►
From YouTube: IETF102-GAIA-20180717-1550
Description
GAIA meeting session at IETF102
2018/07/17 1550
https://datatracker.ietf.org/meeting/102/proceedings/
A
Hello
we're
going
to
get
started.
This
is
the
gaia
group.
I
am
Jane
coffin
and
this
is
Leandro
Navarro,
my
partner
in
crime
for
Gaia
we're
the
co-chairs.
If
you'd
like
to
see
our
Charter
online,
just
take
a
look
under
Gaia.
We
are
and
have
been
rebooting
and
working
with
community
networks.
Local
access
network
types,
people
are
hacking
code
to
come
up
with
interesting
presentations
that
are
going
to
lead
us
to
happily
a
BCP
that
we're
going
to
work
on
with
all
of
you
over
the
next
months
until
we
get
to
Thailand.
So.
A
There's
something
called
a
note.
Well,
if
you
don't
know
what
it
is,
go
online
and
have
a
look,
because
it
would
be
useful
for
you
to
know
what
that
means,
I'm
not
going
to
go
through
it.
The
last
time
I
did
someone
laughed
at
me
and
said:
Jane,
you
don't
have
to
read
out
the
note
well,
so
in
any
way,
in
any
event,
I
also
want
to
be
very
cognizant
of
time,
because
crypto
is
after
us,
so
we're
gonna
try
and
stick
to
our
time
sheet.
There
are
four
people
giving
presentations.
A
There's
one
person
who's,
not
here,
who
couldn't
be
min
either
eyes
remotely
they'll
speak
in
Thailand.
His
presentation
I'll
show
at
the
very
end,
it's
a
link
to
something
he's
done
on
open,
whisp,
so
guy
a
working
group
in
AI
RTF
initiative,
I'm
gonna
beam
through
this
fast
global
access,
is
a
human
right.
Obviously,
as
we
know,
there's
need
for
infrastructure.
I
was
actually
just
up
in
northern
Greece
in
a
community
network
where
people
are
using
a
university
backhaul
to
help
put
in
a
Wi-Fi
network.
A
Amazing,
stuff
and
people
are
rejuvenating
purpose,
they're,
bringing
in
more
business
and
so
we're
here
to
try
and
bring
you
back
down
a
reality.
Some
of
you
work
at
a
super
high
level
from
a
technical
perspective.
You
cannot
have
IOT
without
the
eye.
You
cannot
have
a
smart
city
without
smart,
so
in
some
places
there's
a
lot
of
infrastructure.
That's
not
built.
49
percent
of
the
people
in
the
planet
are
not
connected,
so
there
are
Politan
we're
working
out
some
bugs
here
with
a
new
process.
Obviously
there
are
public
policy
challenges.
A
What
we're
doing
also
from
a
research
perspective,
is
working
with
those
of
you
out
in
the
community
who
are
testing
different
interfaces
looking
at
different
equipment,
how
it's
working
that
project
in
Greece
I
just
mentioned
just
changed
out
all
of
their
gear
to
ubiquity
recently,
because
it's
cheaper,
better,
faster
and
also
just
the
price
point.
Was
there
I'm?
Not
here
to
promote
ubiquity
but
I'm,
just
saying
the
price
points
have
come
down
so
much
that
local
access
networks
are
doable
and
we're
also
here
to
help
foster
more
collaboration.
A
We
all
know
eye-to-eye
contact
working
with
people
not
just
online,
is
also
critical
and
part
of
our
objectives,
which
are
also
in
line
in
our
Charter
to
share
more
information
to
document
it
and
the
documentation
part
we're
hoping
to
Leandra
and
I'll
sit
down
tomorrow
with
the
terms
of
reference
draft
for
you
join
us.
If
you'd
like
we'll,
put
a
note
out
on
the
Gaia
list,
where
we
could
meet
up
and
hack
it
in
real
time,
get
it
out
on
the
list.
