►
From YouTube: IETF105-GAIA-20190723-1330
Description
GAIA meeting session at IETF105
2019/07/23 1330
https://datatracker.ietf.org/meeting/105/proceedings/
A
B
You
can
follow
along
on
line
with
the
agenda
first
up
for
every
meeting
here
at
the
IETF
is
that
the
blue
sheets
are
circulating,
please
sign
in,
and
that
we
have
to
remind
you
of
the
note.
Well,
you
can
take
a
look
online
I've
been
told
that
I
don't
have
to
go
through
the
whole
thing.
So
it's
up
to
you
to
take
a
look,
but
you
should
know
this
by
now.
B
B
Underserved
areas
places
where
you
know
at
least
49
percent
of
the
population
of
the
planet
are
not
connected,
so
we
bring
in
different
experts
to
talk
about
ways
to
look
at
building
anything
from
community
networks
to
internet
exchange
points,
local
wireless
Wi-Fi
mesh,
whatever
works,
and
some
of
the
some
of
the
people
in
the
room
here
are
part
of
the
participate
in
battle
mesh
where
they
hack
mesh
standards.
They've
deployed
community
networks,
which
are
both
wireless
Wi-Fi
and
fiber,
which
is
grief
II,
which
is
the
network
that
Leandro
works
with.
B
We
also
have
with
us
today,
people
from
ripe
ncc
to
talk
about
what
they're
doing
to
help
communities
of
interest
with
network
operator
groups.
We
have
a
colleague
from
a
tribal
community
he
and
the
United
States
that
works
with
19
different
indigenous
tribes
looking
at
connectivity
and
how
to
create
sustainable
networks
that
are
more
community
built,
rather
than
last
mile,
so
I'm
going
to
turn
to
Leandro
for
a
sec
for
a
bit
more
of
an
introduction.
While
we
get
people
queued
up.
C
Well,
hi
everyone.
There
are
several
like
housekeeping
things.
One
is
I
mean
there
are
remote
participants.
Today
there
is
the.
If
you
look
at
the
agenda,
you
will
see
the
in.
There
is
one
link
to
a
etherpad
where
we
collect
notes
so
you're
very
welcome
to
go
there
and
to
contribute
text.
We
try
to
collect
a
summary
of
the
discussion
over
there.
I
will
be
there
taking
notes,
but
I
mean
anyone
else
is
welcome.
C
C
People
can
also
ask
questions
over
there
and
then
just
going
directly
to
the
introduction,
there's
so
many
screens
in
our
table,
so
the
slides
are,
the
ID
will
be
able
to
download
them
later
on
and
the
objective
of
the
interaction
just
to
give
a
bit
of
context
for
those
who
are
new
to
the
group.
This
comes
from
the
web
page
from
the
group
in
in
the
IRT
EFT
on
our
website
and
well.
It
says
that
this
is
started
back
in
2012.
Looking
at
the
the
access
gap
in
and
ways
to
to
deal
with
that.
C
So
maybe,
and
then
this
is
a
list
of
objectives
that,
because
sometimes
we
you
forget
about
the
as
as
we
go
as
we
share
information
as
we
have
presentations,
we
can
forget
about
what
why
we
started
and
when
I
work
with
the
aims.
So,
as
you
can
see,
is
about
increasing
increased
visibility,
any
interest.
C
So
one
of
the
aim
so
of
this
working
group
is
to
share
to
make
visible
different
initiatives
very
diverse,
but
probably
complementary
how
to
create
together,
efficient
that
in
blog
dimples
many
stakeholders
how
to
articulate
and
foster
collaboration
among
them,
including
the
the
different
aspects
which
have
to
do
with
what
ITF
does
also
the
research
part.
What
a
ITF
does,
and
and
also
what
everyone
of
you
do
in
your
own
context
and
community.
C
Very
important
sustainability
is
about
documenting
different
aspects
which
are
critical
for
for
addressing
the
the
unconnected
or
improving
the
connectivity
for
those
who
are
connected
and
then
to
look
at
how
how
the
research
part
that
we
do
can
translate
into
into
the
efforts
that
IDF
does
in
general.
So
this
is
the
list
of
objectives
and
then
for
to
contribute
to
that
today.
This
is
the
agenda,
so
without
much
daud-you
we
will
start
with
the
first
presentation
from
the
remote
presentation
from
ripe
ncc.
It
was
already
one
introductory
or
previous
presentation
in
a
previous
guided
session.
D
Thanks
well
thanks
for
having
me
and
last
time,
I
was
actually
there
in
person
and
I.
Remember
I
was
one
of
the
few
people
who
didn't
present
them
remotely.
So
now,
this
time,
I'm
also
going
to
present
remotely
from
a
tropical
Amsterdam
at
the
moment,
so
I'll
tell
you
to
change
the
slides
right.
Is
that
how
that
works?
D
D
That's
not
our
office,
it's
Central
Station,
but
we're
located
right
next
to
it
beautifully
central,
and
so
damn
we
have
more
than
20,000
members
now
and
which
is
mostly
because
we're
running
out
of
ipv4
and
everybody
is
trying
to
get
the
last
last
blocks
of
before
so,
but
you're
doing
in
many
other
activities
like
and
last
time,
I
presented.
Maybe
you
can
go
to
the
next
and
last
time,
I
present
it
mostly
on
our
data
and
tools
that
did
we
provide
like
as
a
regional
registry.
D
We
have
a
lot
of
data
that
we
collect
from
our
members
as
part
of
the
IP
registry
that
we
run
and
maintain
and
operate,
and
with
that
we
also
offer
a
lot
of
tools,
for
instance
like
ripe,
stat
or
ipv6
measurements
or
routing
measurements
that
you
make
available
to
everybody,
and
we
also
have
another
measurement
platform
that
some
of
you
might
know.
It's
called
web
atlas.
D
D
D
Collector
or
from
a
collection
of
route
collectors
that
we
may
collect
the
data
and
make
available
and
last
time
I
was
mentioning
a
new
service.
We
had
it's
called
Miss
livestream,
where
you
can
basically
follow
route
changes
and
live
in
a
certain
tool.
So,
if
you're
interested
in
any
of
this
I'm
not
going
to
talk
about
that
now,
there
was
all
last
time
and
you
can
look
at
the
slides
from
NASA,
and
this
time
you
can
move
on
please
and
this
time.
D
I
got
to
talk
more
about
what
we
do
for
other
other
services
that
we
provide
to
the
community
next
slide,
please,
where
it's
a
bit
of
a
mixed
bag
now
of
various
community
projects
that
that
we
run
on
basically
supporting
various
parts
of
the
community.
The
first
one
is
for
network
operators,
groups
and
Jane
said
in
her
introduction
that
obviously
one
of
our
main
communities
and
also
most
of
our
members,
our
network
operators
in
there
are
some
form
or
fashion
I
mean.
D
Doesn't
it
can
also
be
enterprises
or
a
government
networks,
but
mostly
people
who
run
networks
and
so
and
more
and
more
over
the
last
few
years
and
and
you'll
probably
notice?
There's
an
increasing
number
of
more
local
network
operators
groups
all
over
the
world,
almost
like
a
natural
evolution
that
they
kind
of
started
with?
Maybe
the
regional
logs
such
as
ripe
and
a
frog
and
Latin
organ
like
what's
a
cult,
it's
called
VOC,
oh
no,
it
was
another
world
in
latin
america
and
also
we
have
right
in
europe
Narrogin
us.
D
National
network
operators
groups,
sometimes
even
citigroup,
select,
meetups
and
such,
and
so
we
thought
it
wouldn't
be
nice
to
have
like
one
site
where
all
this
information
is
collected.
So
we
created
a
map,
you
can
see
there
a
screenshot
and
if
you
hover
over
that
as
a
link
at
the
bottom,
if
you
hover
over
the
link
you
can,
you
can
see
the
information
of
that
well,
not
on
the
screen
right
now.
D
But
then,
when
you
go
to
that
website,
you'll
see
a
link
to
that
particular
network
operators
group,
and
then
you
can
find
information
about
the
next
meeting
where
they
are.
You
know
how
you
can
contact
them
and
and
I'm.
Some
people
use
this
to
find
the
local
Naga
and
then
you
can
start
participate
on
that
same
page.
There's
also
like
a
whole
list
of
these
of
these
groups,
and
it's
not
only
as
you
can
see.
