►
From YouTube: IAB workshop on Environmental Impact of Internet Applications and Systems 4: Next Steps
Description
IAB workshop on Environmental Impact of Internet Applications and Systems 4: Next Steps
Workshop webpage: https://datatracker.ietf.org/group/eimpactws/about/
Papers in GitHub: https://github.com/intarchboard/e-impact-workshop-public
Session 1 (The Big Picture): https://youtu.be/90GxlL34rQ4
Session 2 (What Do We Know?): https://youtu.be/EaNgREHLXRg
Session 3 (Improvements): https://youtu.be/jjEZwuuChZc
Session 4 (Next Steps): https://youtu.be/Pc_XY5sDR58
A
We
don't
have
more
than
90
minutes
today,
so
maybe
we
actually
did
do
get
started
so
welcome
again.
So
this
is
the
fourth
and
last
session.
A
The
discussions
will
continue
after
this
Workshop,
of
course,
but
this
is
the
last
of
the
main
sessions
we're
going
to
talk
about
future
developments
and
next
steps
and
there's
some
ground
rules
that
I
think
you've
seen
by
now.
So
I
don't
plan
on
repeating
them,
but
please
continue
to
keep
them
in
mind.
You've
been
doing
very
well
and
everybody's,
been
explaining
very
carefully
I
think
it's
been
very
understandable
to
everybody.
So
that's
it's
great.
A
Despite
our
backgrounds
and
goals
for
today,
there's
a
little
bit
of
a
look
back
at
the
previous
sessions
and
then
we'll
talk
about
whether
it's
our
future
developments,
the
different
areas,
we'll
talk
about
conclusions
and
next
steps-
that's
discuss
about
them.
We'll
also
talk
about
that.
Workshop
next
steps.
More
concretely,
it's
a
RFC
coming
out
on
on
each
of
the
IAB
workshops.
A
So
we'll
do
that
again,
but
not
only
that
will
let
me
discuss
how
we
continue
and,
as
usual,
Western
slides
from
the
beginning
are
a
useful
reminder
on
what
can
we
do
and
what
are
the
potential
classes
of
conclusions?
What
we
know
or
don't
know,
what
can
we
do
in
terms
of
improvements
and
what
might
be
the
next
steps
and
today's
agenda?
So
indeed
there's
some
summaries
by
myself,
Colleen
and
Eve
on
the
three
different
sessions
we've
had
so
far.
A
We
have
five
talks,
I
think
on
potential
future
development
areas
on
carbon
neural,
networking
technology,
mobile
networks,
what
we
should
not
be
doing,
it's
an
interesting
item
always
and
then
ietf
and
ITF
process
and
I'm
gonna,
try
and
ask
you
guys
to
be
brief
and
to
the
point
we
don't
have
a
huge
amount
of
time,
given
the
number
of
talks
on
the
agenda.
If
you
want
to
actually
leave
some
time
for
discussions
right,
you
take
too
much
time.
I
I
might
actually
try
and
cut
you
off.
A
Then
we'll
have
some
discussion
mostly
about
like
what
is
the
conclusion,
or
what
should
we
be
doing
concretely
and
and
then
I'm
going
to
reserve
it
just
a
little
bit
of
time
at
the
end
to
talk
about
the
next
steps,
work
supplies
calling
anything
to
add.
B
A
Yeah
and
any
any
agenda
batch,
otherwise,
foreign.
A
Brief
summary
of
session,
one
which
was
I,
thought
really
really
interesting,
because
we
had
tons
of
yeah,
basically
different
viewpoints,
I
think
at
least
for
me,
some
of
that
was
was
really
new,
so
it
was
very
refreshing,
but
I
did
spend
a
little
bit
of
time
beforehand,
trying
to
figure
out,
like
you
know,
if
you
look
at
anything
in
this
space,
what
can
we
learn
from
it
and
like
what
are
the
things
I
should
have
on
my
slides
or
or
on
the
notepad
came
up
with
a
like
this
template.
A
This
observations,
useful
tools,
actions
that
we
can
actually
take
and
goals
that
we
should
have
and-
and
the
really
interesting
stuff
also
is
often
there's
some
research
problems
that
clearly
are
not
things
that
we
can
engineering
wise
do
today,
but
what
will
be
fun
if
they
they
resolved
and
the
good
topics
for
people's
students
to
go
and
do
a
PhD
thesis
or
something
write,
the
paper
measurement
so
for
session.
One
I
tried
to
fill
this
in
and
there's
you
know
plenty
of
things.
A
Of
course
I'm
not
going
to
go
through
this
in
detail,
but
just
you
know
the
the
main
feelings
that
this
is
so
much
bigger
than
our
Tech.
Obviously,
of
course,
but
you
know
maybe,
but
for
all
us,
maybe
for
me
in
particular-
I
wasn't
like
necessarily
obvious
before
this
Workshop,
but
that's
that's
the
case
and
that
influences
all
all
parts
of
society
and
there's
a
bunch
of
non-technical
things
also
involved
business,
summarize
them
as
Institute,
but
inequality,
externalized,
costs,
Justice
or
lack
thereof,
and
so
on.
A
So
that
that's
a
good
thing
to
keep
in
mind.
So
this
is
really
Broad
and
also
I
talked
a
little
bit
about
the
improvements,
how
they
come
in
different
forms,
and
we
should
keep
that
in
mind
not
to
get
stuck
on
like
the
thing
that
I
control
that
I'm
going
to
improve
and
solve
everything
with
that.
That's
not
the
case.
Probably
we
need
multiple
things
and
then
useful
tools,
and
this
is
like
a
slightly
different
from
maybe
like.
A
If
we
look
at
some
of
the
other
sessions,
the
usual
tools-
I
can
appear
like
this
thick,
but
but
here
is
like
softer
things
like
solidarity
and
awareness
and
sufficiency
of
you
know
what
what
we,
what
we
have
or
we
should
be
happy
with
what
we
have
and
how
we
should
make
that
actually
last
and
and
serve
our
needs,
rather
than
just
chase
the
numbers,
good
good
things
to
think
about,
and
also
not
waiting
for
the
perfect
solution
that
that's
always
bad.
A
For
almost
anything,
and
particularly
true
of
this
things,
we
can
do
on
the
right
side.
I
have
some
goals,
and
one
of
them
is
continuous
improvements.
So
that's
a
pair
of
course.
That's
also
some
more
concrete,
hard
things
like
this
use
of
renewable
energy
in
in
the
systems
that
we
run
and
and
also
carbon
awareness,
that
you've
talked
about
actions
that
we
should
take
I
think
from
sessor
one
in
particular.
A
It
was
pretty
clear
that,
like
since
we
are
not
alone
in
this,
we
should
perhaps
be
connected
somehow
and
not
to
work
in
isolation,
but
be
aware
of
what's
Happening
elsewhere
and
be
aware
of
what
the
demands
are
and
possibilities
and
working
together
and
so
on,
and
then
the
goals
and
some
continuous
Improvement
also
but
also
I'm
a
little
bit
moving
away
from
this
chasing
bigger
numbers
or,
as
I
put
it
bigger
and
more
bloated
web
pages
every
year
to
actual
increased
usefulness
of
you
know,
whatever
we
do
with
icts,
too
or
or
the
internet
usefulness
to
me
and
to
the
society
at
Large
and
also
declining
emissions.
A
I
think
that's
that's
been
fairly
clear
that
there's
some
some
demand
for
reduction
and
not
just
staying
where
we
are,
but
at
the
same
time
not
compromising
on
this
usefulness
and
then
a
bunch
of
things
that
we
could
do.
Research
on
one
thing
that
was
mentioned
was
effects
Beyond
energy.
That
energy
is,
of
course,
very
high
profile,
but
that's
not
the
only
thing.
It
is
like
really
serious
effects
from
some
of
the
raw
material
uses
and
and
so
on.
A
So
we
do
need
to
think
about
that,
and
that
probably
needs
more
research
and
or
maybe
that's
available
somewhere,
but
I'm
not
aware
of
it
struck
to
me
as
some
of
this
carbon
Awareness
stuff
might
benefit
from
thinking
about
like
how
can
you
do
that
in
a
sort
of
a
multi-value-
and
you
know
variable
trustworthiness
world
where
you
need
to
learn
something
from
the
other
side
of
the
planet?
But
you
don't
know
if
you
can
trust
it.
How
do
we
do
that?
I
mean
this.
A
This
case
a
useful
thing
to
think
about
and
also
business
model
impacts.
Don't
have
a
lot
of
data
on
that
I.
Think,
for
instance,
what's
the
cost
of
advertising
yeah?
Anybody
else
want
to
add
something
or
take
away
or
dispute
something.
That's
on
the
slide.
A
Yeah
I
don't
see
anybody,
it's
something
to
the
microphone
so
yeah,
maybe
maybe
that's
it
for
this
summary
and
then
calling
perhaps
yeah
you
take
over
and
discuss
the
next
session.