A
Have
people
take
a
look
at
it
and
then
move
forward
with
the
BCP,
so
I'm
not
going
to
speak
much
longer,
but
some
of
our
outcomes
are
this
connection
of
people
getting
the
BCP
together.
Talking
to
folks
like
who
are
working
on
hacking
and
Batman
BMW
I
was
sorry.
Bmw
BMX
I
was
just
with
some
of
the
battle
mesh
folks
last
week.
There
are
great
for
around
the
world
where
some
of
this
is
happening
where
we
can
bring
in
some
of
their
findings
as
well.
A
B
Thank
You
Jan
for
traduction,
so
my
name
is
Adisa
nursing
subtly,
so
he's
quite
communicated,
I'm
a
full-time
researcher
at
AAT
right
now
and
also
today,
I'm
come
from
the
I
got
a
fellowship
from
Isis
as
well.
So
thank
I
talked
to
two
key,
the
first
you
to
need
to
attend
this
meeting
so
today,
I
would
give
a
beef
talk
about
a
botnet.
B
All
right:
okay,
it's
okay!
Now,
ok!
So
today,
I'm
gonna
keep
the
big
talk
about
darkness,
which
is
the
continent
work
in
Thailand.
So
this
world
is
combined
many
work
from
the
many
people
from
the
AIT
team
in
the
lab
team.
So
this
is
one
of
the
person
who
work
within
this
project.
So
ok
start
to
start
with
what
is
dark
net.
B
Okay,
so
magnet
is
a
small
community
work
which
is
try
to
provide
internet
connect
to
connectivity
to
the
ahead
of
Thailand,
which
is
the
northern
part
of
Thailand,
is
called
the
alchemist
art
which
is
in
the
parks
Poland.
So
it
is
also
at
the
border
between
Thailand
Obamas
as
well.
So
we
start
this
project
from
the
research
project
in
AIT
and
since
thousand
thirteen.
So
and
currently
we
have
only
it's
about
five
years
and
we
have
led
point
almost
fifteen
Lemoore
authority
already
s
and
we
can
cover
almost
or
one
thousand
user.
B
We
keep
using
our
network
on
the
daily
basis,
alright.
So
at
the
beginning,
that
magnet
is
based
on
modern
TF.
So
we
collect
li
kui's
Dolan's
here,
which
mostly
the
student
from
the
under
mostly
undergrad
student,
to
have
a
cam
like
a
winter
camp
dealing
between
the
December
November,
because
this
is
the
cool
time
in
Thailand
at
that
time.
B
So
really
could
a
student
and
then
the
student,
like
a
cam
and
everywhere
the
student
went
to
to
the
to
the
network
right
to
deploy,
to
help
the
people
install
and
set
up
the
network
and
teach
them
how
to
use
as
well.
That
used
to
be
the
back
of
so
in
terms
of
technologies
is
similar
to
another
continent
work
like
Evie
or
Athens.
We
use
the
wireless
mesh
network,
we
see
based
on
the
Wi-Fi.
So
what
we
do?
We
can't
dimension
by
using
the
OSR
routing
protocols
and
we
share
the
Internet
to
the
Gateway.
B
We
shouldn't
be
a
DSL
or
fiber
instead
of
hardware.
We
try
to
make
it
simple.
We
try
to
use
the
hardware
available
in
the
market
like
the
tp-link
one
that
you
can
find
anywhere.
So
at
our
first
photo
tie.
We
use
the
T
peeling
m320,
the
listen
to
you
use
that
one,
because
it's
quite
is
supportable
and
we
had
the
battery
inside
and
also
we
used
our
on
in
terms
of
when
the
is
a
happening
like.
But
what
happened
like
a
tsunami
earthquake?
B
The
relation
can
take
that
router
and
then
just
can
run
the
images
cinema
immediately
without
manipulation.
So
we
insert
of
15
when
the
neighbor
has
earthquake.