D
As
we
are,
we
are
in
contact
with
a
lot
of
people.
We
sometimes
can
also
provide
sponsorship,
and
obviously
we
have
a
lot
of
expertise
and
an
experience
in
that
connecting
people
and
them
and
helping
people
to
start
such
an
ox.
For
instance,
have
you
recently
also
set
up
a
mailing
list
for
the
NOC
organisers?
That's
the
last
point
on
that
slide.
It's
a
listed
and
we
created.
G
D
Necessarily
for
all
the
participants,
but
more
for
the
for
the
problem,
committees
or
organizing
committee,
so
that
they
can
exchange
some
information
and
it's
amazing.
You
know
how
many
different
shapes
and
forms
these
network
operators
groups
coming.
There
are
some
that
only
meet
like
a
couple
of
hours
in
the
evening.
Every
few
months,
other
stuff,
like
two
full
day,
meetings
for
training
sessions,
everything
and-
and
it's
really
good-
to
see
that
they're
helping
each
other
out
in
exchanging
information
and
also
exchanging,
for
instance,
and
potential
speakers
and
and
content
yeah.
D
You
can
move
on
I
think
another
another
project
we
started
a
few
music
here
and
I.
Remember:
we've
received
I
was
actually
talking
to
Jane
at
the
time,
but
she's
involved
in
a
similar
project
and
from
the
ithe
Internet
Society
perspective.
It's
what
we
call
the
ripe
ncc
community
project
fund
and
it's
basically
a
way
to
I
mean
he
take
membership
money
so
we're
based
on
membership
fees
of
all
these
over
20,000
members.
D
G
D
D
There's
a
next
one
committee
that
evaluates
the
applications,
there's
a
link
that
you
can
go
to
the
page
and
and
submit
your
proposal,
and
there
are
a
number
of
criteria.
Basically,
it
has
to
be
not-for-profit
and
beneficial
for
the
community
and
it
tends
to
be
a
lot
of
open
source,
both
software
and
hardware,
that
we
provide
and
supported
in
the
last
few
years.
D
I
mean
it's:
free
people
can
participate
the
organizing
to
hackathon,
see
here
this
year.
It's
the
tenth
hackathon
we're
actually
organizing
on
Io,
T
and
I
and
security.
So
we
don't
know
quite
yet.
How
is
this
going
to
pan
out,
but
we're
hoping
people
will
come
in
maybe
dissect
some
devices
and
see
how
they
can
be
built
better
and
made
better
and
made
more
secure
over
the
last
few
years.
D
Such
as
measurements
and
tools
and
DNS
and
exchange
points
tools
for
exchange
points
we
had
the
last
instance
was
interesting-
was
on
quantum
internet
and
out
of
that
also
came
that
quantum
research
group.
That's
now
a
part
of
the
I
am
I
RTF
and
I,
don't
know
if
the
meeting
this
week,
but
you
probably
know
that
better.
So
this
one
next
one
again
it's
open
for
applications,
you
can
still
register
it's
an
app
in
conjunction
with
exploiting
and
also
what's
on
that
slide
that
I
wanted
to
mention.
D
D
To
come
to
write,
meeting
and
present
their
research,
and
also
for
this,
the
deadline
of
the
application
is
still
open.
Till
August
of
is
left
space
and,
at
the
moment,
I
think
you're.
Looking
for
candidates
for
the
arrive
meeting
in
rotterdam
and
also
for
a
regional
meeting
in
kiev,
I
believe
in
october
so
go
to
the
website.
Have
a
look
I
mean
it's
a
really
good,
an
opportunity
to
present
your
research
and
to
reach
out
to
operators
and.
D
D
The
ripe
NCC
is
expertise
in
or
anything
I
mentioned.
Let
us
know,
and
the
last
but
not
least,
I
wanted
to
mention
these
ripe
Atlas
ambassadors
that
I
also
I
mentioned
at
the
beginning,
because
there's
also
a
really
good
and
successful
community
project,
and
we
we
almost
don't
have
to
do
anything
with
that.
You
know
there
are
notes
of
these
ambassadors
out
there.
Handing
out
these
probes
also
the
our
ers,
the
other
ARS
helped
a
lot
with
that,
and
we
couldn't
do
it
without
them.
So.
D
B
Are
perfectly
on
time,
Miriam
and
the
floor
is
open
for
about
two
minutes,
for
questions
for
Miriam
and
Miriam
is
just
good
to
see
your
face.
Thank
you.
Thank
you.
Does
anyone
have
any
questions
for
ripe
ncc
and
if
you're,
not
an
ambassador
for
the
Atlas
program,
definitely
contact
right
because
we're
seeing
the
absolutely
important
amalgamation
of
data
that
comes
out
of
ripe
Atlas
for
some
of
the
work
that
we're
trying
to
do
for
more
evidence-based
research
and
more
information
out
there
in
general,
any
questions.
B
K
Miriam,
it's
sang
by
drawing
from
Indian
society,
North,
American,
Bureau
and
I'm
glad
you
mentioned
the
committee
networks
program
that
that
you
have
I
wasn't
following
this
is
a
new
program
or
you
had
it
for
a
couple
of
for
some
time
and
if
you
already
had
it
before,
can
you
share
some
best
practices,
or
some
case
studies
I
think
would
be
very
useful
for
us
to
take
into
consideration.
Thank
you.
You.
K
D
Evaluate
some
applications,
we
have
a
replication
forum
where
people
have
to
submit
the
project
lifetime.
What
the
results
are
this
the
goals?
What
kind
of
budget
they're
looking
for
and
it's
similar
to?
What
was
they
call
it
I
saw
had
something
similar
I
remember
but
of
you.
So
we
also
learned
a
lot
from
how
you
and
this
in
the
past
I
think
maybe
the
chapter
projects
and
I'm
referring
to
and
and
so
it's
a
it's
it's
you
year-long
project,
sometimes
and.
D
D
Also
ask
or
like
a
midterm
report
and
at
the
end,
is
usually
like
a
report
either
at
the
right
meeting
or
on
the
Reb
labs,
blog
and
and
what
I've
seen
and
usually
these
projects
their
lyrical
nicely
and
through
the
community
as
well.
There's
a
lot
of
open
source,
I,
open,
BGP
and
security
projects
and
came
out
of
that.
B
Thank
you
very
much
Mariam,
and
that
does
dovetail
really
nicely
with
that.
There
are
some
other
grant
projects
that
we'll
probably
get
all
that
information
out
on
the
Gaia
list.
Well,
thank
you
very
much
and
we're
gonna
hop
over
to
the
team
from
M
labs,
so
Miriam,
thank
you
and
we'll
have
her
presentation
up.
So
anyone
it's
already
on
the
web.
If
anybody
has
any
other
questions,
you
can
contact
Miriam
directly,
so
M
labs
team.
You
are
up
wonderful
leggy.
G
Thank
you
can
hear
me
all
right:
cool
hi
everyone.
My
name
is
Lonnie
Olson
I,
recently
joined
measurement
lab
as
the
projects
director
myself
and
crystal
our
procurement
management
and
Emilia
will
be
speaking
about
how
we
support
the
internet
search
community.
Our
contact
information
is
here
as
well
as
the
link
to
the
slides,
if
you
like,
any
more
information
after
the
presentation
slide.
So
what
is
my
mama
has
made?
The
few
components
were
supported
by
teams
code
for
science
and
society
Google
and
instance,
planetlab
Princeton's,
planet
bug.
G
We
provide
a
data
repository
that
is,
you
know,
open,
contributed
by
users,
longitudinal
and
drive
for
open
source,
internet
infrastructure
data
and
in
we
deploy,
build
and
run
infrastructure
that
aims
in
the
creation
of
that
data.
For
posit
for
you
next
slide,
you
know,
point
was
simply
fishin
is
too
much
of
the
internet.
Saves
data.
G
We
like
to
make
because
that
we
don't
measure
by
ourselves,
people
measure
the
internet
using
their
their
own
devices
and
our
servers
and
we
collect
it,
comes
from
those
measurements
and
make
it
publicly
accessible,
so
that
all
this
is
done
with
our
principles
in
mind.
We
take
privacy
seriously,
so
all
of
our
data
is
synthetic
data
and
our
measurements
are
active,
meaning
only
clients
initiate
servers,
do
not
rest
on
their
own.