Yeti.
C
One
this
is
Carlos
one,
one
very
quick
way
in
which
I
interpret
you
know.
Some
of
the
early
sessions
is
that
these
two
things
really
number
one
is
that
there's
no
one-size-fit-all,
there's
no
single
Silver
Bullet
for
this,
and
you
know
problem
understanding
as
well
as
solution
is
to
come
in
a
lot
of
different
domains.
Dimensions
areas,
that's
number
one,
and
you
know
number
two
I
feel
that
there's
also
you
know
we
all
bring
just
because
this
is
multidisciplinary.
We
bring
our
preconceptions
on.
C
You
know
what
is
a
high
order
beat
and
I
think
you
know
going
back
to
your
last
point
on
Research.
What
is
really
the
impact
of
actually
making
change
in
particular
areas?
We
all
go
quickly
to
energy
and
to
carbon
and
and
there's
actually
immense
amount
of
data
is
not
the
same
for
a
core
router
done
for
a
mobile
endpoint.
You
know,
even
if
it
is
carbon
contribution
during
the
overall
life
cycle,
so
so
I
feel
that
understanding
is
also.
C
A
D
Thanks
folks,
there
was
one
thing
that
I
kind
of
took
away
from
when
I
was
looking
through
this
one
thing
that
was
possibly
some
research
that
he's
doing
is
to
get
an
idea
of.
D
We
think
that
the
energy
itself
will
be
becoming
less
harmful
in
terms
of
carbon
emissions
for
this,
because
if
you
don't
have
any,
if
you
don't
have
any
assumptions
here,
it's
going
to
be
quite
difficult
to
work
against
the
other
thing
that
might
be
worth
actually
that
I
was
really
thinking
about.
Is
that
other
Industries,
who
have
to
decarbonize
they've,
basically
said
we're
going
to
try
and
decarbonize
by
this
date?
We
think
it's
going
to
cost
this
much
and
that
and
that
works
out
to
be
a
rough
idea
of
how
much
investment
may
be
required.
D
So
for
context
between
now
and
2050,
the
aviation
industry
have
said
it's
going
to
cost
this
many
trillion
dollars.
We
reckon
to
replace
all
the
stock
and
switch
to
a
Greener
Aviation
and
that
works
out
to
be
a
figure
of
like
128
billion
US
Dollars
year
on
year,
between
now
and
2050..
If
you
have
some
numbers
like
that,
then
at
least
it
allows
folks
to
actually
take
a
feed
into
some
of
the
existing
discussions.
D
That's
the
kind
of
the
cop
28
and
the
cop
29s
where
people
say
well,
we
need
this
much
to
actually
stay
on
track
for
what
the
science
has
told
us
to
be
doing
and
I
think
this
is
something
that
would
be
really
really
helpful
when
speaking
to
both
policy
makers
but
also
private
sector
to
say.
Well,
we
need
to
raise
this
money
to
make
this
possible,
even
if
we're
just
going
to
focus
on
training
people
to
use
the
existing
infrastructure
much
more
efficiently.
A
A
really
good
point,
I
agree
that
that's
that's
a
useful
thing
to
have,
and
and
also
that
we
don't
have
it
that
today
on
trying
to
add
these
things,
as
as
we
hear
them
thanks
a
lot.
A
A
B
Yes,
yes,
we
can.
Okay,
so
I
I
can't
see
the
hands
being
erased.
So
if
you
have
questions,
please
do
jump
in
so
session.
Two
was
very
much
about
what
we
know
and
that
there
were
free
talks
initially
by
Michael
Wizzle,
then
Daniel
shine
and
then
Jen's
marmadin
Michael's
talked
focused
on
what
we
know
about
the
energy
usage
and
what
information
exists
about
that
usage?
B
B
B
B
B
He
spoke
about
the
different
factors
influencing
the
energy
usage,
some
discussion
about
how
energy
proportionality
is
pretty
poor
in
the
networks,
how
the
the
energy
usage
doesn't
necessarily
correspond
to
the
amount
of
traffic
and
about
some
of
the
The
Limited
impacts
of
marginal
emissions
and
so
on,
and
carbon
intensity
of
the
energy
Supply
and
its
impact
itself.
B
I
think
some
of
the
interesting
points
to
me
of
this
discussion
were
about
the
the
upgrade
cycle.
The
potential
for
Peak
well
I,
think
mentioned,
was
Peak
shaving
trying
to
shift
or
defer
some
of
the
traffic
to
avoid
the
peak
times
to
avoid
the
usage
getting
to
the
point
where
upgrades
can
be
needed
so
at
least
delay
the
need
to
upgrade
the
infrastructure
and
to
switch
to
a
higher
capacity
and
hence
higher
power
infrastructure.
B
And
then
the
final
talk
was
Jen's.
Talking
about
some
of
the
measurement
data
and
I
think
that
this
one
was
interesting
and
and
perhaps
a
little
bit
controversial
in
the
group
and
his
data
seemed
to
show
that,
while
the
amount
of
data
being
you
used
is
growing
rapidly,
the
energy
consumption
of
data
centers
themselves
has
been
close
to
Flats
and
the
energy
consumption
of
the
networks
is
growing
pretty
slowly
in
comparison
to
the
the
rates
which
is
being
used.
B
B
The
current
emissions
are
still
too
high
and
there's
also
some
concern
that
you
know
we
don't
necessarily
understand
why
the
consumption
is
flat
and
what
what
are
the
engineering
efforts
which
are
causing
it
to
stay
that
way
and
how
to
accelerate
those
engineering
efforts
to
beat
the
curve
so
rather
than
having
flat
or
slowly
growing
energy
consumption
have
declining
energy
consumption
over
time.
B
What
Headroom
we
have
and
how
we
can
yeah
move
on
to
sort
of
beat
the
curve
for
a
rather
than
just
keeping
up
with
the
traffic
and
there's
some
discussion
at
the
end
about
what
we
don't
know,
and
this
seemed
to
relate
to
breaking
down
the
energy
use
by
different
applications,
different
parts
of
the
network
Etc
rather
than
just
overall
aggregate
measures.
So
so
we
can
start
to
understand
which
parts
of
the
system
we
need
to
optimize
to
reduce
the
the
usage.
E
F
So
session
three
was
yesterday
and
we,
oh
maybe
let
me
go
into
full
screen
mode.
F
We
had
five
terrific
talks
and
some
great
conversation,
the
first
of
which
centered
around
the
metrics
Alex
gave
an
incredibly
thorough
assessment
of
the
metrics
needed,
and
then
we
had
Suresh
giving
some
general
thoughts
on
Solutions
and
trade-offs
with
Russ,
following
that
by
including
a
routing,
a
very
detailed
routing
perspective.
Louise
spoke
to
us
about.
F
The
prospects
were
multicast
and
Brendan
spoke
about
data
formats
and
I
felt
there
were
many
takeaways
in
the
discussions,
just
a
few
of
which
I've
highlighted
here
as
I
stated,
I
thought
that
Alex
Alex
has
talked,
did
a
great
job
to
sort
of
categorize
the
metrics
needed
for
devices
paths
flows
and
the
overall
Network
and
pointed
to
some
concrete
next
steps
that
are
available
to
us
to
get
started
like
Yang
models
and
protocol
extensions.
F
The
second
talk
was
interesting
because
it
very
clearly
pointed
out
to
us
that,
while
we
focus
on
Scopes
one
and
two
that
scope
three,
which
is
the
usage
of
all
the
equipment
that
many
of
our
companies
produce,
if
that
far
eclipses
the
totals
and
the
sum
of
the
totals
of
Scopes
one
and
two,
we
should
also
consider
some
inclusion
of.
F
What's
being
what's
emergent
as
a
scope
for
which
is
kind
of
a
handprint
effect,
which
is
what
might
be
the
savings
of
using
technology
towards
others
Footprints,
so
that
was
kind
of
an
interesting
perspective.
I
thought
and
another
point
that
was
made
was
around
how
there's
so
many
sustainability
best
practices
that
perhaps
we
should
begin
to
collect
them
into
a
document.
There's!
No!
F
You
know
we
should
start
doing
that
sort
of
straight
away
when
we
got
to
the
presentation
about
about
routing
what
was
interesting
was
you
know,
Russ
pointing
out
to
us
that
merging
it's
really
important
for
us
to
merge
multiple
metrics
into
one,
because
when
we
start
to
do
these
joint
optimizations,
it's
a
well-known
NP,
complete
problem
to
have
multiple
metrics,
and
so
the
convergence
into
one
metric
is
something
we're
going
to
have
to
take
strategically.
F
It
was
also
interesting
to
hear
from
the
control
plane
perspective
the
levers
we
have
to
either
reduce
to
remove
redundant
links
or
remove
equipment.
