[ih] Ethernet, was Why TCP?
jack at 3kitty.org
Wed Aug 31 21:29:50 PDT 2016
There was a paper, also in that early 80s timeframe, where someone
analyzed the wire-level protocols with queuing theory and other magic,
and came to an elegant mathematical conclusion. Unfortunately I can't
remember whether it applied to Ethernet or Alohanet, or what the elegant
conclusion was, or the title/author. It proved mathematically something
like "the system remains stable as long as you keep this key timer below
a value of 1/e". Hence there would be no performance collapse on
Ethernets if you followed the rules.
Curiously, in the neighboring world of TCP and Internets at the time,
there was a similar conclusion reached about stability albeit not very
well quantified or elegant. Something like "everything will be fine as
long as there is a lot more capacity in the network than the amount of
traffic being carried".
Another one was that packet loss rates of 1% on a TCP connection were OK
and could be handled with no problems. As I recall, there was no
blackboard full of equations involved. We just went around the room and
1% seemed reasonable to most people.
Rough consensus and running code.....
On 08/31/2016 07:09 PM, John Levine wrote:
>> So if one wanted a cheaper mechanism, tossing out the extreme discipline
>> made sense to me. We always hear about computing/storage tradeoffs.
>> This was a cost/channel-efficiency tradeoff.
> It is my recollection that at the time a lot of people thought that
> Ethernet sounded too good to be true. If it were heavily loaded, all
> those collisions would surely cause a storm of interference and
> performance collapse, unlike a token ring that shared the capacity in
> a predictable way.
> As we all know, Ethernets worked just fine. A lot of people didn't
> believe it until they saw it, and sometimes not even then.
> internet-history mailing list
> internet-history at postel.org
> Contact list-owner at postel.org for assistance.
More information about the Internet-history