The
experiments
are
curated
and
reviewed
by
our
experts.
G
Our
expert
review
committee,
which
are
and
they're
mostly
developed
by
academics
and
those
measurements,
come
from
the
community
meeting.
Anyone
can
develop
an
edge
client
anyway.
Again
this
new
platform
friend
pressed
against
the
top
one
and
we
believe
in
open
it.
So
that
would
mean
all
of
your
data
is
within
some
distance
to
zero
and
all
our
code
that
we're
gonna
miss
for
sure
there's
a
choice,
and
this
is
useful.
Our
community
involves
a
wide
variety
of
folks.
G
L
M
L
D
L
So
a
bit
about
our
platform
infrastructure
for
those
who
may
not
already
be
aware
of
them
lab.
We
place
our
servers
in
transit
data
transit,
connected
data
centers
so
connected
to
peering
locations
and
internet
exchanges,
which
is
a
different
role
in
a
way
that
other
speed,
testing
and
other
internet
measurement.
Public
internet
measurement
platforms
place
servers.
So
we
are
not
inside
of
the
last
model
networks,
but
we
are
outside
of
that
too.
Then
you
have
some
information
routes
and
performance
to
locations
outside
of
the
last
mile.
L
The
widest
used
is,
of
course,
our
network
diagnostic
tool
that
was
originally
from
internet2
and
then
paris
traceroute
is
also
used
quite
heavily
by
our
team.
In
as
a
matter
of
course,
we
published
all
those
data
sets
in
bigquery
to
the
next
slide,
so
signing
up
for
our
discuss.
Group
gives
anyone
query
access
for
free
and
the
bigquery
web
interface
in
other
locations
available
to
query
our
bigquery
data
sets
is
how
most
researchers
interact
with
our
data
and
on
the
next
slide,.
L
We
also
publish
all
our
switch
utilization
data
openly
as
well,
so
one
can
examine
the
throughput
of
our
of
each
location
where
that
collects
data
on
to
the
next
slide.
In
terms
of
accessing
our
data,
there
are
several
ways
to
do
that:
we
publish
our
source
data
and
that's
all
basically,
the
raw
packet
captures
and
traces,
the
meta
that
occur
each
time
a
test
is
conducted
to
one
of
our
servers,
the
visualization
website,
which
is
an
aggregation
of
our
NDT
data
based
on
location
and
and
by
provider.
L
Bigquery,
of
course,
which
I
just
mentioned,
that
parse
data
is
the
primary
way
people
access
our
data
and
then,
of
course,
there's
a
variety
of
other
ways
to
access.
And/Or
generate
data
which
we're
going
to
talk
about
here
a
moment.
So
we
just
run
through
these
real
quick
next
slide.
The
source
data
is
openly
available
and
it's
stored
in
compressed
TGT
GZ
format
in
Google,
Cloud
storage.
It's
primarily
used
by
a
small
number
of
researchers
and
partners
for
exam
example.
L
Someone
recently
wanted
to
calculate
actual
Vitter
according
to
RFC
of
various
tests,
from
a
certain
geography,
as
opposed
to
what
the
entity
client
gives
back
to
the
user,
which
is
an
estimate
of
jitters
so
to
do
that
they
needed
to
access
and
parsha
the
raw
trace
files.
On
the
second
next
slide,
we
published
a
visualization
website
which,
on
our
website,
it's
the
visualizations
link
in
the
menu
that
is
meant
to
being
an
accessible
front
door
for
the
general
public
to
see
where
they
are
in
the.
L
See
what
our
data
says
about
their
locations
and
ISPs
and
that's
updated
approximately
monthly
and
then
in
the
next
slide,
the
primary
way
as
I
mentioned
to
access
our
data
is
in
bigquery,
there's
a
Quick
Start
Guide,
which
gets
anyone
set
up
with
a
gmail
account
to
query
that
data
for
free
variety
of
documentation
is
available,
and
our
support
addresses
here
on
this
slide.
This
is
the
primary
way
we
interact
with
researchers
who
want
to
use
our
data
on
to
the
next
slide.
L
Please
so
because
it's
a
research
focus
group
and
you're
interested
in
what
this
data
can
tell
you.
So
we
do
have
a
select
number
of
publications.
These
are
both
Oh
Demick
in
non-academic,
listed
on
our
website
and
I've
just
pulled
a
few
of
them
more
recent
ones
here
that
are
interest,
some
of
which
are
coming
from
researchers
and
academic
units
and
some
which
are
not
from
a
variety
of
fields
as
well.
L
L
The
other
way
that
searchers
interact
with
measurement
lab
is
to
host
an
experiment
so
or
a
test.
So
any
of
the
tests
that
are
hosted
on
our
platform
are
applied
by
someone
who's
designed
a
network
measurement
experiments,
and
we're
excited
that
in
this
calendar
year
we
will
complete
our
migration
from
a
planet.
L
System
on
our
servers
to
the
docker,
kubernetes
managed
platform,
and
there
will
be
new
opportunities
to
host
new
experiments
in
different
contexts,
and
so,
if
you're,
a
researcher,
that's
developing
Network
measurement
experiments
you
can
apply
to
hosts
with
M
lab
and
I
noted
that
one
that
will
one
new
experiment
that
will
be
out
of
the
platform
as
the
whee-hee
app,
which
is
a
net
neutrality,
focused
focused
test.
We're
excited
about
that
new
platform
and
in
the
next
few
slides
I'm
going
to
talk
about
how
we
currently
support
researchers,
so
there's
the
general
support
and
documentation.
L
So
that's
the
primary
way
if
you
email
support
myself
or
Letty
or
someone
else
on
the
team,
we'll
get
back
to
you
and
answer
your
question
we
generally.
This
is
light
support
and
how
to
get
you
access
to
query
the
data
for
free
or
to
answer
your
questions
as
you
begin
to
explore.
It.
We've
also
begun
in
more
recent
last
year,
so
more
recently
to
do.
Presentations
in
combination
with
partners
at
different
research
communities
so
listed
here
are
some
that
we've
done
over
the
last
year
or
so,
which
were
highly
interesting
and
valuable.
L
To
make
connections
with
other
folks
using
data,
there
was
some
really
interesting
conversations,
for
example,
with
the
the
folks
at
the
research
ICT
Africa
group
at
AFRICOM,
afrinic
and
AIS
to
talk
about
leveraging
multiple
types
of
datasets,
for
example,
using
picking
using
them
in
combination.
So
we've
had
some
conversations
about
how
to
use
rec
data
and
M
lab
data
together,
and
so
those
conversations
will
continue
to
happen.
We've
also
started
to
sponsor,
on
a
small
scale,
some
specific
research
events
and
to
support
hackathons
like
the
sitcom
hackathon
and
booty
fest.
D
B
B
L
L
I'll
I'll
wrap
it
up.
Thank
you
for
that
time.
Jack
Jane,
so
another
way
when
you're
supporting
research
is
partnering
on
a
smaller
scale,
on
specific
research
initiatives.
So
in
the
u.s.
right
now
your
partner
on
a
research
program,
that's
focused
on
US
public
libraries
as
an
anchor
institution
to
produce
a
premise-based
device
that
can
run
our
tests
automatically
and
produce
a
visualized
display
of
that
data.
L
Next
slide
and
community-based
research
is
another
area.
I
wanted
to
highlight
separately,
because
not
only
are
we
supporting
research
on
the
academic
side,
but
also
community-based
data
gathering
initiatives
and
visualization
initiatives,
so
one
reach
that
one
is
a
app
that
integrates
NDT
tested
and
in
the
US,
the
National
Association
of
counties
built
this
app
and
are
using
this
to
gather
better
leaf
geolocation
data.
In
order
to
understand
broadband
speeds
in
rural
areas.
Next
slide,
then,
we've
participated
and
ported
of
community
based
mapping
and
surveying
initiatives.
L
This
one
some
began
several
years
ago
with
the
city
of
Seattle
Washington
and
several
other
rural
counties,
the
idea
being
to
place
a
survey
and
a
web-based
NDT
test
in
a
website
that
allows
users
to
both
answer
questions
about
their
broadband
speeds
and
to
protest
to
gather
that
data
next
slide.
And
so
we've
worked
with.
L
Institutions
doing
this
type
of
work.