We
send
those
louder
to
help
themselves
to
create
the
industry,
situate
emergency
networks
as
well.
Also,
we
also
have
the
another
model.
We
should
be
bigger
that
provide
better
coverage
areas
and
for
the
outdoor
like
we
have.
We
want
to
provide
them
like
a
longer
link.
We
use
the
ubiquity
one,
which
is
I,
think
everyone
know
about,
and,
of
course
we
use
the
Raspberry
Pi.
B
This
is
cool,
it's
cheap
and
to
run
our
services
are
so
inside.
The
village
and
I
wouldn't
is
not
to
mention
that
we
also
use
sorts
kind
of
DIY
stuff,
like
a
walk
that
we
use
them
like
yeti-like
dao.
I
do
it
yourself
support
to
park
for
the
point-to-point
link
and
party
for
antenna
that
we
can
attend
idea.
They
wanted
a
coal
ash
alley
or
the
wireless
links
as
well.
I
couldn't
remember
exactly
how
many
perform
ended
we
can
improve,
but
I
think
is
about.
If
you
have
interested
in
that,
I
can
have
some.
B
We
need
some
measurements
for
that
a
lot.
How
can
improve
the
connectivity
by
using
that
work?
So
in
terms
of
services,
we
develop
many
services
as
well
like
we
used
like
open,
saw
to
like
partially
to
implement
the
authentic
Indian
system,
natural
for
network
monitoring,
and
also
we
did
develop
our
own
chat
application
called
ambush
shot,
so
the
damage
has
it
like.
We
are
connect
to
the
lie,
which
is
I,
think
the
law
application.
B
This
is
developed
by
the
Japanese
company,
but
was
mentioned
that
in
Thailand
we
don't
use
people
don't
use
whatsapp,
but
people
use
lie
as
the
identification.
So
in
this
application
the
relator
can
use
their
mobile
phone
connect
to
that
and
they
can
chat
with
the
local
locally,
but
also
they
can
connect
to
the
lie
user
as
you're
outside.
Apart
from
that,
we
also
develop
some
kind
of
the
video
Sri
means
as
well.
We
cache
the
high-definition
called
videos,
mostly
like
education
content,
to
the
to
the
louder
that
is
show
before
and
also
sealant
listen
Lee.
B
We
also
developed
kind
of
what
I
for
IOT
of
channel
ins,
which
is
like
modeling
the
hairs
as
well
to
measuring
the
pollution
that
happening
in
those
area.
So
it's
so
good
right,
so
putana
just
keep
working
for
some
time
and
then
it's
quite
okay.
But,
however,
once
we
launched
the
network
about
two
to
a
year
and
we
try
to
sit
and
think
carefully
if
we
further,
it
is
quite
difficult
to
make
a
road
network
scale
and
also
it's
quite
very
different
challenges
as
well
to
how
to
make
the
network
sustained
by
yourself.
B
So
that's
why
we've
had
to
think
and
what
to
some
kind
of
the
engineering
that
we
will
try
to
come
up
with
a
new
model
contact
so
in
this
model
is
not
based
on
volunteers
anymore,
it's
the
one
until
isn't
based,
but
we
still
involved
many
partner.
So
one
of
those
is
like
we
opened
a
company,
a
social
enterprise
company
whose
call
net
to
homes
which
is
gonna
be
like
that
got
the
lice
and
Michael
Mike
Hawaii
license
as
well.
So
the
net
to
home
will
help
us
in
terms
of
running
the
services.
B
Do
the
Billings
or
do
kind
of
managing
the
networks
as
well,
but
near
to
Houma,
so
get
some
support
from
the
Titian
Technic
foundation
into
the
equipment.
So
the
teach
any
foundation
do
a
lot
of
fundraising
to
get
a
support
for
their
to
help
or
to
be
the
network,
and
also
we
have
in
the
languages
I'm
working
for
we
see
kind
of
taking
care
of
ID
a
section
and
also
we
have
a
local
community
so
like
local
people.
