[ih] Design choices in SMTP

[touch at strayalpha.com]

> On Feb 7, 2023, at 11:35 AM, Grant Taylor via Internet-history <internet-history at elists.isoc.org> wrote:
> 
> On 2/7/23 11:37 AM, Guy Almes via Internet-history wrote:
>> <> and the speed of light got worse.  I mostly mean this in the sense that it didn't get any better, while processing and transmission both got much faster.
> 
> This is an interesting take on -- what I believe is called -- the propagation delay becoming a larger proportion / percentage of the delay as things around it got better / faster.

Everybody talks about the speed of light, but nobody every does anything about it!

>> But it's also true in the literal sense that wide-area networks were largely built with microwave in the 1980s (or geosynchronous satellites), while they are built with fiber optic circuits now. And the speed of light through fiber is about 2/3 the speed of light through air (or space).
> 
> This floored me the first time I heard about it.

Latency is both multidimensional and those dimensions combine in nonlinear ways.

Propagation delay is only one of those dimensions, the other parts of which include symbol rate (which affects message rate), encoding latency (symbol layer, encryption, and error correction), whether the channel supports pipelining and/or parallelism, as well as server access and computational delays. Not to mention that none of this is a property of a given channel - it’s a property *of a given message* (or exchange) over a given channel.

Note that 2/3 (67%) is the ratio for fiber to free space, fiber is basically the same propagation speed as coax and most twisted pair. Twin-lead is actually faster - 80%. The only common media that comes close to air/vacuum is open-ladder wire.

Joe