This
slide
show
cases
the
Michigan
initiative
that
is
focused
on
homework
gap
and
giving
the
test
as
an
assignment
to
students
next
slide,
and
this
there's
another
initiative
in
Louisville
Kentucky
that
started
in
Louisville
Kentucky.
That
is
now
blossoming
into
a
larger
initiative
from
the
technology
Association
of
Oregon.
L
C
L
G
Also,
we're
really
trying
to
move
towards
a
more
community,
oriented
governance
structure
and
thinking
about
how
we
can
respond
best
and
provide
the
best
service
for
the
people
that
are
using
the
measurements
and
the
data.
So
if
you'd
like
to
be
involved
or
hear
more
about
that,
we
would
really
like
in
your
venue
and
specifically
in
this
room.
B
B
Okay,
I
think
we
thank
you
very
much.
There's
no
one
approaching
the
microphone.
Thank
you,
lady
and
Chris.
The
slides
are
on
the
web,
so
if
anybody
wants
to
double-check
and
reach
out
to
the
EM
labs
community,
please
do
and
thank
you
for
what
you're
doing.
We
know
that
the
community
is
appreciative
of
a
lot
of
the
research
and
looking
at
the
broadband
Maps.
B
We
always
makes
me
smile,
because
the
next
presentation
from
our
colleague
Matt
ran
Tannen
who's
with
the
tribal
communities
will
show
us
where
there
are
some
very
big
gaps
in
connectivity
and
the
certain
parts
of
the
world.
So
thank
you
very
much
and
Matt's
up
will
pull.
It
will
pull
up
your
slides.
Thank
you.
B
D
H
P
The
clicker
okay,
so
welcome
Matthew
Ann
Tannen
I,
currently
hold
the
position
of
Director
of
Technology
for
twenty
Native
American
tribes
in
Southern,
California
and
I've
been
working
with
them.
I'm
starting
my
19th
year
with
them.
I
also
run
partnering
in
business
development
for
arcadian
infra
comm,
which
is
a
fiber
builder
in
the
United
States.
It's
about
18
months
old.
P
It's
been
in
the
public
since
January
I'll
explain
a
little
bit
more
about
that,
and
then
you
can
see
a
few
of
the
other
things
that
I've
been
involved
with
currently
involved
with
him
involved
with
back
in
the
past,
so
we'll
just
jump
into
it
since
we
have
10
minutes
or
so
so.
I
want
to
talk
about
what
the
previous
administration
used
to
call
technology
deserts.
P
P
So
it
shows
you
know
clearly
a
infrastructure
going
to
an
area
with
not
a
lot
of
people,
but
it
is
a
way
to
connect
to
people
and
I
just
want
to
I'll
move
along
until
we
get
into
actually
I'm
gonna
go
forward
and
then
come
backwards
again.
So
we
have
land
that
looks
like
this
in
the
tribal
regions.
Now
you
remember
the
you
know,
there's
also
in
Canada
634
First
Nations
in
in
the
US
there's
573
federally
recognized
tribes
and
these
tribes
were
moved
to
areas
where
the
general
population
didn't
want
to
live
right.
P
So
they
pushed
people
into
the
base
of
the
mountains.
They
pushed
people
away
from
the
beach
away
from
the
lakes
away
from
where
they
normally
were
inhabitation
and
and
they
were
pushed
to
these
regions
that
didn't
have
services
and
didn't
have
opportunity.
So
it
was
a
responsibility
of
the
federal
governments
to
be
able
to
connect
those
people
and
bring
the
resources
to
them
because
they
put
them
there
and
that's
not
obviously
happening
so.
The
tribes
themselves
have
been
working
to
solve
this
problem,
but
some
of
the
geography
looks
like
this
and
then
aerial.
You
know.
P
You're
looking
down
in
valleys
that
are
not
visible
from
other
locations,
except
for
mountaintops
and
yeah,
and
and
so
this
is
the
kind
of
solution
that
makes
a
lot
of
sense
in
in
tribal
communities.
I
would
say
one
size
does
not
fit
all
every
reservation.
Every
tribal
reserve
is
different,
some
of
them
are
on
the
flattest
land.
You
can
imagine,
looks
like
somebody
ironed
it.
Some
some
of
them
are
on
the
the
most
rock
rocky
jagged.
You
know,
geography,
that's
available,
and
so
digging
fiber
in
a
lot
of
situations
is
not
permissible.
P
Just
because
geography
digging
fiber
is
typically
not
permissible,
because
when
we
varied
people
and
had
cultural
ceremonies
and
things
and
recorded
that
that
was
taken
away
and
and
those
records
were
destroyed.
So
a
lot
of
the
where
somebody
is
buried
is
not
recorded,
and
so,
when
you
dig,
you
typically
run
into
artifacts
and
different
things
that
are
tough
to
deal
with.
P
So
you
don't
want
to
to
dig
on
tribal
land
unless
you
really
have
a
good
knowledge
of
what's
going
on
so
a
lot
of
the
things,
because
we
are
in
remote
locations
and
we
are
away
from
where
the
services
are
delivered.
We
have
to
provide
our
own
services
as
well.
So
solar,
solar
power
is
a
big
power
source,
wind
power
and
generator
power,
a
big
power
source
to
survive
and
provide
us
connectivity
at
the
locations
where
we
need
to
distribute
our
network.
P
And
yes,
it
snows
in
San,
Diego
I
like
putting
that
one
up
there,
because
San
Diego
its
beaches.
So
we
have
locations
that
are
very
beautiful
and
very
remote.
We
see
the
below
there,
a
homestead
on
one
of
the
reservations
and
and
how
far
away
is
from
other
existing
infrastructure
and
wireless
is
the
way
to
communicate
with
those
folks
there's
some
more
tower
locations
that
have
been
built
over
the
years,
really
small
repeaters
to
be
able
to
reach
out
further
and
further
into
community
and
then
I'll
get
to
a
map.
There
we
go
now.
P
I
can
start
talking
about
some
of
the
core
problems
that
have
happened.
So
in
the
United
States
20
years
ago,
ish
about
seven
networks
were
built
across
east
to
west
west
to
east,
whatever
you
want
to
do,
and
they
were
valued
on
miles
built,
fiber
miles,
built
right.
That
was
the
value
of
the
company,
that's
where
they
got
their
money.
They
got
their
loans,
their
evaluation,
so
to
go
across
the
country
as
fast
as
you
could
building
fiber,
you
avoided
things
like
the
Grand
Canyon
or
the
Rocky
Mountains,
or
you
know
a
giant
lake.
P
You
didn't
go
through
those
things
because
they
slowed
you
down
or
you
had
to
go
around
them
to
make
to
make
it
easier
to
build
fiber
miles.
You
also
avoided
Native,
American
reservations,
I'm
sure
it's
the
same
in
Canada,
but
this
is
the
US
map
and
you
see
along
these
paths.
This
is
the
fiber
paths
that
exist
in
the
United
States.
This
was
done
out
of
I,
think
Oregon,
State
University.
P
The
map
overlays
and
I
worked
with
the
Obama
administration,
with
a
CTO
of
the
United
States,
to
define
this
very
map
with
the
carrier's
data,
actual
KMZ
files
laid
over
top
of
tribal
of
reservation
boundaries
to
determine
how
many
missing
middle
mile
fiber
miles
there
were.
There
are
8,000
miles
of
missing
middle
mile
fiber
to
to
just
connect
the
lower
48
tribes
and
that's
320
reservations,
so
8,000
miles
are
missing
just
to
get
middle
mile
fiber
to
320
reservations
in
the
lower
48
alaska
is
a
whole
nother
problem.
So
you
see
the
big
red
circles.
P
Those
are
what
they're
calling
in
the
connectivity
deserts.
Those
big
red
circles
are
full
of
tribes,
they're,
also
full
of
sparsely
populated
rural
America,
small
town
USA
people
that
have
been
left
behind.
There
is
no
service
in
these
areas,
a
lot
of
times
no
cellular
service.
So
what
do
you
do
solve
that?
You
can
build
a
community
Network
from
the
ground
up
which
we've
done
at
the
tribal
digital
village
network
in
Southern
California.
P
But
then
our
biggest
problem
was
in
the
last
six
years
to
try
to
figure
out
how
to
get
transport
to
the
rest
of
the
world
at
an
affordable
cost,
because
we
didn't
have
access
to
fiber.