Or
is
there
a
question?
F
Okay,
anyway,
that
we
have
several
levers
for
reducing
power
usage
and
all
of
which
can
be
done
in
a
manner?
That
is
time
variant,
which
was
a
topic
that
came
up
multiple
times
yesterday
and
in
previous
sessions
around
you
know.
If
we
know
beforehand
apriori
that
links
will
come
and
go
that
we
can
schedule,
we
can
have
a
more
robust
routing
algorithm
that
can
schedule
its
knowledge
to
match
the
coming
and
going
of
predictable
behaviors
about
these
links.
F
It
was
interesting
to
get
an
update
on
multicast
that
it's
perhaps
time
to
reconsider
its
usage,
because
in
the
past
we
have
focused
on
the
Simplicity
of
the
algorithm
rather
than
the
efficiencies
of
the
algorithms,
and
that
many
of
the
problems
that
were
considered
old
challenges
have
new
Solutions,
most
importantly,
creating
a
stateless
multicast.
F
So
router
state
was
a
problem
in
the
past
for
the
wide
area
Networks,
and
also
that
the
performance
of
implementing
multicast
in
user
space
is
now
competitive
or
is
at
least
reasonable
enough,
then,
and
as
compared
to
say,
being
required
to
do
this
in
kernel.
Space
and
I
also
found
it
very
instructive
that,
in
multiple
of
the
talks
yesterday,
there
was
the
emphasis
that
we
should
be
reusing
techniques
that
we
have
experience
with.
F
From
these
more
constrained
contests
constrained
networking
context
as
well
as,
almost
importantly,
you
know
the
wireless
context
where
we
do
have
to
already
consider
low
power
and
that
we
should
remember
that
our
the
very
edges
of
our
topology
Are
Wireless,
it
was
fascinating
also
to
hear
about
data
formats,
I
think
who
knew
that
that
there's
typically
a
30
energy
reduction
when
using
binary
over
text.
It
was
good
to
see
the
quantification.
F
You
know
intuitively.
You
would
assume
that,
but
it
was
good
to
see
a
number
I.
Also
appreciated
that
we
heard
you
know
we
were
courage
to
stop
thinking
of
the
network
as
an
infinite
resource
and
therefore,
along
with
that,
stop
ignoring
data
formats,
because
even
if
you
have
these
small
reductions
in
the
aggregate
they
add
up
and
the
example
was
given
that,
in
contrast
to
video,
which
we
know
is
the
main
type
of
traffic
on
the
network
and
accounts
for
a
large
percentage,
maybe
90,
plus
of
traffic
on
the
network.
F
We
still,
we
send
emails
all
the
time
and
the
what-ifs
of
if
we
had
a
super
representation
of
mime
for
all
those
mime,
encoded
emails
that
we're
sending
every
day
many
times
a
day,
and
there
was
also
some
in
the
discussion
at
the
very
end
of
the
presentations
concession
that
we
recognize
that
at
least
the
current
state
of
monitoring
is
not
very
efficient
in
and
of
itself.
But
we've
got
to
start
somewhere
and
that
we,
you
know
it's
important
for
us
to
understand
approximately
where
energy
is
consumed.
F
To
understand
where
the
high
impact
place
is
for
us
to
focus
our
attention,
including
on
monitoring
itself.
Eventually,
those
are
some
of
my
takeaways,
but
I
would
welcome
others
to
jump
in
and
contribute
their
impressions.
A
I,
don't
see
any
raised
hands
either.
I
just
had
a
one
comment
myself
that
that
that
that
the
TVR
type
of
time
varying
action
is,
in
my
mind
at
least
it's
sort
of
a
part
of
a
bigger
thing,
or
at
least
a
related
thing
that
that
is
about
ability
to
handle
things
on
a
particular
time
and
then
sleep
at
other
times,
and
that's
not
just
a
routing
thing.
A
It's
it
sort
of
goes
across
implementation,
Technologies
and
and
and
protocols
and
design
of
Link
layers
and
details
of
linked
layers
and
so
on.
So
in
4G
and
5G,
for
instance,
there's
some
some
details
that
lead
to
very
different
power
usage,
and
some
savings
can
be
done
in
5G
that
were
not
possible
in
40
because
of
these
things.
A
So
maybe
that's
the
thing
that
could
be
also
noted.
Okay,.
G
Was
just
curious,
I
liked
what
I
heard
when
you
said
that
documenting
best
current
practices
is
sustainability
is
was
one
of
the
conclusions
but
I
didn't
see,
the
new
slide
did
I,
miss
it
or,
and.
B
F
G
Okay,
if
you
need
volunteers
to
help
out
to
that,
I
can
help
absolutely.
H
I
H
The
point
it
was
like
not
just
to
do
the
the
best
practices,
but
also
talk
about
the
trade-offs
right,
I
think
that
was
like
you
know.
One
thing
I
kind
of
wanted
to
add,
because
there
might
be
other
things
that
for
people
to
consider
so,
like
not
just
say
like
hey,
this
is
good,
and
this
is
what
you
need
to
kind
of
I
would
say,
balance
it
against
right,
so
I
I
think
a
lot
of
the
things
we
saw
have
some
kind
of
trade-offs.
H
It's
not
like
always
like
a
green
thing
is
event.
Always
right,
so
I
think
that's
something
I
kind
of
wanted
to
emphasize
a
little
bit
on,
but
I
think
Eve,
like
summarized
it
extremely
well
right
like
and
yeah
I
would
love
to
have
some
help
on
that.
If
people
want
to
help
black
out
certainly
help
writing
something.
E
F
J
A
F
And
I
will
say
that
you
know
I
I
tried
my
best
to
put
it
all
on
one
slide,
as
requested
and
I'll
try
not
to
linger
on
the
things
that
we
have
stated
a
lot,
but
suffice
to
say
you
know
there
was
lots
of
discussion
on
the
whether
energy
usage
should
be
increasing,
remaining
the
same
or
being
reduced
and
I
wanted
to
go
back
to
sort
of
the
urgency
of
some
of
the
organizations
like
the
UN
ipcc,
the
intergovernmental
panel
on
climate
change,
world,
the
resources
Institute
that
you
know
when
you
look
at
the
sixth
assessment
that
came
out
over
the
last
year,
it
makes
it
very
clear,
we're
sort
of
in
code
red.
F
You
know
we
all
of
the
advices.
You
know
this
is
the
decade
where
we
need
to
make
a
difference,
and
although
there
are
these
goals
to
say
you
know,
50
reductions
by
2030
and
vest
is
more
even
quicker
reductions
by
10
per
year.
These
are
recommendations,
but
everything
points
to
that.
F
We
need
to
front
load
this
decade,
so
we
should
be
assertive
and
aggressive
and
remember
the
urgency
of
this
task
so
that
that's
what
I
wanted
to
say
there
amidst
all
the
emails
that
are
flowing
about
that
for
carbon,
aware
networking,
we
we
want
to
take
all
of
these
good
metrics
that
we
already
have
around
performance
of
our
of
the
internet
of
ICT
of
networks
specifically,
and
we
want
to
augment
them,
and
in
particular
we
we
gravitate
to
energy
because
it
feels
like
it's
in
our
wheelhouse.
It's.
F
You
know
it's
kind
of
next
of
kin
for
performance
metrics
that
we've
typically
focused
on
at
least
for
the
core
Network.
We
know
that
in
the
wireless
realm
you
know
it's
very
much
front
and
center,
but
let's
augment
them
with
more
information
about
carbon
and
more
environmental
metrics,
and
one
of
the
phrases
that
my
colleague
and
co-author
don
memphas,
as
often
said
was.
We
cannot
Energy
Efficiency
our
weight
in
at
zero
greenhouse
gas
emissions
by
2050,
which
is
also
one
of
the
recommendations
which
some
have
said.
F
You
know
that's
very,
very
late
in
the
game,
but
even
so
we
can't
we.
We
need
to
have
the
tandem
task
that
we
need
to
be
looking
at
Not,
Just,
Energy
reductions,
but
carbon
reductions
and
other
environmental
impact
reductions.
So
this
is,
of
course,
just
focusing
on
the
carbon
part
of
it,
but
we-
the
first
task
we
need
to
do-
is
to
deliver
on
Telemetry
extensions
so
that
we
can
assess
the
pain
points
that
are
happening.
F
We
need
to
make
those
the
Telemetry
real
time
and
we
need
to
quantify.
Maybe
as
we
go
along
I
mean
we
want
to
understand
what
we're
using
and
also
what
we're,
how
we're
reducing-
and
this
has
to
happen
at
the
platform
level,
the
component
level-
it's
not
sufficient
to
just
be
doing
this
for
CPUs.
F
After
all,
we
really
want
to
help
out
applications,
so
we
need
to
on
a
thread
by
thread
basis,
understand
what
is
our?