So
normally
we
try
to
train
them,
how
to
a
family.
B
B
So
we
do
internet
team
because
of
their
over
the
new
service
and
for
that
try
to
help
network
more
sustainable
as
possible
and
at
the
same
time,
so
right
now,
and
also
the
internet
also
try
to
provide
technical
support
top
alluding
to
in
the
trainings
as
well
to
the
local
communities.
So
this
model
is
the
network.
Isn't
this
is
not
free,
but
no
code,
the
local
community
and
they
have
the
user
right.
They
had
to
pay
a
little
money
for
the
monthly
fee.
B
So
this
money
try
to
mostly
go
to
cover
like
to
get
internet
fee
that
had
to
go
to
Internet
and
also
some
cause
for
the
maintenance
as
well.
At
the
same
time,
the
net
to
home
use
try
to
utilize
that
money
to
make
the
network
more
sustained
by
first
of
all,
they
push
the
money
to
to
increase
the
development
to
to
intern
at
teams
and
also
they
fight.
B
They
also
supplied
equipment
for
in
terms
of
installations
as
well,
but
in
order
to
utilize
that
money
more
wisely
and
make
a
network
more
sustained,
the
we
try
to
push
the
Larry
nail
back
to
the
local
communities
as
well.
So
what
we
try
to
do,
since
we
have
a
lot
of
training
to
train
the
local
share,
the
common
champion,
Rider
called
local
technicians.
So
now
we
keep
them
the
installations
horse
as
well.
B
So
the
local
technician
can
have
some
kind
of
some
money
that
to
help
the
network
for
installation
so,
and
that
is
the
part
of
business
model
that
we
have
right
now,
so
the
model
is
working
quite
well,
so
I
may
have
looking
quite
a
phone
for
some
time,
but
in
terms
of
research
that
I'm
working
on
it's
or
the
idea.
So
what
happened
so
we
had.
This
is
some
kind
of
problem.
It
is
a
problem
that
we
found
like
we
did.
B
Some
arises
by
looking
to
the
traffic
measurement
and
also
some
social
interviews
in
15
I.
Think
IP
sent
this
guy
a
sort
of
15
in
London
as
well.
So
what
we
did
we
have.
We
want
to
look
into
some
kind
of
social
communication
what
people
are
using
in
the
village
and
we
found
people
using
Facebook,
which
is
quite
fitting
homo
and
also
used
la
application.
As
I
mentioned.
This
is
number
one
chat:
application
using
in
Thailand
before
using
Instagram
and
email.
So
what
we
fall.
B
We
found
that
80%
of
the
user
that
using
in
village
life,
so
they
have
the
contact
inside
the
village
and
look
into
the
the
traffic.
That
is,
if
the
message
exchanging
most
like
twenty
percent
of
twenty
percent
and
a
message
that
keep
it
local
but,
however,
look
into
a
detector
that
we
have
all
the
relays
have
at
their
own
gateway
and
they
have
to
go
to
the
internet.
But
we
don't
have
any
local
link
to
how
to
link
them
and
keep
the
traffic
local.
That's
the
problem
that
we
found.
B
B
So
the
first
thing
that
we
have
to
do
for
this
project,
so
we
try
to
do
a
lot
of
measurement
or,
in
some
spectrum,
try
to
build
the
database
for
the
nation
that
for
the
nationwide
database,
why?
So?
We
got
a
lot
support
from
ICT,
as
you
would
mentioned,
and
Micro
Center
knows,
and
another
mono
for
the
tool
for
the
spectrum
measurement,
and
this
is
some
example
that
we
did
be
the
measurement
in
in
the
park
net,
and
we
can
see
that
more
spectrum
are
quite
free
at
the
moment.
B
So
as
a
next
step,
we
want
to
do
to
demonstrate
a
benefit
or
to
MySpace
by
linking
the
village
by
using
the
T
Y
space
today
that
expresses
architecture
that
we
are
doing
right
now.