We
were
paying
$99
per
megabit
per
month,
u.s.
dollars.
So
obviously
you
can't
scale
that
and
survive.
So
for
the
last
six
years,
we've
been
navigating
the
policy,
the
state
government
of
California,
the
different
providers
in
the
region
and
solving
this
problem.
P
There
is
another
way
to
look
at
this
problem
and
and
to
solve
it
as
well,
and
that
is
one
of
the
things
that
I
am
involved
with
the
Arcadian
infra
comm
company.
So
the
other
way
to
do
it
is
to
take
the
people
who
are
driving
the
path
of
the
Internet
and
who
is
that
right?
Is
that
18
t
Verizon
any
of
the
telcos?
No,
it's
no
longer
the
carriers,
it
is
the
over-the-top
hyper
scalar
data
companies
right,
it's
Google,
Amazon,
Microsoft,
Facebook
and
Apple.
P
Those
five
are
driving
where
fiber
needs
to
be
put
based
on
placement
of
data
center.
They
set
a
data
center
down.
They
need
three
minimum
three
diverse
routes
out
of
that
data
center
to
be,
you
know,
within
their
network
plan,
they're
looking
for
four.
So
if
you
look
at
the
black
lines
on
the
map,
those
have
been
built,
rebuilt
and
rebuilt,
and
people
are
still
building
on
them
and
there's
no
diversity
to
those
routes.
You
know
the
fiber
gets
cut
in
that
location.
P
That's
all
laying
in
the
same
ditch
if
a
backhoe
rips
open
an
entire,
ditch
and
cuts
all
the
fiber
that
whole
regions
down.
If
you
look
at
the
United,
States
Denver
and
El
Paso,
if
you
chop
out
fiber
and
Denver
and
El
Paso
East
does
not
talk
to
west
until
you
get
it
fixed
right.
There
are
no
real
routes
that
go
back
and
forth
across
the
US,
so
it's
time
to
draw
new
lines
on
the
map.
How
are
we
gonna?
Do
that?
P
So
then,
let
me
flip
back
up
to
you.
I'm
gonna
go
backwards
for
a
second
and
that
one
okay
cool,
so
I
just
wanted
a
few
minutes
to
wrap
up.
I
just
wanted
to
talk
about
the
lack
of
opportunities
in
the
u.s.,
the
changing
environment
that
you
know,
working
with
different
administrations
and
things
like
that
really
fast.
P
P
If
I
have
a
giant
area
and
I
have
one
house
connected
in
that
area,
that
entire
area
gets
counted
in
that
report,
so
it's
false
and
it
doesn't
have
the
true
data
so
that
true
data
we're
trying
to
work
with
the
general
Accountability
Office
to
define
what
that
true
data
looks
like
we're
trying
to
get
the
FCC
to
change
their
data
recording
methods.
Everyone
is
to
get
the
reality
of
that
to
be
able
to
define
better
what
it
looks
like
on
tribal
reservations,
Indian
communities,
American
communities
do
not
participate.
P
Typically
in
surveys,
they
don't
typically
want
to
be
counted.
Every
time,
they've
been
counted
in
the
past,
they've
been
eliminated.
So
every
time.
How
many
are
you
you're?
A
hundred
okay,
we're
gonna,
kill
50
of
you
right,
because
that
you're
too
many
so
it
you
know
what
there's
a
there's
a
stigma.
There
there's
a
history
there
that
they
don't
like
to
participate
in
every
time
they
get
counted.
Something
bad
happens,
so
it
needs
to
be
an
understanding
of
you
know,
you're
being
misrepresented.
P
You
know
we
need
to
collect
the
information
so
that
we
can
understand
how
to
better,
represent
you
in
this
fiber
plan
and
and
build
out
plans.
So
there
are.
There
are
some
obstacles
to
education
and
awareness
to
be
overcome,
that
we
understand
that
that
Wireless
is
a
great
opportunity
as
a
stopgap
or
as
a
half-measure
to
get
to
fiber,
and
it
may
be
a
complete
and
total
solution
if
the
fiber
is
impossible.
P
But
we
are
striving
to
get
fiber
to
each
of
these
communities,
at
least
into
the
community,
to
be
able
to
have
bigger
transport
and
be
able
to
support
more
throughput
to
the
end-user
at
what
we
call
the
first
mile,
not
the
last
mile.
The
first
mile
is
the
one
that
counts
and
that's
pretty
much
it.
If
anybody
has
any
questions.
M
Question
this
is
Dino
Elizabeth
I
met
us
you
over
to
top
vendors
that
are
trying
to
bring
fiber
through
you
mentioned
that
they
may
have
to
connect
libraries
in
public
places
and
stuff
is:
are
they
resistant
to
that
or
are
they
just
bringing
one
fiber
strand
down
it's
responsibility
of
the
individuals?
That's
happened
to
it,
or
do
they
not
allow
it
so.
P
They're
not
even
actually
involved
in
that
component
of
it.
So
when
the
build
happens
from
data
center
to
data
center
they're
just
giving
getting
delivery
from
each
end
in
the
middle,
we're
dragging
the
line,
the
company's
dragging
the
line
into
the
region
and
then
offering
up
connectivity
at
that
location
and
and
then
we're
not
building
out
the
community
side.
We're
partnering
with
folks.
That
would
do
that.
Build
out
and.
M
P
There
demarked,
like
so
like
Google
or
Facebook's
responsibility
or
ours,
okay,
okay,
so
so
the
the
over
the
top
players
are
just
concerned
about
traffic
from
A
to
B
and
that
transport
long
haul
and
we
will
be
managing
loops
and
small
accessibility.
You
know
huts
and
handholds
along
the
way
of
the
route,
to
be
able
to
support
those
communities
and
be
able
to
give
either
lit
services
or
strands
of
fiber
to
a
third
party
that
will
be
building
that
component
out
and
managing
that.
B
C
Q
Okay,
so
thank
you
for
having
me
here.
I'm.
Not
though
I
am
professor
at
the
university
of
Venice.
Up
to
May,
I
was
a
researcher
at
the
University
of
Trento
in
Italy
and
I
was
the
work
package,
technical
coordinator
or
a
project
that
is
called
net
Commons.
He
ended
in
2019
in
March
and
we
light
it
at
the
University
of
Trento
with
other
partners
like
UPC
and
I'm.
Q
Q
Q
Q
What
we
are
the
components
that
we
use:
open,
the
open
data,
surface
models,
path,
loss
models
and
an
engine
that
simulates
the
growth
of
the
network.
So,
let's
let
me
just
go
through
the
three
points,
so
yeah
that
just
just
a
little
warning
we
did
this.
The
code
is
already
on
github,
it's
a
bit
done
in
a
rush
for
at
that
time
you
know
but
see
we're
working
on
it,
and
there
are
a
lot
of
things
that
we
can
actually
improve.
But
we
are
happy
to
I'm
really
happy
to
receive
any
kind
of
feedback.
Q
This
is
Florence,
it's
a
city
in
Italy
we're
starting
with
data
from
Italy,
because
you
know
we
know
the
language.
You
know
how
to
interpret
other
things
and
yeah,
and
this
is
the
pink,
is
ground
level
white,
it's
buildings,
and
this
is
a
hill
that
is
just
in
front
of
my
house
and
specifically,
and-
and
this
is
the
same
thing
from
another
3d
point
of
view,
so
those
the
mountains
here,
they're
there,
the
buildings,
the
the
thing
on
top
over
there.
Q
Q
This
was
on
this
was
in
Indian,
because
Leon
has
a
very
nice
in
France
as
a
very
nice
data
set
with
a
lot
of
precision,
and
basically
this
is
the
same
link
that
we
were
testing
using
different
kind
of
datasets.
The
first
one
is
the
most
precise
one.
The
second
one
is
the
down
sample
and
the
third
one
is
it's
a
pretty
famous
produced
database
that
is
caused
SRTM,
that
you
can
get
it
for
free
from
the
nasa
it's
for
over
the
world.
But
it's
precision.
Q
So
we
collected
data
from
a
few
of
those
gears
and
so
given
a
path
loss,
we
can
actually
try
to
understand
what
would
be
the
bitrate
that
those
two
specific
devices
could
actually
negotiate
and
since
we
know
the
price
of
these
devices,
we
can
also
estimate
what
is
the
best
couple
of
devices
to
make
that
specific
link,
and
this
is
something
that
we
are
still
working
on.
