[ih] AOL in perspective

Olivier MJ Crépin-Leblond ocl at gih.com
Thu Sep 4 18:50:26 PDT 2025



On 04/09/2025 14:36, David Sitman via Internet-history wrote:
> Daniele asked me to try to distribute this to the list because his attempt
> appears to have been unsuccessful.
>
> David
>
> -----Original Message-----
> From: Daniele Bovio [mailto:Bovio at aol.com]
> Sent: Wednesday, August 20, 2025 3:17 PM
> To:internet-history at elists.isoc.org
> Subject: [ih] AOL in perspective
>
> A couple of comments about AOL:
>
> I was hired by AOL in 1995 to build and manage the network operations for
> Europe, in order to allow the AOL-Bertelesmann joint venture to launch and
> operate AOL in UK, Germany and France. The first question I was asked by
> Suk S. Soo, ANS director, during my interview was: how do you make money on
> the Internet? Coming from the academic world as former Technical Director
> of the European Academic Research Network (EARN) and member of the EAT
> (EBONE Action Team) the group which designed and implemented the first
> pan-European Internet Backbone, you may imagine my answer was less than
> adequate, as for the first time I was confronted with the concept of a
> "commercial" Internet.
> As I quickly learned the AOL business model was to sell content/media to
> the residential market via a different mean than TV by allowing the
> households PCs to connect via the telephone network. The problem was that
> in 1995 there was no ISP in Europe yet able/interested to offer internet
> access to the residential market and so we had to build everything from
> scratch, i.e. a dial-up access network and transatlantic WAN circuits
> connecting to the US AOL data centers and, via ANS, to the US Internet.

The situation in the UK was a little more advanced with a handful of 
small ISPs like Demon Internet that had started its services in 1992. 
That was partly due to PIPEX that had started offering Internet Service 
in 1990 and was ready to sell to budding ISPs who had only one Internet 
option: to connect to the US.

Working on proprietary communication networks that were used for banks 
(and using X.25 + Leased Lines), I remember visiting a data centre in 
1993 or 1994 leased/owned (?) by WorldCom, that was located a couple 
hundred metres North East of Old Street roundabout (later called 
"silicon roundabout" and now no longer a roundabout), where one floor 
was used by the City's banks proprietary networks and servers which were 
already pretty full. We walked up to the next floor which was largely 
empty. The raised flooring had gaping holes here and there in order to 
pull cables through and in one corner of the room were no more than six 
or seven stacks. Three were empty. One stack was full, with some kind of 
PSU at the bottom, a couple of CISCO routers and three or four racks of 
USR modem cards with leds flashing, wired quite randomly - and a Dymo 
generated label saying "COMPUSERVE". Next to that stack was another 
stack that also had its PSU plus CISCO routers but only two racks of 
modems - probably no more than 16 or 20 modem cards in total... and a 
label "AOL" - even though the service was still sold as "America On 
Line". All of these modems were connected via a spaghetti-like mass of 
cables to the two other racks next to them on which rows and rows of 
standard BT telephone sockets had been fixed on a plank of chipboard and 
the telephone numbers scribbled using a pencil by each socket. "There's 
something like 40 telephone lines here", our host said.

We laughed at how artisan this all looked compared with the floor we 
were working on, where structured, coloured and labelled orthogonal 
cable routing was already in effect, with a mix of 10base2, 10baseT, 
whether through Ethernet or Krone blocks. But then the bank's floor was 
the "expensive" floor and the future Internet floor was the "cheap" 
floor... or at least that's what it looked like at the time. When I 
asked our host "so when are we doing to connect the floors together?", 
he laughed and said "are you joking? Banks don't want to get connected 
to this mess. Imagine the security risk!"

It took quite a few years to finally make the connection between the two 
floors - I was not around when it happened, but I am sure it was a big deal.

> So, indeed, the main target of AOL was not to be an ISP, but becoming one
> was an essential component of its business plan, and therefore, given its
> popularity, AOL became quickly one of the largest ISPs in the US and in
> Europe.

It also became the world's largest distributor of good quality coffee 
mug coasters in the shape of 3.5in floppy disks and CDs. :-)

>
> That the main target of AOL was not to be a pure ISP in the long run was
> clearly apparent when in 1998 AOL acquired CompuServe via a 3-parties
> agreement with WorldCom: AOL acquired the CompuServe subscribers and sold
> to WorldCom ANS and the entire CompuServe Network Services department.

Seeing how the services essentially used the same facilities, it 
probably made sense technically too.

>
> Network being a fundamental pillar of its business however AOL kept
> in-house the management of all the network infrastructure in the US and
> Europe, so still acting as an ISP to its users, but the cost was huge,
> particularly for
> Europe: in 1998 AOL spent $37M for transatlantic circuits and was
> projecting an expenditure of $58M by Y2K.

At the time I visited the data centre, I was quoted £100K per annum for 
a 64K line from Old Street to MAE East so I can imagine the costs.

Thanks for sharing your struggles in dealing with European TelCos... I 
can see we young ones at the time were not the only ones struggling with 
them...

Kindest regards,

Olivier

-- 
Olivier MJ Crépin-Leblond, PhD
http://www.gih.com/ocl.html



More information about the Internet-history mailing list