So
we
try
to
connect
the
four
villages
to
the
T
white
space
link
and
equipment.
We
gonna
use
is
going
to
be
the
cost
which,
based
on
ITV
802.11
AF,
which
this
is
quite
similar
to
the
Y
files
as
well.
B
Apart
from
that
is
not
only
their
the
device
based
on
the
hell
out
in
some
of
the
backhauling
try
to
connect
the
relationships
is
far
away,
but
looking
to
some
example
here
about
in
terms
of
coverage
earlier,
you
can
see
that
the
yellow
square
is
a
college
and
that
we
can
provide
by
the
mesh.
But
we
also
have
a
lot
of
requests
from
the
people
who
live
in
the
next
circle,
which
is
not
really
far
from
from
our
network.
B
But,
however,
we
try
many
solution
by
the
multi
hall,
communication
or
Wi-Fi,
even
though
the
work
with
the
loan
and
communication,
but
the
proven
is
very
poor.
So
how
to
solve
that
one,
since
we
already
have
a
license
from
from
the
liquidator
right
to
use
the
TV
ban.
So
we're
looking
to
another
solutions
as
well
by
working
with
Microsoft's
at
UK.
Try
to
use
it
out
a
small
sale
solution
which
is
also
running
over
the
T
wish
ban.
B
Why
connected
to
the
mobile
core,
which
is
the
VM?
So
we
have
to
run
now?
We
don't
need
any
further
equipment
for
the
for
the,
for
the
mobile
call
equipment
is
EPC,
so
everything
is
running
based
on
the
software
is
running
in
the
cloud,
but
and
for
the
cloud
that
we
use
is
Microsoft
card,
there's
a
contest,
one.
It's
been
in
Singapore
and
Hong
Kong.
So
this
is
a
concept,
one
that
we
can
file,
and
this
is
some
performant
benchmarking
that
we
have
to
be
in
camp.
B
They
did
in
canvas,
so
so
they
have
the
particular
bandwidth
in
service
25,
75,
25
in
terms
of
download
and
and
upload
the
active
user
going
to
have
16
users
at
the
same
time,
and
the
college
is
my
criticals
as
well.
We
have
a
four
kilometres
diameters
and
also
the
good
thing
that
had
to
mention
is
like
the
user
who
use
a
modified.
It
says
they
can
keep
that
equipment
inside
the
house,
but
they
still
got
the
signal.
B
B
Okay,
okay,
just
to
have
some
more
I.
This
is
a
key
texture.
I
think
I,
don't
have
much
time
to
code
to
go
for
that,
but
I
always
publish
this
one
to
the
interview
as
a
poster,
so
you
have
want
to
go
for
detail
just
the
paper.
I
think
it's
going
to
be
all
about
soon
and
finally,
I
would
like
to
some
kind
of
key
takeaway.
What
being
the
time
I
become,
the
successful
I
think
the
keys.
The
key
thing
is
going
to
be.
B
The
stock
operation
be
partner
that
I
already
mentioned,
I'd
be
team
from
the
net
to
home
companies
and
the
form
the
low
engagement
with
local
communities
and
also
are
the
key
takeaways.
Another
key
takeaway
is
like
the
take.
It
should
be
simple
enough
that
the
local
people
can
know
and
understand
what
is
going
on
and
they
can
use
and
they
can
typically
produce
and
also
we
also
need
to
engage
to
the
local
actors.
B
We
can
not
only
giving
them
a
technology,
but
we
have
to
push
them
some
kind
of
benefit
like
lever
news
or
as
we
have
here,
you
can
see
that
the
guy
over
there
his
name
is
cool
some
porn,
so
he
trained
them
at
the
beginning,
but
now
he
he
were
with
us
and
then
he
can
keep
that
local
this
job
as
a
second
job.