So
it's
just
a
very
simple
web
interface
where,
basically,
you
actually
pick
two
points
in
space.
Q
You
choose
how
long
out
all
your
poll
where
you
can
mount
your
devices
would
be,
and
you
pick
the
devices,
and
this
is
going
to
tell
you
if
there
is
an
off
site
and
it's
going
to
give
you
a
rough
estimation
of
the
bit
rate.
That's
gonna
be
negotiated
depending
on
a
lot
of
heuristics
that
I'm
not
going
to
then
I'm
not
going
to
go
through
here.
Q
So
now
we
have
a
way
to
estimate
if
a
link
is
possible
and
we
were
discussing
at
the
bottom
mesh
in
Paris
with
with
other
communities
other
wireless
communities,
how
the
way
that
they
could
use
it,
and
there
were
other
communities
that
were
doing
similar
stuff
is
in
different
data.
So
this
we
had
quite
a
good
reception
from
them
and
but
then
we
try
under
research
point
of
view.
Q
Q
Is
you
just
pick
the
one
that
is
giving
you
the
best
bandwidth
on
that
link,
which
is
it's
just
the
most
natural
thing
that
you
do?
Actually
we
will
imagine
if
we
can
do
something
better
than
that,
for
instance,
since
we
have
a
metric
that
is
going
to
determine
when
the
growth
will
stop
well,
you
can
actually,
you
could
actually
pick
the
neighbor
that
is
actually
producing
the
metric.
That
is
bad.
Q
That
is
bad
for
the
nephew,
so
the
one
that
the
neighbor
that
will
actually
give
you
the
best
metric
in
your
network
evolution-
and
this
is
just
one
of
the
things
that
that
we
can
do.
Okay,
I
have
this
number
some
numbers
here,
but
I
don't
want
to
go
much
into
the
details.
What
I
want
to
show
you
here
is
that
you
have
a
minimum
diff
minimum
bandwidth
per
user,
and
this
is
the
number
of
users
that
you
can
achieve
number
of
nodes
that
you
can
achieve
with
respecting
that
stopping
condition
in
values
Yubin
environment.
Q
Yet
we
know
it
and
and
this
with
a
greedy
approach,
so
you
get
between
150
and
200
nodes
for
the
minimum
bandwidth
for
one
megabit
per
second,
and
if
you
use
the
network
approach,
you
can
more
than
double
it.
Basically,
so
if
you
tell
if
you
give
the
instrument
to
the
community
and
they
have
to
make
this
decision,
you
can
actually
help
them
to
make
the
network
scalable
or
well
scale
better
and
yet.
Q
Or
less
everything,
so
we
did
this
tool
to
model
the
growth
of
the
network
and
the
methodology
can
actually
replicate
it
using
data
from
any
place
and
we
are
giving
away
the
code,
and
we
are
also
trying
to
understand
if
you
can
include
datasets
from
other
countries,
so
the
people
on
just
use
our
servers,
which
are
not
actually
know
public
now,
but
they're
gonna
be
and
yeah,
and
then
we'd
have
a
lot
of
ideas
to
keep
developing
this
tool
in
several
directions.
Just
a
few
thing
is
one
thing
now
we're
picking
nodes
at
random.
Q
It
would
be
Lin
really
nice
if
we
could
actually
pick
notes
with
some
other
criteria.
So
we
could
take
data
from
national
surveys
and
understand
who
is
living
in
that
building
more
or
less,
because
data's
are
pretty
detailed,
and
so
you
can
understand
what
in
average
e.
You
have
people
with
this
education,
this
age,
and
so
they
might
be
more
likely
to
want
this
link.
And
if
you
get
this
kind
of
data,
then
you
can
actually
try
to
understand
where
your
network
cables
not
just
a
random
topology.
Q
But
in
that
specific
city,
where
it
is
it
going
to
go
so
which
area
in
that
city
are
going
to
be
served
or
are
not
going
to
be
served
and
that's
it
credits
code
by
myself,
Gabriela
Jamie
bad
code
by
myself,
a
good
code
by
gabriela
and
daniela,
made
the
interface.
And
then
you
find
a
lot
of
details
in
in
this
paper,
which
has
was
actually
accepted
ago.
R
Q
Do
I
get
them
from
public
administrations
like
in
in
Tuscany?
There
is
there's
a
website
that
gives
you
the
delighted
map
in
in
France
I
think
that
each
part
of
the
nation
is
covered
by
other
map
and
you
can
get
them,
and
some
of
them
are
also
even
integrated
in
OpenStreetMap
in
some
places.
So,
but
this,
for
instance,
in
the
guys
from
New
York,
they
have
the
whole,
not
the
ladder
map,
they
have
the
whole
3d
model
of
the
whole
city.
So
they
do
something
similar.
B
I
Joseph
that
economists
foundation
any
pushback
from
other
suppliers.
I
Q
I
I
Q
That's
that's
a
good
question
and
I
mean
I
can
ask
you.
I
can
see
you
on
the
basis
of
the
project
that
we
did.
While
we
were
talking
with
many
different
communities
and
it
depends
a
lot
on
the
place,
for
instance
in
Italy,
it's
actually
illegal
right
now
to
do
trunking
on
five
gears,
so
Niyaz
I'm
bandwidth,
which
should
already
killed
the
whole
thing
itself,
something
that
we
discovered
six
months
ago.
In
other
places,
there
is
like
small
operators
actually
behave
well
with
community
networks.
Q
Q
B
Through
Friday,
great
so
Chris
you're
at
and
to
the
gentleman
that
just
asked
the
question,
some
of
us
have
targets
on
our
backs.
Because
of
this
there
is
a
lot
of
pushback
around
the
world
for
the
small
networks,
which
is
really
it's
really
strange
to
us,
because
there
are
small
community
networks,
servicing
communities
of
5,000
and
under
generally,
where
the
return
on
investment
for
the
bigger
companies
isn't
met.
So
we
don't
quite
understand
why
the
model
isn't
accepted.
I
R
Hi
everybody
I'm
Curtis
Amaral
assistant,
professor
at
the
University
of
Washington.
The
purpose
of
this
talk
is
a
little
bit
of
grenade,
throwing
in
a
light
light
very
high,
be
supportive
way.
So
this
is
a
really
early-stage
project
that
we're
just
spinning
up
in
Seattle
and
what
I'd
like
to
do
is
be
told
it's
a
bad
idea
if
it's
a
bad
idea-
and
hopefully
it's
a
good
idea-
and
maybe
that's
what
we'll
get
out
of
this.
R
So
yes,
hello,
I'm,
Curtis,
I'm
part
of
the
ICT
D
lab
at
the
University
of
Washington.
This
is
a
long-running
research
lab
focused
on
the
intersection
between
technology
and
poverty.
Professor
Richard
Anderson
is
the
other
co-lead
of
the
lab.
He
runs
financial
services
and
health
services
and
I
do
universal
internet
access
and
conservation.
So
my
lab
for
quite
a
while
now
has
been
working
on
community
networking,
and
we
were
already
talking
about
this
on
a
brief
introduction.
Just
in
case
some
people
didn't
haven't
been
following
us
for
a
while.
R
These
are
networks
built
owned
and
operated
by
the
citizens
and
users
in
a
participatory
and
open
manner.
We
fee
being
the
juggernaut
in
the
space
at
thirty
five
thousand
nodes,
just
this
amazing
effort
in
Catalonia,
but
there's
many
many
of
these
networks
now
for
iPhone
kalter,
mundi,
NYC
mesh,
all
these
sort
of
networks,
spinning
up
throughout
the
world
and
all
built
on
these
eight
or
to
11
mesh
protocols
which
we're
all
hearing
a
little
bit
about
there,
and
this
is
there's
a
whole
bunch
of
reasons
for
this.
R
Unless
you
operate
an
unlicensed
spectrum,
largely
we're
at
the
IETF,
we
understand
IP
networks,
these
are
IP
networks
and
so
we're
able
to
use
these
in
sort
of
an
intuitive
way.
My
group,
in
particular,
has
been
working
on
a
subset
of
community
networks
known
as
community
cellular
networks.
