[ih] How long did it take?

[Michael Grant]

This isn't directly related but more stuff I remember from that time 
period around X.25.

Around the time of "The Event", around 1988/1989 (the OSI trade show 
that happened exactly once in Baltimore, MD), I worked for COS.  I was 
in charge of the testing and integration lab at COS from the technical 
side--meaning I wasn't management, I was the guy who did all the work.

In coordinating all of the vendors and such, we were depended greatly on 
email and phone calls.  Email back then was basically our uucp link to 
Rick Adam's box, I think it was still seismo at the time and later he 
broke away from that and it became Uunet, or maybe Uunet already existed 
by then.  Later after The Event, towards the end of COS, I did a tiny 
bit of hourly work at Uunet to help put together some T1 channel banks 
on the side during their initial build-out.  At the end of COS, I had a 
choice to go to Uunet or Sun and I chose Sun.

Anyway, I had been complaining bitterly to my management within COS 
about the lack of email and was told that I needed to get X.400 working 
and that under no circumstances would we be allowed to connect up to the 
internet (we had uucp but management did not really understand what it 
was since they used the Wang system for internal mail).  In the lab we 
had two X.25 networks.  Telenet and one other which I am spacing on the 
name at the moment.  Both were free for us to use because they were part 
of the lab donated by vendors.  Both were connected by 56K sync serial 
leased lines back to the vendor.  (We eventually did get a slip line to 
Uunet but it wasn't until nearly the end).  From memory, you could not 
get a connection from one to the other, like X.75 either wasn't fully 
working yet or it had not been turned on as some feature, I don't 
remember, but I do remember you couldn't get from one to the other.

Both of these things were sitting, essentially unused in the COS lab and 
I was told to use them to "connect us up to the OSI internet and get 
X.400 mail working and stop asking about the internet".  So I had this 
sort of ongoing side project (a one man project) to try to do this.  I 
set up ISODE on a Sun.  I messed with Sendmail.  I think it was Peter 
Vanderbilt (sorry if I misremember the name) who fed me a sendmail.cf 
which with some help we doctored up to get an X.400 gateway to something 
at Sun.  And here's where X.25 comes in.  I think it was point to point 
over the X.25 network with the X.25 address embedded inside the very 
long OSI address if I am not mistaken and I had that hard-wired into the 
config somewhere on one of the Sun's and the Sun was plugged one of the 
sync serial ports on the X.25 concentrator in the lab.  I think Peter 
and I passed approximately 1 email back and forth across this link and 
that was probably the only use it ever saw.

This had me wondering how on earth would one use X.25 as "the OSI 
internet" when X.25 had addresses that looked like phone numbers and 
there was nothing like ARP on a wide area network and there were no wide 
area routing protocols and ES-to-IS was only some vague dream, there 
were no host names, not even a hosts.txt file floating around, and 
nobody could tell me how any of this stuff fit together.  And not to 
mention the TP0-TP4 non interoperability issue.

I failed to see how X.25 could be used as an internet.

Not only that, I couldn't understand how you could do an internet with a 
charging model where you paid for packets and connection time like long 
distance calls.  Nobody was going to pay to let someone else download 
something from them.  I tried to explain this to the people I was 
working for and got blank stares or things like someone would set up 
some way to pay.  I tried to explain to them the lack of host names, 
lack of routing, lack of any sort of address resolution and it all fell 
on deaf ears and was told to go away and do my job.

It was quite obvious to me back then that no single vendor was going to 
maintain some sort of hosts file.  There was no way to do any sort of 
default route and I saw no one putting together any sort of global 
routing of OSI addresses.  There was no way to use X.25 as an internet.  
I never could see the future here.  I mean, it was usable if you had 
some private network or some private point-to-point link you needed for 
some reason, but as some public internet like the Internet, nope, even 
back then, I could not see how X.25 was ever going to do that.  I tried 
in vain to try and explain this to the people I worked for and I was 
told I didn't know what I was talking about and that I would see, that 
all these problems were being worked on and would be solved.  I tried to 
get people at COS interested in being that overall organization to do 
some of these things but absolutely nobody would listen to me.  Perhaps 
just as well, it might only have prolonged what was already a stupid 
mess.

Anyway, those are some thoughts I had from that time period, not really 
history per se.