On Tue, Mar 21, 2023 at 10:11 PM Christian Huitema <huitema@xxxxxxxxxxx> wrote:
On 3/21/2023 6:39 PM, Raghu Saxena wrote:
> That's a pretty interesting note, and one which I think deserves a
> proper discussion. Considering how much more energy use is going towards
> the internet everyday, I do believe IETF plays a crucial role in
> increasing the efficiency (energy wise) of the internet.
>
> For instance, by moving towards HTTP/3, if we reduce the RTT to
> establish a connection, that's less CPU time spent on establishing the
> connection, less electromagnetic radiation (fiber optics / wireless)
> required, since now x less packets need to be transmitted etc.
>
> I do think using available funds for further advancing the goals of IETF
> directly would be a better use - you could argue that IETF is producing
> carbon offsets as a direct consequence of its work!
Efficiency increases are an obvious good thing but in the past,
efficiency increases have been matched by increases in Internet usage,
such as more videos or larger images, and I am not sure the total energy
consumption decreased over time. If it did decrease, that means saving
energy, but most of the energy saved comes in the form of electricity.
Only a fraction of the electricity is produced by burning carbon based
fuels, and that fraction is supposedly decreasing over time as cleaner
techs replace burning coal, oil or gas.
Regarding the emissions associated with a phone or a cpu during its lifetime, I think that most of these emissions happen during the manufacturing and transport of the device/hardware - as opposed to the device running. That is more or less what you are saying below.
We could argue that tools such as video conferences do directly save
carbon emissions by reducing the need for physical travel. That's
plausible, but very hard to measure because the Internet also enables
making friends in remote places, which is a very good thing but
ultimately encourages more traveling.
And then there is a counter argument that the Internet enables more
international commerce, e.g., someone in America buying widgets from a
producer in China through Amazon or Ali-Baba. These items have to be
shipped across continents, so in a sense the Internet contributes to
increased carbon emissions. I have no idea how all that balances, and I
would indeed be delighted to read a good study. But that seems a job for
ISOC, not the IETF.
As far as the IETF is concerned, the direct challenge ought to be
whether we can be as efficient while traveling less, or at least by
relying less on long distance air travel. We did run a two years
experiment during Covid, and the results have been mixed: work kept
progressing, but some measures like number of version-00 drafts did
regress. We have to do better!
I am questioning the metric here - especially if that is the only thing we have found - as a 00 draft is probably the easiest way to justify the need to travel ;-).
-- Christian Huitema
Daniel Migault
Ericsson