These
are
community
networks,
utilizing
the
cellular
protocols
that
other
than
the
Wi-Fi
protocols.
We
build
off
of
some
open
implementations
of
the
cellular
standards,
open
BTS
and
Osmo
com4,
the
2g
stuff,
open
air
interface,
SRS
LTE
for
the
LTE
stuff.
There's
a
couple
examples
of
these
networks.
R
Many
many
less
than
the
you
know:
211
networks,
rhizomatic
of
being
a
juggernaut
in
this
space,
then
Oaxaca
our
own
networks
in
the
Philippines
and
Indonesia.
Similarly,
my
group
has
recently
not
that
recently,
probably
the
last
couple
years
transitioned
over
to
focusing
on
LTE,
specifically
a
project
called
community
LTE
or
coltie,
which
Spencer
right
there
runs,
and
so
we've
been
deploying
that
into
our
communities.
Recently,
there's
a
network
up
in
Oaxaca
I
have
a
student
there
right
now,
as
well
as
a
network
up
in
Indonesia
and
hopefully
we'll
be
scaling
from
there.
R
These
networks
have
largely
focused
on
sort
of
traditional
rural
access
problem
stuff,
like
what
Matt
was
talking
about
just
saying:
hey
there
areas
without
connectivity,
so
there
is
a
good
fit
for
that
right
and
then
we
come
up
with
a
normal
rural
access
problems.
Hey
you,
don't
have
backhaul,
let's
do
local
services
all
these
kind
of
things,
but
our
group
has
sort
of
been
iterating
a
little
bit
and
we
started
asking
ourselves.
The
following
question:
is
community
cellular
appropriate
for
cities
as
well
right?
So
you
look
at
our
work
on
community.
R
So
it's
all
rural,
but
you
look
at
mesh
networks,
things
like
NYC
mesh.
You
don't
get
a
lot
less
rural
than
New
York
City,
and
so
we
were
trying
to
see
if
maybe
this
technology
could
be
applicable
in
that
domain
as
well,
and
so
I'll
go
through
a
little
thought.
Experiment
on
Wi-Fi
networks
versus
cellular
networks,
so
the
biggest
Wi-Fi
network
I
can
find
in
the
world
is
a
thing
called
fawn.
Has
anyone
in
the
room
used
Fon,
zero,
ok,
white
one?
It's!
R
It's
reported
this
twenty
three
million
hot
spots
throughout
primarily
Europe,
which
is
probably
why
it's
a
little
bit
of
an
unfair
question
to
ask
here
distributed
around
and
the
idea
is
basically
similar
to
your
Xfinity
Wi-Fi
at
home
right.
So
the
way
Fon
works
is
they
put
their
software
onto
a
router?
You
run
the
router
and
by
being
a
Fon
customer,
you
can
roam
on
to
other
Fon
routers
in
your
town
right,
and
so
this
is
I
think
the
largest
Wi-Fi
network
on
earth.
R
This
is
what
Fon
coverage
looks
like
in
downtown
London,
so
23
million
hotspots,
but
you
can
kind
of
see
some
of
the
limitations
of
trying
to
build
city
scale,
Wi-Fi,
networks,
right
and
so
there's
a
whole
bunch
of
these
first
Wi-Fi
is
just
bad
at
city
scale.
You
can
see
all
these
little
dots
are
little
dots
and
that's
the
nature
of
Wi-Fi
right
Wi-Fi
has
been
selected
to
operate
in
these
bands
that
have
low
propagation,
and
so
it
doesn't
go
very
far.
R
Wi-Fi
is
really
bad
at
mobility.
It
was
just
never
designed
into
the
protocol
mobility
being
the
ability
to
move
between
different
access
points.
So
you'll
lose
your
connection
and
you
lose
everything
as
you
go
along.
This
is
solved
by
some
of
the
centralized
Wi-Fi
stuff,
but
you
don't
see
that
that
commonly
and
again
the
limited
coverage
area
makes
it
just
hard
right.
You're
gonna,
run
into
dead
start
pretty
fast,
leaving
your
house
even
if
handover
worked,
and
lastly,
Wi-Fi
is
bad
at
Spectrum
coordination.
R
We
see
this
with
the
mesh
network
work
if
the
network
is
really
really
dense.
Communication
happens
if
the
network
is
really
really
sparse
or
you
lose
connectivity
between
the
hops.
So
it's
really
hard
to
build
these
kind
of
city
scale,
Wi-Fi
solutions
and
now
we'll
give
an
example
of
a
city,
sale,
Skelly,
a
cellular
solution.
This
is
British
Telecom.
This
is
British
telecoms
map
of
London.
The
green
areas
are
areas
with
good
coverage.
It's
all
of
them.
It's
the
river,
it's
everything
right,
and
so
why
is
this
easier?
R
Well,
there's
a
lot
of
business
reasons
which
I'm
at
the
IETF
so
I'm
going
to
punt
on
those
it's
actually
a
little
unfair
to
punt
on
those,
because
that's
a
whole
side
discussion,
but
we're
community
networking
people,
so
we
can
punt
on
this
in
the
short
term.
Similar
is
really
good
at
white
area
right.
It's
designed
for
this
it's
kilometres
of
coverage
and
that's
what
it
was
built
for
in
the
start.
So
there's
great
of
mobility.
It's
a
level
2
element
of
cellular.
It's
like
a
core
requirement.
R
The
2g
standard
was
built
specifically
to
allow
people
to
move
on
trains
in
France,
like
that
was
in
the
specification
to
be
able
to
do
that.
Lte
actually
includes
peer-to-peer
mobility
primitives
in
the
protocol,
so
the
Eno
B's,
the
access
points
talk
directly
to
each
other
to
do
mobility
and,
lastly,
it's
good
at
spectrum
coordination.
This
is
literally
what
cellular
means.
So
there
is
spectrum
coordination,
it's
how
to
coordinate
spectrum
inside
of
a
city,
so
we
don't
waste
any
of
it
and
then
there's
all
these
technologies.
R
As
its
said,
on
top
of
that
songs,
self-organizing
networks
being
the
sort
of
standard
to
be
able
to
coordinate
spectrum,
so
white
cellular
networks
are
good
for
city
scale
coverage.
Okay,
this
shouldn't
have
been
too
much
of
an
army
with
anymore.
Why?
What's?
Stopping
us
community
networking
people
from
building
this
stuff?
Well,
there's
sort
of
standard
set
of
answers
to
this
problem.
Spectrum
right,
so
Lola
loses
license.
Spectrum
and
spicen
spectrum
is
brutally
difficult.
Well,
it
turns
out,
that's
been
shifting
it
specific
quite
a
bit
recently.
R
First
with
the
cvrs
band,
this
is
a
citizen's
broadband
radio
system
band
48
LTE.
That
is
a
neat
tri-level
licensing
regime
that
has
frequency
open
for
small
users.
You
don't
forget
a
telecom
to
get
that.
In
fact,
there
are
always
bands
available
everywhere
for
non
telecom
users,
so
that's
great
and,
moreover,
the
cellular
people
are
creeping
in
on
our
business
and
so
now
that
there
is
actual
cellular
protocols
to
run
in
Wi-Fi
bands.
This
is
LTE
you
and
LTE
LA.
R
Those
are
designed
to
coordinate
with
the
existing
Wi-Fi
installations
to
allow
LTE
to
op
right
there.
So
cellular
spectrum
is
getting
a
lot
easier
and
has
in
the
last
five
years.
Okay,
cellular
hardware
is
super
expensive
right.
Well,
not
since
LT
e
2g
access
points
were
like
tens
of
thousands
of
dollars.
Ten
years
ago,
there's
been
a
big
drive
to
commoditize
those,
and
so
now
2g
access
points
or
5k.
An
LTE
node
B
that
you
can
run
that
we
run
in
Papua
is
$2,500.
R
A
new
radio
5g
is
pushing
that
even
farther,
so
they
built
a
commodification
about
a
kind
of
commoditization
model
into
there.
That's
making
us
all
much
more
affordable
as
we
go
forward.
As
we
mentioned,
telecoms
are
kind
of
difficult
to
work
with
and
so
well,
what
are
we
gonna?
Do
we
need
telecoms
right?
No
LTE
is
IP
based
entirely.
It's
just
a
big
Wi-Fi
network
and
so
do
what
you
do
in
your
Wi-Fi
network
run
things
over
the
top
run.