What
is
the
impact
of
applications,
and
that
really
was
the
thrust
of
this
Workshop,
or
at
least
the
title
for
this
Workshop
is
the
internet
applications
and
services.
So
we
need
to
get
to
that
level
of
specificity,
and
we
also
need
to
consider
and
track
secondary
effects
like
Cooling
and
so
forth.
F
In
monitoring
electricity
consumed
and
carbon
intensity,
which
I
both
I've
put
them,
both
underneath
the
umbrella
of
carbon
intelligence,
we
need
to.
We
want
to
to
Monitor
and
track
them
at
the
end,
points
the
sources
and
destinations,
but
also
we
want
to
leverage
in-network
Telemetry
to
understand
the
hop
by
hop
assessment
as
well,
and
we
know
that
there
that
all
of
our
constituents
who
happen
to
be
actors
like
the
developers,
the
users,
the
stakeholders,
The
Operators.
F
And
so
it's
a
somewhat
of
a
sorrowful
state
that
we
find
some
of
our
tools
like
trace
route,
which
used
to
help
us
to
understand
by
Hound
accounting
of
other
kinds
of
metrics.
Wouldn't
it
be
great
to
have
a
tool
that
allowed
us
out
of
band
to
look
at
the
network
and
have
some
observability
of
it
with
regards
to
carbon,
both
energy
and
carbon,
and
in
the
earlier
talk
on
in
the
first
session,
we
did
outline
more
about
you
know.
F
The
end
goal
is
many
different
facets:
kind
of
a
toolkit
of
carbon
awareness
throughout
the
stack
and
one
of
the
cultural
I
guess
alignment.
The
cultural
alignment
that
we
have
with
the
ietf
is
that
we
have
many
working
groups
in
which
we
could
probably
be
doing
some
of
this
work,
both
at
the
routing
and
transport
levels
for
traffic
engineering
and
so
forth.
F
But
again,
I
want
to
go
back
to
the
fact
that
we
need
to
sort
of
accelerate
our
efforts,
accelerate
deployment,
and
to
do
that,
we
really
need
to
have
some
rigorous
testing
and
certification
potentially
so
that
we
can
attest
to
the
measurements
that
we
get
or
even
the
attest,
to
the
approximations
that
we
can
account
for
the
and
and
understand
how
how
we
audited
the
network,
and
we
need
to
be
able
to
scale
this
up.
F
I
worry
that,
culturally,
we
don't
have
the
apparatus
in
place,
certainly
not
in
the
ITF
for
things
like
certification
or
validation,
to
the
degree
of
record
that
we
might
need
it,
and
so
that
that's
an
open
for
us
to
consider
who
might
be
our
partners,
which
is
exactly
what
I
say
in
this
last
point,
which
is
it's
become
very
clear
in
all
of
the
discussions
that
we
need
to
be
collaborating
across
sort
of
all
corners
of
the
internet
and
those
who
have
a
stake
in
it,
and
especially
those
who
are
working
hard
in
other
com
standards.
F
Organizations.
We've
heard
about
the
itu.
We
we
have
heard
about
you
know
like
their
their
other
organizations
in
the
mobile
space.
I
suspect
will
hear
about
some
of
them.
There
was
a
great
report
that
just
came
out
this
week
from
the
ngmn
about
commitments
in
the
mobile
space.
You
know,
so
we
need
to
be
partnering
to
make
sure
that
we've
we
have
and
I
think
it
was
prunilla
who
suggested
we
needed
a
gap
analysis.
F
Others
saying
we
need
best
practices
and
even
shared
terminology,
so
we're
speaking
the
same
language
I
feel
I'd,
be
remiss
if
I
didn't
underscore
that
we
really
need
to
understand
the
socio-technical
implications,
because
we
want
to
affect
change,
but
we
want
to
affect
the
right
kind
of
change
which
in
turn,
are
going
to
affect
the
policies
that
are
created
by
governments
and
then
also
we
want
the
technology
we
create
to
be
adopted,
and
so
we
need
to
understand
what
motivates
people
to
care
and
to
adopt
technology.
If
we.
F
F
There
was
an
interesting
paper.
I
would
recommend,
reading
that
Bruce
Nordman
submitted
about
how
could
we
influence
the
re-architecture
of
this
electric
grid,
drawing
upon
what
we
know
from
the
internet?
F
Understanding
pricing,
understanding
business
models,
because,
after
all,
the
electric
grid
is
going
to
have
to
supply
four
times
as
much
electricity
as
it's
producing
now,
in
order
to
make
even
the
transition
just
to
the
electrification
of
transportation
and,
as
we
know,
that's
going
to
imply
massive
disruption
to
the
infrastructure
and
there's
this
opportunity
to
affect
change
there
and
to
be
involved
there.
So
there
you
have
it.
Thank.
A
I
think
we
sort
of
out
of
time
for
questions.
Maybe
perhaps
we
should
move
on
to
the
next
talk,
Suresh
or
Carlos,
which
one
of
you
is
doing.
It.
H
A
H
A
Good
yeah
thanks
so.
H
Just
like
Gary
and
calling
us
to
summarize
like
pretty
much
the
day,
three,
the
the
improvements,
the
the
implementation
technology
and
protocols
so
I
think
like
a
lot
of
this,
is
going
to
be
overlapping
with
what
Eve
said,
but
I
just
do
want
to
put
up
the
silly
and
things
that
came
up
not
only
on
day
three
but
also
before
that's
related
to
like
you
know
what
we
can
do
pretty
quickly
on
the
implementation
and
Technology
side.
So
one
thing
that
kind
of
came
up:
a
lot
of
talks
was
like
you
know
they.
H
We
need
to
measure
like
what
we
have
to
do
like
you
know
how
to
improve
things.
So
I
think
the
metrics
and
networking
was
like
pretty
high
up,
like
you
know,
quite
a
few
like
you
know,
Alexa
stock
or
talk
like
you
know,
even
the
like,
you
know
earlier,
like
you
know,
I
think
Eve
stock,
like
you
know,
a
lot
of
it
actually
focused
on
how
to
measure
things
and
one
of
the
key
things
was
like.
How
do
you
standardize
this
terminology
and
what
is
getting
measured
so
like
we
don't?
H
We
all
talk
about
the
same
thing
when
we
are
comparing
different
products
or
different
networks,
and
so
on
to
how
to
talk
the
same
language
and
and
I
think
this
is
another
key
thing
like
it
came
again
in
multiple
talks
like
and
how
do
we
improve
non-networking
things?
H
So
we
are
networking
people
we
can
improve
that,
but
and
there's
like
two
aspects
like
now:
how
do
we
kind
of
take
care
of
like
our
value
chain
like
now
where
things
are
coming
from
so,
like
you
know,
I
think
Alex
was
talking
a
little
bit
about
you
know.
How
do
we
actually
like
look
at
the
power
and,
like
you
know,
on
on
the
chat
side,
yeah
and
Marisol
are
talking
about
it
as
well.
Like
you
know,
how
do
we
actually
take
the
whole
product
life
cycle
into
consideration?
H
Right,
like
you
know,
how
do
we
take
the
packaging,
the
modularity,
and
we
also
talked
about,
like
you
know
how
things
can
be
like
more
reusable
right
like
how
can
things
a
bit
more
flexible
at
the
cost
of
like
having
a
longer
life,
product
and
I?
Think
it's
like
I
think
also
dovetails
into
what
Yari
said
a
little
bit.
H
You
know
how
do
we
like
stick
with,
like
you
know
something,
that's
good
enough,
let's
go
forward
right
and,
and
these
all,
without
fail
into
having
slides
a
little
bit
more
flexibility
built
in
if
it
doesn't
cost
too
much
right
and
and
kind
of
like
keep
things
for
longer,
and
the
other
angle
was
really
the
energy
of
our
routing
staff.
So
there's
like,
like
simple
things
like
you
know,
picking
more
energy
efficient
routes.
H
If
you
have
the
information
and
also
time
variant,
routing
that,
like
you,
know,
Russian
alvarovs
like
talk
us
about
so
there's
like
kind
of
compromises
in
there
again
as
well
and
I,
think
that's
like
a
theme,
I
kind
of
want
to
summarize
it
at
the
end.
But
the
theme
was
like:
okay,
like
you
know,
if
you
start
doing
stuff
like
you
know,
you
know
turning
off
links
or
turning
off
routers
and
so
on.
H
Like
you
know,
the
convergence
like
kind
of
takes
a
hit
right
like
so
it's
not
going
to
be
like
a
zero
cost
to
do
this
and
I
think
like
earlier
talks,
I
think
it
was
Daniel
who
was
talking
about
Duty
cycling
things
right
like
that,
so
it
kind
of
fits
into
the
same
thing,
so
the
PVR
can
either
be
due
to
constraints
in
the
the
technology
itself
could
be
something
that
you
actually
want
to
do,
which
is
Duty,
cycling
and
I.