For
him
too,
for
living
and
last
mentioned
some
time
is
just
going
since
we
run
the
first
motor
of
four
thousand
thirteen,
and
so
that's
two
and
two-thirds
up
2016
right.
B
B
I
would
like
to
thank
to
our
team
and
Professor
cancer
now
who's,
one
of
pioneers
and
he's
running
the
laps,
the
spirit
of
the
mon
coeur
north
opinion-
and
this
is
all
the
team
that
how
is
including
me-
and
I
would
like
to
thank
you
to
our
sponsor
technique,
foundation
and
BTC.
We
said
hi
like
you
later,
I
CTP
came
with
where
I
used
to
work
before
a
micro
service
educates
as
well
that
give
us
some
funding
in
terms
of
equipment
for
LTE
and
finally,
I
would
like
to
die
I
soft
as
well.
B
To
give
me
the
friendship
to
come
here
that
that
this
is
were
too
much.
So
even
is
very
good
for
me
to
learn
about
what
IETF
and
then
keep
the
work.
So
last
but
not
least,
you
know
that
the
native
in
Bangkok
right,
so
we
also
have
an
icing
and
at
the
conference
as
well,
we
call
a
cement,
a
responsive,
I,
say
calm,
so
the
deadline
for
some
reason
in
September
in
August,
but
just
don't,
go
the
conference
just
be
right.
Next
to
the
idea,
just
ITF
economies
and
Friday
right
so
and
on
the
Monday.
C
B
A
B
A
Just
a
side
note
on
the
project
that
other
Soren
mentioned
with
the
TV
white
space
use.
We've
had
a
very
interesting
time
when
we've
been
working
in
some
organizations
to
talk
about
the
importance
of
the
use
of
TV
white
space.
In
fact,
we've
had
some
massive
opposition,
and
so
what
you
see
is
active
use
of
TV
white
space
and
working
with
the
regulator.
A
D
Afternoon,
everyone
Oh
quiet
Phelan,
so,
as
Janus
said,
I
am
I,
am
I'm
Harper
I'm,
the
director
of
next
generation
leaders
at
the
Internet
Society
I'm
responsible
for
fellowships
so
congrats
to
Paris,
Lauren,
I'm,
also
responsible
for
capacity
building
online
training.
Some
face
to
face
as
well
as
blended,
so
I
would
have
presented
on
this
a
project
a
few
years
back
when
it
was
in
the
conception.
State
blueberry
was
just
a
fledgling
project,
but
basically
a
bit
of
background.
D
So
we
started
in
2015
a
partnership
between
ie
and
Tripoli
and
Internet
Society,
really
to
to
scale
or
or
Wireless
fork,
fork
fork,
communities
or
project
which
was
based
in
India,
I.
Think
where
we
had
done
a
lot
of
work
with
face
to
face
the
training,
but
it
did
not
scale
well,
so
we
decided
we
would
develop
this
training,
so
we
have
a
train-the-trainer.
D
It
was
also
about
creating
communities
of
practice
in
terms
of
community
networks
really
of
fostering
the
build
of
no
not
for
a
for-profit
wireless
in
the
networks,
and
it's
part
of
the
internet
society
is
2000
and
gay
teen
campaigns.
We
really
focused
on
building
out
community
networks
and
increasing
our
access.
D
So
just
a
bit
about
the
development
process.
We
we
engaged
a
content
development
partner
to
to
help
us
to
build
in
interactive
and
training
way
of
learning,
pedagogy
effective
in
the
instructional
design,
etc.
We
we
work
very
closely
with
the
I
Triple
E
standards
group
to
review
and
refine
the
contents,
so
that
we
can
we
could
develop
something
that
was
fit
for
purpose.
D
We
did
a
pilot
into
2016,
really
validate
or
a
learning
objective
and
to
really
see
how
the
community
responded
and
that
was
extremely
successful
and
then
to
increase
our
scale.