R
What's
up
and
you're
running
a
wide
area
network,
without
the
telecoms
necessarily
being
able
to
get
into
your
way,
you
don't
need
phone
numbers
and,
lastly,
cellular
equipment
is
really
hard
to
use
and
this
is
actually
I
think
the
most
exciting
one
and
sort
of
brought
me
into
this
room
to
talk
about
this
is
that
you
know
LTE
isn't
designed
for
the
little
guys
right.
Yeah.
Everyone
in
here
can
run
a
Wi-Fi
hotspot,
but
you
can't
do
that
with
LT.
Necessarily
this
is
shifting
right
now.
R
One
data
point
is
the
Wireless
ISP
model
in
the
u.s..
There
are
hundreds
of
wireless
ISPs
in
rural
America
running
LTE
equipment.
They
do
that
they
hand
you
the
radio
receiver
and
it
has
a
sim
card
built
in
and
you
just
have
access
in
the
wide
area.
So
this
is
already
happening
so
sort
of
hacker
ish
folks
are
running
these
networks
at
scale
in
the
US.
A
second
thing:
that's
happening
is
private
LTE,
and
this
is
all
IOT
buzz,
but
it's
actually
getting
built.
R
So
these
are,
let's
say
you
own
a
factory,
and
you
want
to
do
do
IOT
in
your
factory.
Well,
what
you
can
do
is
install
a
small
cell
phone
network
just
in
your
building
against
smaller
and
smaller
organizations
participating
in
building
out
this
kind
of
networks.
Lastly,
carrier
aggregation
is
I,
think
the
most
exciting
thing.
This
is
going
to
be
a
huge
element
of
the
middle
term
future
and
into
new
radio.
R
What
it
is
is
in
buildings
like
this,
the
owner
can
install
LTE
in
the
building
and
what
they'll
do
is
they'll
loan,
that
out
to
the
carriers,
the
VMO
knows
and
charge
a
little
bit
of
overhead
on
that.
So
basically,
every
office
building,
every
residential
metropolis
building
is
going
to
have
building
owned
network
in
it,
and
for
that
to
be
a
community
network
is
not
significantly
different
than
what
it
is
now.
You
just
pointed
at
a
different
carrier:
aggregator,
okay,
so
this
is
kind
of
the
point.
R
I
think
this
is
a
great
opportunity
right
now
just
start
to
build
out
community
LTE
in
cities,
so
I
went
too
fast.
So
what's
the
plan,
what
are
we
doing?
What
is
it?
What
is
what
are
my
students
actually
building?
It's
it's
largely
this
right.
So
our
technology
agenda
is
the
thing
we're
calling
federated
LTE
and
the
big
idea
is
we're
using
distributed
ledger
as
the
backend
to
store
a
whole
bunch
of
the
core
network
required
people
know
amber
we've
sort
of
hedged
our
bets
on
doing
amber,
esque
stuff
by
thinking.
R
Lte
is
the
better
access
medium
for
all
the
reasons
I've
given
earlier.
The
big
idea
here
is
make
it
so
that
anybody
can
join
this
Federation.
Let
you
stand
up
a
new
LTE
network.
You
pointed
at
the
right
thing:
you
join
the
Federation.
All
the
sim
cards
work,
all
the
roaming
works,
all
that
kind
of
stuff,
and
instead
of
just
doing
the
billing
as
the
Federated
element,
which
is
what
amber
did
before
we're
gonna
try
to
do
everything
else.
You
saw
authentication
and
mobility
and
spectrum
coordination.
R
Similarly
in
the
sky,
in
a
way
that
everybody
can
peer
together.
In
a
distributed
decentralized
fashion,
so
this
isn't
a
hugely
technical
talk,
so
I'm
gonna
hand
away
of
all
of
that,
but
we
are
building
things
and
we're
deploying
things,
and
so
this
is
our
planned
deployment
here
in
the
fall
in
Seattle.
So
we
have
access
to
two
rooftops
from
the
University
of
Washington.
R
Think
it'll
be
the
first
community
cellular
installation
in
a
city
that
I
know
of
a
key
point
here
and
kind
of
one
of
the
reasons
to
be
here
is
that,
as
a
nature
of
a
federated
network,
anyone
here
is
welcome
to
join
and
we'd,
be
happy
to
point
you
on
how
to
do
this
kind
of
stuff.
You
stand
it
up
and
we'll
send
you
some
SIM
cards
and
you
can
roam
onto
our
network
and
we
can
roam
on
to
yours
and
hopefully,
everybody's
friendly
and
happy
all
right.
Thanks,
excellent.
N
N
R
The
way
that
it
works
technically
is
that
your
radio
will
connect
to
whatever
the
core
network
is
somewhere
right
and
that's
the
organization
you're
talking
about
that
has
some
sort
of
privacy
guarantees,
and
so
even
in
the
roaming
network,
in
like
you're,
a
carrier
aggregation
side
of
a
building.
If
you're
an
AT&T
customer
they'll
build
a
pipe
for
you
to
go
to
AT&T,
and
then
your
your
data
is
all
sensibly
private,
at
least
the
AT&T.
R
Now
in
our
system,
we're
actually
removing
a
whole
bunch
of
this
for
a
bunch
of
technical
reasons,
users
are
saying
we
should
be
doing
with
the
internet,
doesn't
do
end
and
privacy
on
basically
everything.
It's
only
handling
IP
traffic.
If
you're
sending
me
unencrypted
things
over
that,
it's
not
going
to
help
you
or
it's
not
gonna.
It's
not
gonna,
be
a
good
idea,
but
yeah.
R
Sure
so,
traditionally,
backhaul
limitations
have
been
a
huge
deal
in
rural
access,
specifically
royalty,
our
network
and
Bob,
who
has
a
three
megabit
per
second
backhaul.
Our
front-hall
connection
on
that
thing
like
that.
For
knows
the
wrong
word,
but
the
connection
to
the
user
isn't
like
a
hundred
megabits,
and
so
it's
just
the
mismatch.
There's
a
huge
deal.
The
nice
thing
about
what
we're
doing
here
is
the
City
of
Seattle
in
particular,
has
fiber
all
over
the
place,
so
we're
just
plugging
into
that.
We're
gonna
be
good
to
go
so
in
general.
R
S
R
M
R
It's
mostly
so,
in
addition,
is
the
sort
of
reality
facing
version
of
it
right.
So
if
you
have
an
apartment,
building,
people
are
going
to
put
in
their
own
their
own,
their
own
Wi-Fi,
but
what
they
want
to
do
is
have
like
people
coming
into
the
building
have
to
sell
over
the
whole
way
through
and
in
this
whole
business
pipeline,
to
be
able
to
get
that
money
out
of
AT&T
and
into
the
the
the
building
owners
and
I
think
in
the
future.
M
R
M
R
There's
a
we
should
talk
about
it
offline
because
I
have
a
whole
bunch
of
like
fun,
Finn,
sim
hacky
solutions
to
solve
some
of
these
things
at
the
end
of
the
day.
Hopefully,
what
it'll
be
the
Federation
is
another
element
of
the
Emma
nodes.
You
can
connect
you.
So
if
you
have
a
federated
sim,
you're
able
to
you
know,
roam
on
to
that
network
and
participate
in
the
same
way,
but
yeah.
R
So
this
so
the
band
48c
BRS
band
is
currently
has
a
set
of
nationwide
licenses
that
have
been
doled
out.
So
we
have
a
partner
who's
gonna.
Let
us
use
that
in
april,
that's
gonna
roll
over
to
the
actual
SAS,
the
spectrum,
not
auction
system,
maybe
an
allocation
system
which
does
sort
of
an
automated
system
of
grabbing
spectrum,
in
which
case
we're
going
to
be
deploying
a
thing
that
lets
us
do
that,
so
that
we
grab
the
spectrum
as
it
goes.
B
C
D
D
C
How
to
upgrade
the
document,
how
to
keep
it
up
to
date
and
what
I
think
it's
now
one
we
need
to
the
three
so
I
think
it's
much
harder
for
everyone
like
those
who
came
and
they
interesting
ideas.
Those
who
have
some
things
to
discuss
after
the
meeting
and
well
I
hope
you
enjoyed
I
hope.
The
information
is
useful
for
the
m-40
percenters.