaways to ease the
our line application through the ponderous bureaucracy of the Zimbabwe
They responded with a box full of
Qantas pens, caps, bottles of wine, etc and I got to work, calling in
turning on the charm and dispensing giveaways to anyone who might be
help get all the ducks in a row for us. I borrowed an EEPROM programmer
and reprogrammed a spare
to make it talk IPARS. I sweet-talked
our configurations people at head office to get the network
in half the usual time.
There was one rather large fly in
the ointment: Just before the show
started, I was getting married at a resort about 300 Km away from
Harare. So with some trepidation, I
exhorted my staff
to follow up everything that needed following up, and went and got
married. Fast forward a couple
of days. I returned to Harare in a state
of newly wedded bliss, with one spare day before we left on honeymoon,
to my wife that morning (the day before the show kicked off): “Dearest,
I’m just popping into work to check
if my Qantas terminal is up and running. Okay?”
So at 9
am that day I was back in at work expecting maybe the odd minor
problem. And my staff had done a good job. The line was in, the modems
the terminal was being polled and all looked good. Except that the
terminal was not getting any
responses from the Qantas host. The travel show was in
a large conference hall and there was only one working phone there,
happened to be on the other side of the hall from the Qantas stand.
Cellphones were not available in Zimbabwe in
So to get them to make test
entries on their terminal, I had to phone that number, ask whoever
get someone from the Qantas stand to phone me back , ask them to make
for a few minutes, and then monitor their line with our protocol
monitor. Yes, we could see their entries,
responses. So we then had to get the
next upline communication center,this case Johannesburg, to monitor,
and repeat the whole
get entries made on the terminal. Again
they saw entries but no host responses.
My boss was understandably not keen on maxing out his international
so all this had to be done by messaging rather than just phoning, which
rather difficult to get the monitoring done at the same time as the
were being made. By this time it was
getting near lunchtime. I phoned my new wife and said I was having
would she mind if I missed lunch? More
groans. And so we managed over the next
few hours to monitor the traffic from our little terminal all the way
our network through Johannesburg, London and Sydney and finally got
confirmation from our Sydney center that they could see entries going
Qantas host system, but no replies.
network operator in Sydney sent me a message, saying that he had sent a
to the Qantas network center but there was as yet no one there who
further (there was a 10-hour time difference) and that he was going
but had let the new operator know about the problem. He also helpfully
gave me a phone number for
the Qantas data center. By now it was nearly 5
PM and in full desperation mode I asked my boss if I could phone Qantas
in Sydney. “Go for it,” he replied.
I phoned Qantas in Sydney and got a
friendly Aussie lady on the line. I
explained the problem and asked if she had any ideas. “Well let me
check my configuration”, she
said. We had about 5 address parameters
for the communications centers where the traffic entered and exited our
network, and our configuration message used a horrible mishmash of
Hexadecimal and Octal values. We went
the configuration and confirmed all the values until we got to the
Cluster Identification (TCID) value.“That’s 43 Hex,” she said. “No,”
I replied, "That’s 043 OCTAL."
is.” she said. “Sorry for that. Hang on a sec.”
heard tapping on a keyboard on the other
side of the world. “OK”, she said. “Try that now.”
I had to go through the whole phoning
rigmarole again. Finally the Qantas rep
at the travel show got back to me. “It’s
working!” he said. “I can sign in!” I said. “I’m out of here!” 20
years later, I’m still married, but I sure know the difference between
David Ashton is an Australian engineer who says this experience gave him
the satisfaction of a job done on time, a happy customer AND a happy boss--at the cost of a slightly grumpy wife.
ylshih...I think that AC "Gimli Glider" was on that "Air Crash Investigations" programme...have you ever noticed that it's almost always a combination of 2 things going wrong?
Most of you would have seen that T-shirt / Bumper sticker:
There are 10 kinds of people in the world....those who speak binary and those who don't!
Thanks all for the nice comments.
Great story and an understanding wife!
Bases/units have caused many problems. The Air Canada "Gimli Glider" (767 airplane that had to glide into a rural airport in Canada) incident arose in part, after a chain of other events, due to a mixup between gallons and liters during the switchover to metric.
It sounds like you really went the extra mile to ensure customer satisfaction! It reminds me of the advertisement that I once saw for a mail sorting machine that moved envelopes at 300,000 km / second. Pretty impressive that they were moving "snail mail" at the speed of light. If the truth were to be told, I believe that they really moved the mail at 300 cm / second. Getting units right can be so pesky.
It's really mind-boggling to think of what my hourly rate would be if I were paid for the actual time that I spent troubleshooting a problem to get to a fix. The challenge is to avoid going down a rathole.
David Patterson, known for his pioneering research that led to RAID, clusters and more, is part of a team at UC Berkeley that recently made its RISC-V processor architecture an open source hardware offering. We talk with Patterson and one of his colleagues behind the effort about the opportunities they see, what new kinds of designs they hope to enable and what it means for today’s commercial processor giants such as Intel, ARM and Imagination Technologies.