H
Think
another
concept
that
came
off
was
like:
is
it
possible
to
timeshift
demand?
So
this
came
up
in
like
multiple
things
like
you
know?
How
do
we,
like
kind
of
like
you,
know,
not
hit
the
Peaks
right
because
the
the
function
of,
like
you
know
the
the
Network's
spending
energy
is
not
like
pretty
it's,
not
linear.
H
So
there's
like
whole
bunch
of
like
discontinuities
and
how
do
we
actually
keep
stuff
by
time
shifting
demand,
so
we
don't
have
to
grow
as
quickly
or
wait
till
the
I
think
like
it
was
vesnar
who
talked
about
a
little
bit
right
like
wait
for
the
improvements
to
catch
up
right
like
before.
We
do
the
upgrades
to
the
network
capacity
and
on
the
multicast
side
again
like
there's,
like
kind
of
I
I
heard
like
two
different
things
that
are
like
kind
of
opposite
to
each
other.
H
So,
like
you
know,
how
can
we
use
multicast
to
improve
things
right,
like
you
know,
kind
of
like
CDN
distributions
and
stuff
and
so
on,
but
also
like
stuff
on
the
side?
Like
you
know,
kind
of
Pascal
talking
about
like
you
know
how,
like
you
know,
multicast
affects
like
a
lot
of
the
constraint
networks,
so
I
think
it's
kind
of
again
boils
down
to
this
kind
of
trade-off
thing
and
so,
following
on
into
Data
formats
like,
and
this
is
like
really
good
presentation
as
well.
H
Like
you
know,
the
c
bar
Json
kind
of
comparison
and
there's
like
two
things,
I
kind
of
came
off
from
it
like
one
of
them
is
the
readability
versus
efficiency
trade-off
like
have
to
be
made
sometimes
like.
If
you
really
compress
things,
is
it
gonna,
make
it
less
readable
or,
and
does
it
matter
right,
I
think
that's
something
we
need
to
think
about
and
also
how
much
is
the
magnitude
of
the
impact.
So
we
talk
about
like
30
right,
like
you
know,
what
is
that
30
off?
H
So
if
it's
just
going
to
be
a
manifest
for
some
video,
that's
already
compressed
like
you
know,
how
much
does
it
save
and
going
up
to
like
high
level
thoughts?
I
think
this
is
like
an
interesting
thing
that
came
up
in
in
chat
and
I.
Think
Rob
talked
about
it
a
little
bit.
Is
there
a
way
for
us
to
kind
of
share
the
energy
impact
to
the
user?
H
So
I
think
this
is
like
a
very
difficult
problem,
but
it's
a
very
interesting
problem
for
us
to
look
at
so
it's
kind
of
like
a
longer
term
thing.
How
do
we
actually
show
the
users
like
what
they
are
like
really
resulting
in
energy
impact
right
and
would
that
change
user
Behavior?
That's
something
to
kind
of
get
to,
and
and
the
follow-up
to
that
is
is
good
enough
good
enough.
H
So
is
there
like
you
know
if
if
people
are
happier
with
like
you
know,
lower
definition,
video
or,
like
you
know,
lower
latency
or
like
lower
availability
like
should
be
able
to
push
towards
that
rather
than
trying
to
always
maximize
on
all
these
axes,
it's
another
thing,
high
level
we
need
to
think
of,
and
another
interesting
thought
that
came
up
like
mostly
in
the
chat
and
discussions
was
like.
If
we
design
for
constraint,
is
it
going
to
be
equally
good
for
non-constraints?
H
So,
like
you
know,
kind
of
think
of
the
most
constrained
environment
like
the
most
energy,
restrictive
and
design
for
that
and
and
try
to
reuse
them
for
non-constraint,
because
it
should
work
right.
It's
like
a
like
a
hypothesis.
We
need
to
kind
of
test
if
that's
true
or
not,
right
and
and
the
last
thing
I
want
to
kind
of
leave
with-
and
this
is
kind
of
I
talked
to
it
earlier
about
this.
H
H
Yeah
so
I'm
I'm
done
with
the
diary
so
I
kind
of
like
rush
through
it.
So,
like
you
know,
if
you
have
time,
I
can
take
questions.
A
Yeah
I'm
almost
off
the
opinion
that
we
go
to
the
end
of
the
presentations
and
then
we
discuss
so
Beatrice
I
think
you're.
Next.
E
K
Okay,
I'll
start
with
the
where
we
are.
We
are
currently
shifting
from
the
monolithic
architecture
of
4G
going
towards
the
service-based
one
of
5G,
in
which
network
elements
are
natural
functions
which
are
usually
virtual
machines
or
containers.
K
K
K
The
next
steps
that
we
have
to
take
I
think
start
with
observability
because
with
as
we
said
earlier,
we
can't
control
what
we
don't
measure
so,
especially
here
I
think
standardization
bodies
such
as
ietf
can
make
a
difference.
We
need
first
of
all
to
Define
which
metrics
to
consider
and
then
also
we
have
to
relate
these
metrics
to
individual
vertical
applications.
This
is
a
non-trivial
task,
since
we
have
in
5G
at
least
one
virtualization
layer.
K
Once
we
have
the
metrics,
we
can
make
the
network
energy
or
carbon
aware,
and
we
can
exploit
this
awareness
in
order
to
buy
incentivizing
economically
the
stakeholders
who
consume
less.
For
example,
another
use
of
the
metrics
that
we
have
to
do
is
the
optimization,
and
we
see
the
trend
of
Energy
Efficiency,
optimization
and
optimization
in
general
is
following
artificial
intelligence.
Now,
in
the
position
papers
that
were
presented
in
this
Workshop
I
found
four
proposed
Solutions.
K
Regarding
the
Energy
Efficiency
of
mobile
networks,
the
first
one
was
proposed
by
a
paper
from
Anonymous
authors,
who
propose
a
method
to
collect
and
calculate
the
end-to-end
Energy
Efficiency
of
edge
Services,
based
on
the
traffic
paths
from
end
users
to
Edge
service
replicas,
the
other
three,
the
other
three
are
present
in
our
position
paper,
and
if
you
want
further
details,
you
should
you
can
find
them
there.
K
Regarding
observability,
we
propose
the
energy
aware
back
pressure
whose
aim
is
to
collect
energy
kpis
at
different
levels,
such
as
that
vertical
application,
Level
at
slice
level
and
also
at
overall
Network
level,
and
to
make
all
the
stakeholders
aware
of
such
kpis.
K
Finally,
the
last,
the
last
method
we
propose
is
green
elasticity,
which
consists
in
the
dynamic
and
adaptive
Hardware,
assisted
offloading
of
network
functions
and
or
vertical
application
components
in
case
of
a
very
heavy
workloads
with
a
very
high
requirements
in
terms
of
latency,
for
example.
Thank
you.
If
you
have
any
questions,
please
ask
them.
L
Yeah,
do
you
see
my
screen,
and
do
you
hear
me
yes
good?
So
this
is
a
presentation
trying
to
summarize
what
we
should
not
do.
I'm
trying
to
summarize
what
has
been
mentioned,
both
in
the
paper
and
in
previous
discussions
and
a
lot
of
things
have
been
about
energy
inefficiency
as
several
people
talking
about.
We
should
not
use
bad
sources
of
energy
with
some
different
definition
of
what
is
green,
and
then
we
have.
We
should
not
do
Luxury
consumers
more
basically
unnecessary
things
that
are
not
needed
should
not
use
systems.
L
They
use
a
lot
of
resources
either
when
you
create
them
like
water,
land,
minerals
or
create.
A
lot
of
e-waste
should
not
do
digital
consumerism.
This
might
be
very
bad
condition
during,
for
example,
mining
of
things
or
much
pollution
during
the
production,
or
it
might
be
e-based
at
a
7-4.
Even
when
it's
formally
stand
for
recycling
should
not
do
things
that
have
a
negative
impact
on
society
should
not
use
numbers
that
don't
add
up
a
little
bit
different
aspects.
L
You
should
not
send
too
many
acts
or
in
general,
not
some
too
many
messages
and
we
should
not
fly
which
use
ICT
systems
instead,
some
slides
on
specific
topics-
Energy
Efficiency,
has
been
discussed
a
lot.
A
clear
blue
hanging
fruit,
for
what
not
to
do
here
is
crypto
assets
with
proof
of
work.
They
use
energy
like
a
small
country.
They
could
also
be
seen
as
having
a
negative
impact
and
being
unuseful.
L
L
Aha,
you
did,
but
here
you
do
you
hear
me
now:
yeah,
yes,
okay,
I,
I.
Yes,
another
discussion
is
to
not
use
bad
energy
sources.
People
has
what
should
we
do?
