[ih] Ethernet, was Why TCP?

[Brian E Carpenter]

On 02/09/2016 00:19, George Ross wrote:
>> But cheap and cheerful won the day in many campuses, before business
>> even knew that they needed a LAN. As you say, Ethernet only penetrated
>> business seriously was when UTP came along.
> 
> Going by our (Edinburgh dcs) experience, what came before wasn't really
> suitable for serious business roll-out.
> 
> Thick yellow cable was hard to install, and then drilling it for vampire
> taps took a bit of skill and wasn't easy in an overhead cable basket with a
> load of other stuff round about.  And it quickly ran out of bandwidth, and
> installing more was a pain.
> 
> Thinnet was much easier to install, but rather fragile.  We had our techs
> crawling through offices about once a week trying to find the latest fault
> caused by feet, dripping bicycles, hum-loops from contact with the
> plumbing, and so on.  It was easier to install, but we got to the stage
> where we simply had too many separate wires to feed them all through all
> offices, and that put constraints on how people could be assigned desks.

That's absolutely true, but I can tell you that if we hadn't installed
kilometres of Cheapernet at CERN at negligible cost, without ever needing
to make a serious budget request, we wouldn't have had the physicists
(i.e. the users) on our side when we requested a budget of many millions
to recable the entire site with UTP5. By that time (~1995), they were completely
dependent on a site-wide LAN. So that turned out to be the biggest single
funding request I ever wrote, and the quickest to be granted.

So I think the progression Ethernet -> Cheapernet -> 10baseT -> 100baseT
was the only way it could have happened, in academia. And getting back
to the origins of this thread, it was closely linked to the progression
from Proprietary -> Multivendor in the protocol world, where the main
advantage for TCP/IP in the mid-1980s was that it came free with BSD Unix
and especially with SunOS, and ran over Ethernet, just when Unix
workstations were invading our world.

I don't necessarily agree 100% with Ben Segal's view of history, but I
think this is very interesting nevertheless (written in 1995):
http://ben.web.cern.ch/ben/TCPHIST.html

   Brian

> UTP was a *major* improvement on all of that.  It was easy to install, much
> more robust, and simple to re-patch as folk moved office.  Our hub site
> rapidly became a mess of knitting, though, and on more than one occasion we
> had to take everything down over a weekend to re-patch neatly from scratch.
> 
> Soft-configurable VLANs finally made all of this easy to manage.  Install
> and patch once, and almost never have to go back into the IT closets.
> 
> (We had to re-install most of our original UTP, though, because it was cat3,
> a lot of it was over-length, and because 10baseT was so robust we had
> doubled-up ports using the "spare" wires.  But that's another story...)
> 
> --
> George D M Ross MSc PhD CEng MBCS CITP, University of Edinburgh,
> School of Informatics, 10 Crichton Street, Edinburgh, Scotland, EH8 9AB
> Mail: gdmr at inf.ed.ac.uk   Voice: 0131 650 5147   Fax: 0131 650 6899
> PGP: 1024D/AD758CC5  B91E D430 1E0D 5883 EF6A  426C B676 5C2B AD75 8CC5
> 
> The University of Edinburgh is a charitable body, registered in
> Scotland, with registration number SC005336.
> 
> 
> 
> 
> _______
> internet-history mailing list
> internet-history at postel.org
> http://mailman.postel.org/mailman/listinfo/internet-history
> Contact list-owner at postel.org for assistance.
>