[ih] Early History of the Internet

[Dave Crocker]

On 1/9/2024 8:14 PM, John Levine via Internet-history wrote:
> Were you suprised at how well IP scaled
> with only minor tweaking along the way?


I thought the collected set of revisions, done in the latter 1980s, as 
the Internet scaled, were considered substantial. Seemed that way to me, 
at least.


And to the extent anyone thinks the congestion control and 
retransmission details provided for TCP are actually minor: Around 1988 
we took over maintenance of a PC stack, from a recently departed 
engineer, and eventually got very serious customer reports that it was 
flooding the their network with packets.  Upon inspection, I discovered 
that the TCP the engineer had written used a fixed, very short timer for 
retransmission.  Given very high reliability and very high bandwidth of 
a LAN, this hadn't been a problem.  As soon as the LAN was accessing the 
wider, slower, less reliable Internet -- as was becoming more common by 
then -- this timer's model didn't work so well...


d/

-- 
Dave Crocker
Brandenburg InternetWorking
bbiw.net
mast:@dcrocker at mastodon.social