What
is
bad
and
what
is
what
we
should
use
instead
has
been
different,
green
and
clean
and
might
not
be
well
defined.
Renewable
might
be
a
little
bit
of
Target
and
maybe
not
enough
might
be
better
to
be
more
concrete
and
I.
Don't
use
systems
that
energy
sources
with
large
greenhouse
gas
emissions.
L
Other
things
that
are
bad
with
with
energy
sources
is
maybe
the
second
largest
bad
thing
is
maybe
air
pollution.
Then
you
have,
of
course,
other
bad
things
that
can
be
bad
with
energy
sources,
also
like
use
of
resources
or
other
forms
of
pollution
and
and
waste,
and
has
been
discussion
that
we
should
have
carbon
and
we're
networking.
Maybe
that
can
handle
other
aspects
in
the
future.
L
Also,
if
that
is
developed,
this
is
a
bit
hard
for
the
IDF
to
control
it's
easier
to
control
the
first
aspects,
especially
about
protocols
how
any
efficient
they
are,
but
in
general
we
should
consider
life
cycles
of
the
systems
as
as
a
whole,
then
my
last
slide
is
about
E-Waste
over
consumption.
Digital
colonialism
related.
L
As
over
consumption
in
the
production,
production
often
is
in
third
world
countries,
while
the
products
are
used
in
the
in
the
industrial
rich
countries,
and
the
same
is
true
with
e-based
a
lot
of
E-Waste
even
understand
for
recycling
in
the
western
bird,
they
end
up
in
a
non-responsible
way
in
developing
a
countries,
so
we
should
avoid
things
that
produce
a
lot
of
E-Waste.
At
least
that
cannot
be
recycled
in
a
responsible
way.
G
M
M
M
This
I
added
kind
of
the
and
towards
sustainability.
That
was
my
my
contribution
to
the
workshop
earlier
and
in
I
have
actually
been
on
this
slide.
I
have
just
questions:
I,
don't
have
a
single
answer
and
a
single
proposal
for
you
for
everybody,
but
rather
have
been
following
the
discussion
on
the
list
and
in
these
meetings
and
and
made
some
notes
on
on
that
on
on
the
sustainability,
IDF
and
all
so
not.
M
J
M
If
we
go
a
bit
more
into
the
to
the
topic,
then
I
have
two
slides
of
early
questions,
because
I'll
be
working
on
this.
Let's
say
Energy,
Efficiency
and
and
sustainability
Topic
in
ICT
for
over
10
years,
and
it's
always
the
kind
of
challenge
in
what
is
coped
and
what
is
actually
what
people
see
as
the
scope
and
what
they
are
looking
at.
For
example,
energy
has
been
mentioned.
Many
times
are
we
talking
about?
Are
we
measuring
electrical
energy?
M
Are
we
measuring
you
know,
I,
see
this
use
of
all
possible
energy
sources
on
this
planet,
I'll
be
talking
about
renewable
energy
or
just
energy
usage
as
a
whole,
for
example.
If
everything
is
renewable
in
the
future,
is
the
problem
solved
as
they
did
in
the
previous
slide,
because
you
could
see
that
you
know
there's
no
such
thing.
M
As
you
know,
emission
free
energy
there
are
less
than
less
and
more
energy
kind
of
emissions
from
an
energy
source,
but
all
energy
sources
need
to
be
built
and
deployed,
and
so
forth
or
I'll
be
talking
about
embodied
energy
or
just
the
operational
side.
I've
seen
many
discussions,
you
know
in
this
last
week
on
on
the
topic
or
I'll,
be
talking
about
carbon
emissions
or
energy
carbon
emissions.
M
Are
we
talking
about
corpse
one
two
three,
and
there
was
scope
four
might
mentioned
today,
even
so
this
seems
to
be
totally
dependent
on
who's
talking
and
or
are
we
talking
about
coming
back
to
the
you
know,
embodied
energy
and
are
we
talking
about
and
actually
the
previous
presentation
had,
this
you
know
waste
and
so
forth,
so
I'll
be
talking
about
the
whole
life
cycle
of
ICT,
of
equipment
of
systems
of
Networks
or
just
you
know,
something
related
to
operational.
Operational
side
seems
to
be
that
people
are
talking
about
different
things
and.
M
Talking
about
you
know,
for
example:
Hardware,
then
again,
is
it
just
operational?
Is
it
the
whole
life
cycle?
That's
been
these
discussions
on
the
on
the
mailing
list
on
this
idle
consumption
or
or
Hardware
vendors
implementing
you
know
energy,
proportional
systems
or
they're
load-based,
or
is
there
law
based
consumption
or
everything
is
flat,
so
it
doesn't
really
matter
whether
you
are
sending
one
bit
or
10
gigabits.
M
M
Me
it's
an
open
question
and
my
last
slide
again
continue
the
same
Chain
of
Thought
or
flowing
out
or
are
we
talking
about
on
one
hand,
Hardware
on
one
hand,
software,
or
do
we
be
discussing,
for
example,
some
implementation
guidelines?
Some
you
know
ways
to
do
more.
Clever
implementations,
more
energy,
efficient
or
carbon
saving,
depending
on
depending
on
your
metric
and
kpi,
so
is.
Would
that
be
in
scope?
I
don't
have
the
answer
just
raising
out
questions
that
I
see
are
need
to
be
discussed
or
are
we
simply
talking
about
architectures
and
their
impacts?
M
How
different
things
are
there
was,
for
example,
these
different
ways
of
routing
routing
have
been
mentioned
many
times.
So
is
that
something
that
would
be
a
scope
or
not
in
scope?
Or
are
we
maybe
focusing
on
protocol
level
things
again?
You
know
you
could
you
could
you
could
do
one
or
both
or
any
any
combination,
then
coming
back
to
the
scope
for
and
in
general
I
was
simply
talking
about
footprint
or
often
people,
especially
in
the
in
the
let's
say.
Ict
sector
in
the
business
side
often
wants
to
Advocate
the
handprint
side.
M
J
M
Opinion
the
open
question,
then,
are
we
talking
about
simply
short-term
things?
You
know
just
use
renewable
energy
and
problem
solved,
or
are
we
talking
about
long-term
questions,
for
example
just
using
a
less
absolute
amounts
of
energy,
because
all
energy
sources
need
to
be
built,
so
it's
not
necessarily
simply
about
just
use
Renewables,
and
you
know,
problem
solved
or
or
or
things
like
that,
so
in
that
sense,
and
our
energy
sources
in
general
are
even
in
the
its
scope
to
consider.
You
know
what
energy
sources
companies
are
using.
M
I
mean
there
are
countries
who
have
you
know
very
little
renewable
sources
and
then
there's
countries
like
nowhere
which,
as
far
as
I
understand,
was
99
Renewables.
So
that's
no
way
need
to
do
anything
because
they
are
running
on
Renewables
so
again
and
at
the
end
it
seems
to
be
that
you
know
the
final
question:
is
there
probably
in
the
first
place,
to
consider
by
the
IAP
or
ietf,
because
on
100,
if
you
look
at
these
different,
you
know
Trends
and
Trends
and
and
kind
of
views
to
the
Future.
A
A
Some
of
some
things
are,
of
course,
outside
the
scope
of
ietf
to
make
improvements
in,
but
we
still
probably
need
to
understand
that
those
things
are
happening,
but
I'll
open,
a
fraternal
discussion
now
and
yeah
like
conclusions
and
proposals
for
next
steps
and
and
such
I
I
will
take
at
the
end
a
little
bit
of
time.
For
my
view
where
we
are
I,
took
some
notes
during
the
session,
but
go
ahead.
A
J
A
This
is
not
we're
not
alone
in
that
respect,
but
if
anybody
from
The
Operators
actually
wants
to
speak
up
and
and
is
present-
and
please
please
speak
up-
we
do
did
have
some
some
data
from
them,
but
it's
always
the
yeah
I
work
for
a
vendor.
We
try
to
innovate
on
stuff,
but
somehow
sometimes
we're
too
eager
be
much
better
to
hear
from
the
people
who
actually
run
things.
N
Is
if
you
already
have
any
ideas
if
or
how
we
should
continue
the
discussion
outside
of
of
your
answer?
I'd,
certainly,
you
know
been
seen
welcoming.
N
You
know
people
in
the
Ops
area
working
group
just
when
there
are
things
that
can
be
put
into
the
form
of
ietf
submissions
which
may
or
may
not
be
appropriated
to
continue
the
discussion
best,
but
yeah.
So
I
think
you
know
short
of
really
talking
about
subjects
right
now.
I
was
wondering
about
that
primarily.
A
A
Some
of
the
things
that
being
talked
about
will
progress
independently,
probably
does
some
metrics
work,
TBR
working
group
will
will
go
ahead
and,
and
many
other
things
will
happen
either
at
the
ITF
or
elsewhere
say
implementation
improvements,
I
think
we
should
also
continue
this
a
bit
more
widely.
A
Maybe
there's
some
other
things
that
we
should
be
doing,
maybe
get
the
Ops
area
involved.
I
think
we
should
continue
the
good
discussion
on
the
mailing
list
in
the
in
this
Gathering
and
perhaps
open
it
up
for
everybody
once
design.
N
Yeah,
if
that
can
be
transformed
into
an
open
mailing
list
with
you
know,
with
an
archive
or
so
that
would
certainly
be
the
lease
overhead
at
least
controversial
approach.
N
The
the
the
the
form
that
I
did
like
in
in
the
iitf
of
course
are
I.
Think
I'd
already
said
that
before
the
special
interest
working
groups,
which
are
not
that
much
tied
to
the
deliverables
that
we
usually
expect
from
working
groups
or
research
groups,
but
there
are
more
to
to
bring
communities
together.
A
J
D
Hi
FX.
Thank
you.
There
was
one
thing.
A
reflection
I
wanted
to
share
that
is
probably
unique
or
interesting.
Specifically,
the
network
context
is
that,
among
all
the
papers,
I
saw
I
think
Beyond
a
reference
from
Bruce
Norman's
paper.
Talking
about
a
power
price
index
as
I
want
a
way
to
kind
of
communicate
things
like
the
cost
and
carbon
intensity
of
electricity.
D
The
looking
for
power
and
different
prices,
different
kind
of
carbon
intensity
in
different
parts
of
the
world
and
I
feel
like
because
the
internet
has
a
kind
of
different,
unique
different
economic
model,
with
ideas
like
peering
and
different
stewardship
of
resources,
I
figured
that
might
be
quite
a
fertile
ground
for
research
that
other
sectors
and
other
places
might
not
be
so
well
placed
to
do
where
there's
probably
a
chance
for
the
ITF
to
share
some
really
groundbreaking
stuff
that
could
be
adopted
in
lots
of
places.
D
A
Sounds
like
an
exciting
topic
and
see.
O
Rob,
yes,
so
thanks
I've
put
my
comments,
sort
of
mainly
in
the
chat.
Actually,
my
question
is
also
back
to
Turles
is
whether
we
should
be
trying
to
set
up
a
working
group
in
the
short
term,
so
I
think
there's
been
lots
of
very
interesting
discussion
here.
Some
of
it
I
think
is,
is
things
that
are
quite
actionable
in
the
short
term
to
do.
I
think
sureshions.com,
he's
proposing
your
suggestions.
It's
quite
easy
to
act
on
quite
quickly,
generating
metrics
is
quite
easy
to
act
on
quite
quickly
and
there's
other
things.
O
I
think
we've
discussed
where
we've
seen
wider
issues.
That
might
not
be
something
you
can
solve
in
the
iitf
at
all.
I.
Do
wonder
whether
creating
a
working
group
in
the
ITF
in
the
short
term
would
be
a
good
idea
to
act
as
a
focal
point
for
these
discussions.
I
do
have
a
question
whether
that
stands
up
as
being
a
talking
shop
or
and
or
what
sort
of
actionable
work
it
can
do,
but
similar
is
I
think
as
Tara
said
about
like
these
special
interest
groups
and
iot
Ops
I.
Think
it's
a
bit
like
that.
A
O
Yes,
necessarily
Pete,
you
had
to
do
the
work
there,
so
it
Ops
or
to
try
this
coordinate
work
with
another
working
group.
So
in
some
cases
that
that
aspect
of
building
like
the
routing
working
around
I,
think
it's
fine
and
the
aspect
of
like
longer
term
versus
short
term
I
think
would
have
to
have
a
goal
of
only
be
looking
at
work
within
the
ITF.
A
short,
shorter
term
goals
and
the
longer
term
stuff
you'd
say
would
be
in
the
irtf
would
be
my
thoughts.
E
Yes,
thank
you
so
I
had
a
comment,
a
reflection,
which
is
general
and
especially
I,
came
to
think
of
it
when
listening
to
Eva
and
Joker,
so
I
think
Ava
mentioned
something
around
The
Gap
Band
analysis.
E
So
you
referred
back
to
me
as
well,
I
think
in
addition
to
to
or
when
entering
a
collaboration
with
other
organizations
or
starting
some
Gap
analysis
analysis,
it
would
probably
be
good
to
sort
of
curve
out
a
niche
for
for
ietf,
because
there
is,
there
are
so
many
initiatives
going
on
all
over
the
place
and
I
think
what
we
have
during
these
days.
We
have
been
high
level.
E
We
have
been
on
details,
we
have
been
going
wide,
we
haven't
been
going
narrow
and
we
have
been
on
the
principal
level
on
the
Practical
level,
so
I
think
it
would
really
be
good
to
to
maybe
sort
out
the
niche
on
on
what
I,
what
what
could
be
the
main
aspects
or
or
the
main
value
that
ietf
could
contribute.
What
what
could
you
do
that?
E
No
one
else
could
do
and
the
the
second
would
be
like
on
the
advocating
plan,
which
is
things
that
are
maybe
more
high
level
but
but
which
might
be
ongoing
in
different
places,
but
you
want
to
push
a
certain
direction,
so
so
I
thought
it
could
be
good
that
that
could
be
a
good
way
of
thinking
to
to
to
find
the
position
in
this
case
already
quite
complex
landscape.
H
Yeah,
thank
you
thanks,
I
think
really
good
point
I
think
like
Rob
made
a
really
good
point
like
have
a
focal
point
may
not
have
to
produce
work,
but
at
least
like
where
people
can
go
and
talk
so
totally
up
for
that
and
I.
Think
one
more
thing,
I
wanted
to
add
is
like
there's
some
stuff
that
we
cannot
do
right
like
I.
Think
like
vanilla
brought
it
up
like
Robert
a
bunch
of
people
brought
it
up.
I
think
the
IAB
can
be
the
the
group
that
kind
of
dispatches
it
off
to
somebody
else.
H
So
we
probably
have
some
relationships
with
some
other
sdos
like
where
this
work
can
be
done.
So
if
there's
something
we
can
figure
out,
I
don't
know,
send
a
leg
or
anything
like
that.
We
are
actually
talking
to
I
think
it
might
be
a
good
thing
to
kind
of
like
also
pass
information
to
other
seos
on
things
that
could
be
useful
for
us
that
we
don't
either
have
the
expertise
or
don't
want
to
do.
F
Sorry
I
think
you
Bruce
has
his
hand
up.
P
Ahead,
oh
yeah.
Thank
you
very
much.
This
has
been
a
great
session
this
last
four
days.
I
would
just
like
to
raise
two
points.
One
is
that,
as
I
mentioned
a
few
days
ago,
the
energy
use
of
all
electronic
devices
is
probably
about
an
order
of
magnitude
larger
than
Network
equipment,
and
there
are
things
that
the
ietf
and
the
IAB
could
do
to
help
reduce
the
energy
use
of
those
Internet
connected
electronic
devices.
P
P
But
I
think
the
thing
that
we
could
do
that
would
be
the
greatest
effect
on
total
carbon
emissions
globally
would
be
to
help
our
electricity
system
move
from
the
19th
century
to
the
21st
century,
our
electricity
technology,
whether
it
be
wide
area
network,
how
electricity
networks
organized
to
even
more
importantly,
how
this
organized
locally
within
buildings
is
as
hard
really
changed
in
many
important
respects
since
the
1880s,
and
it
desperately
needs
an
upgrade
and
I
think
we
really
don't
appreciate
the
the
fact
that
the
internet
technology
was
a
revolution
that
overthrew
our
unitary
phone
system
to
move
to
a
network
model,
and
we
desperately
need
to
do
the
same
thing
for
electricity
and
we've
done
this.
P
So
we
know
we've
learned
a
lot
about
how
to
sort
of
reinvent
a
technology
on
a
networking
model
and
the
electricity
people
really
have
a
lack
of
a
vision
in
this
area
and
desperately
need
the
help
of
the
people
on
this
call.
So
so,
please,
please
help
out
that's
there's
far
more
carbon
at
stake
in
in
improving
our
electricity
system,
so
we
can
integrate
more
Renewables
and
reduce
costs
than
everything
else.
We
could
do
so.
Thank
you.
I
Yes,
please
thank
you.
Thank
you.
Yes,
I
joined
the
the
the
the
the
praise
for
the
workshop
quite
quite
interesting,
quite
wide
area
of
topics
that
we
discussed
quite
in
lightning
on
on
the
on
the
topic
of
collaborating
with
other
stos
I
I
think
it
makes
quite
good
sense
to
consider
one
of
them
as
as
the
itu
and
Etsy
as
vanilla.
I
As
mentioned
on
the
itu
side
study
group
13
focus
on
future
networks,
looking
at
expansions
of
existing
architectures,
some
new
protocols
that
could
be
considered
and
study
group
five,
which
looks
at
the
environmental
aspects
of
of
icts
I,
think
it
would
be
quite
critical
and
it's
an
expectation
for
for
the
ietf
who
have
has
the
the
protocol
experience
around
the
IP
stack
experience
that
that
such
knowledge
comes
from
from
the
ietf
and
and
also
for
reasons
of
continuity
and
also
for
backwards
compatibility.
I
I
Ietf
will
be
the
best
place
to
to
in
to
ensure
that
whatever
protocols
expansions
are
being
considered,
they
would
be
interoperable
and
backwards
compatible
and
can
co-exist
with
with
existing
architectures
one
area
it
that
could
be
at
risk
is
to
to
with
lack
of
interoperability
if
these
Protocols
are
done
somewhere
else
that
fragmentation
of
the
network
would
be
would
be
or
could
could
take
place.
I
So
I
think
these
are
areas
to
to
consider
and
I
share
the
The
View
that
collaboration
in
this
vast
space
is
necessary,
because
no
entity
has
the
whole
experience
of
the
of
the
value
chain,
but
I
think
also.
We
need
to
focus
on
where
we
have
the
expertise
and
can
lend
our
expertise
to
other
stos
and
collaborate
on
that
for
in
that
front.
A
Yeah
good
points,
indeed,
collaboration
is
the
way
to
go
and
finding
the
our
niche
it's
pretty
little
about
it.
A
I
think
it's
is
key
and
we
don't
own
the
whole
whole
problem
on
part
of
it
and
even
for
the
part
that
we
are
working
on,
we
need
to
maybe
sometimes
understand,
what's
going
on
in
other
other
parts
as
an
example,
if
we
want
to
play
tricks
on
our
side
on
like
when,
when
things
are
up
and
down
and
so
on,
we
would
actually
have
to
understand
what
the
league
layers
do
and
whether
that's
beneficial
or
not
or
on
under
what
circumstances.
A
So,
there's
a
lot
lots
to
coordinate,
but
maybe
even
yeah,
possibly
some
meetings
where
there's
more
joint
participation
on
in
the
space
that
might
be
useful
thing.
Looking
at
the
time,
I
think
I
don't
see
anybody
else
having
their
hands
up.
If
not,
then
I'll
just
go
briefly
to
my.
J
O
A
A
I
think
I
think
you
see
my
screen.
Yeah
I'm,
gonna,
I,
I
I,
will
have
sort
of
a
technical
conclusion,
so
I'll
cover
that
last,
but
just
that
the
next
steps
in
the
workshop
we
talked
about
that
a
little
bit
already
when
Taurus
was
asking
about
this,
but
discussion
on
the
way
should
continue
we'll
work
on
a
draft
report.
A
If,
if
you
are
interested,
please
contribute
or
we
don't
probably
need
50
people
helping,
but
we
would
appreciate
a
couple
like
everybody
or
a
person,
at
least
for
for
each
of
the
sections,
and
the
sections
are
roughly
like
our
systems
where
something
like
that
would
probably
be
helpful,
we'll
be
editing
it
with
Colin
I
I
think,
but
that
help
would
be
appreciated.
We'll
have
some
discussions
at
the
ietf
I.
Don't
exactly
know
what
the
answer
is
in
terms
of
a
new
working
group.
I.
A
Think
that
needs
more
more
discussion.
We
also
heard
about
yeah
even
and
sures
I
think
you
talked
about
this
sustainability
considerations
and
trade-offs
documents.
That's
a
that's,
clearly
a
possibility
and,
as
has
been
mentioned,
there's
some
some
things
that
are
in
progress,
or
at
least
should
be
in
progress.
A
A
I
I
I
keep
thinking
that
we're
not
doing
enough.
If
this
is
all
what
we
do,
it's
not
probably
enough,
we'll
probably
have
to
do
some
other
things.
So
if
you
have
ideas,
please
so
just
maybe
given
the
time
suggest
them
on
the
on
the
chat
window.
A
Yeah
before
I
go
forward.
Colleen
did
you
have
anything
else
to
say.
B
Yeah,
so
it's
a
wrap
up.
From
my
point
of
view,
I
mean
that
this
has
been
an
interesting
Workshop.
Certainly
a
lot
of
things
to
think
about.
B
I.
Think
a
lot
of
these
short-term
things
things
we
can
do
engineering
wise.
That
can
happen
recently
quickly.
B
Clearly,
a
much
interesting
research
questions,
there's
a
bunch
of
things
which
are
in
scope
for
the
ITF
Community,
a
bunch
of
things
which
are
a
lot
broader
scope,
I.
Think
as
as
Eve
said,
we
need
to
remember
the
urgency
right.
It's
easy
to
just
say:
we
need
to
do
more
research
and
clearly
we
do
need
to
do
more
research,
but
I
think
we
need
to
build
on
the
momentum
and
the
act
and
maybe
develop
some
standards
in
the
ITF
made.
B
A
Thanks
I'm
gonna
show
one
more
thing,
but
in
the
in
the
in
the
background
or
in
the
meanwhile
we'd
like
to
know
how
this
Workshop
went
and
given
that
we're.
A
Of
time
quickly,
then,
maybe
we're
not
going
to
do
this
verbally,
so
you
could
do
two
things.
You
could
comment
in
the
chat,
we'll
save
the
chat
log
or,
if
you
feel
like
you,
can
also
email
me
and
call
in
and
we'll
collect
the
the
feedback
and
and
try
and
build
a
build
and
understand
what
we
did
well
or
what
could
be
improved.
A
Certainly,
from
my
perspective,
the
spinos
teller
Workshop
really
appreciated
the
diversity
of
backgrounds
and
opinions
and
and
learning
stuff,
at
least
personally
gained
in
this
work,
so
that
that's
great
I
think
there's
there's
indeed
momentum
and
urgency,
and
it
doesn't
mean
that
we
have
to
like
do
huge
things
right
now,
but
it
does
mean
that
we
have
to
do
some
things
every
year
and
keep
improving
I
think
that's
that
that
is
the
way
to
go
and
anyway,
so
the
one
one
more
thing
that
I
wanted
to
show
and-
and
this
isn't
complete
by
any
any
chance.
A
But
I
was
doing
that
on
the
background,
while
you
guys
were
talking
so
please
excuse
me,
but
I
did
take
some
notes
on
like
what
are
the
observations.
One
of
us
on
this
continuous
Improvement
and
influence
understanding
how
important
can
help
the
society
understand
what
the
situation
is
and
why
is
it
like
like
it
is,
and
how
could
we
affect
that
and
understand
the
trade-offs?
A
There's
a
bunch
of
useful
tools?
Much
of
the
discussion
was
focused
on.
You
know,
naturally,
on
the
ietf
protocol
design
issues
and
some
of
it
was
was
specifically
unrouting
and
it's
probably
a
little
bit
broader
than
that
that
we
have
to
worry
about
at
least
in
terms
of
like
in
our
day,
jobs
worry
about.
Implementations
worry
about
Renewables,
as
well
as
proper
Countryside
actions
that
we
should
take
continue.
The
discussion
connecting
the
ietf
to
others
I
think
that's
sort
of
a
long-term
thing.
A
How
can
we,
you
know
not
just
have
one
meeting
with
somebody
else,
but
but
make
sure
that
we're
on
the
map
and
understand
what's
needed
and
have
the
discussion
partners
where
that
actually
matters,
maybe
ITF,
I'm,
sorry,
IEEE
and
gtpps.
As
sort
of
a
link
layer
organizations
come
to
mind,
we
should
pick
the
low-hanging
fruits,
the
avoiding
the
lengthy
formats,
the
crypto
assets
and
a
non-renewable
energy.
A
A
You
know
we
have
researchers
participating
in
Academia
and
there's
a
lot
of
interest
in
and
funding
on
this.
These
topics,
and
we
we
can
actually
do
multiple
things
at
the
same
time
and
and
Academia
in
particular-
is-
is
more
geared
towards
understanding
than
the
necessarily
fixing
products.
For
instance,
the
sponsored
research
that
needs
doing
I
think
we
should
dive
into
that
somehow
and
yeah
I
I
think
that's
more
or
less
it.
Clearly.
A
We
need
to
write
up
what
happened
in
the
workshop
and
summarize
it
more
nicely
than
than
the
slide,
but
a
lot
of
really
good
material
and
lots
of
leads
for
us
to
actually
go
and
pursue
and
leads
for
different
time
scales
and
short
term
and
long
term.
So
that's
also
really
good,
because
we
also
need
the
thing
that
we
need
to
do
in
in
five
or
ten
years.
So
I
think
it's
all
good
call.
In
any
final
words,.
A
A
It
up
and
and
figure
out
what
we
will
do
in
addition
to
the
mentioned
things
at
the
next
ITF,
so
do
stay
tuned.