We
did
a
translation
into
Spanish
and
we've
also
made
sure
that
we
had
in
the
Bedok
many
processes
so
that
we
can
ensure
that
the
course
continues
to
meet
the
needs
of
that
the
communities
as
well
as
to
help
us
continuously
and
improve
it.
D
So
the
course
content
is
pretty
the
diversity
looked
at
our
radio
physics,
let
that
sight
sight
that
I'm
planning
in
terms
of
power
looking
at
back
on
Hall
ground
in
erecting
different
types
of
knives
and
Canada,
etc.
We
do
a
lot
of
work
on
networking
as
well
IP
addressing
routing
excetera
security
of
wireless
networks,
access
point:
we've
been
focusing
a
lot
on
more
of
the
low-cost
and
open
source
and
platforms.
D
So
we've
done
some
work
on
a
libre
router,
as
well
as
micro,
to
take
devices
as
well
as
some
some
troubles
student
of
war
or
less
and
then
that
networks
in
the
initial
stages
we've
had
a
few
partnerships
where
we
thought
we
can
make
the
most
impact.
So
we've
worked
with
after
Chicks,
where
we're
really
focusing
on
building
capacity
and
empowering
women
in
tech
across
the
African
region
done
some
work
with
the
free
and
open
source
software
foundation
of
Africa
as
well.
D
The
course
is
primarily
for
the
capacity-building,
but
we've
also
seen
a
need
for
the
course
around
workforce
readiness
really
to
to
help
people
to
develop
skills
that
are
transferable
to
the
workforce.
We've
done
some
work
with
a
deaf
where
we've,
because
that
has
a
number
of
love
technicians
who
volunteer
and
who
don't
really
have
a
strong
understanding
of
how
to
build
out
these
networks.
So
so
we've
used
the
training
to
prepare
them
before
they
deploy
to
the
field.
D
We
was
working
closely
with
nogs
in
both
African
and
Latin
America,
as
well
as
we've
been
doing
some
work
with
welcome
chapters
so
progress
to
date
for
2017,
with
really
our
first
full
year
of
delivering
the
coop
the
course.
So
we
train
twenty
trainers
who
can
then
go
back
to
their
regions.
Are
there
look?
There
are
no
cows
and
and
train
additional
groups.
We
delivered
our
online
course
to
approximately
150
persons.
D
We
had
a
lot
of
webcast
and
could
do
various
building
where
we
invited
practitioners
for
from
work
from
Latin,
America
and
Africa
to
speak,
to
participants
to
do
knowledge,
exchange
and
lessons
learn,
and
we
also
did
a
collaborative
leadership
exchange
on
community
network
that
NIC
that
not
for
those
who
are
not
so
familiar
with
work,
collaborative
leadership
exchange.
It's
pretty
much.
An
unconference
that
we've
delivered
at
a
number
of
venues
are
ugf.
D
D
In
2018
this
year,
we've
really
integrated
our
work
into
I
socks,
Community,
Network
campaigns,
moderator
training
is
currently
in
progress.
The
first
set
of
moderator
and
training,
retraining,
15
individuals
from
13
countries
in
q1
we
trained
a
hundred
and
five
individuals
from
21
plus
countries
and
in
q2
we
trained
at
128
kinda
vigils
from
25
plus
countries,
and
we
will
be
opening
up
a
powerful
applications,
probably
in
the
next
week
or
two
for
coutry,
for
Q
tree
done
training
and
that's
in
English
and
Spanish.
D
D
We
can
allow
communities
to
self
manage
the
delivery
of
the
course
and,
from
that
perspective,
we're
looking
at
we've
designed
the
course
to
be
portable,
so
it's
standards-based,
so
it
can
be
surfaced
on
any
learning
management
system,
we're
developing
and
almost
finished
an
offline
version
that
can
be
run
from
either
an
and
XE
XE
cutable
on
window,
though
it
can
be
ran
from
a
USB
stick
so
that
the
training
can
be
dealing.
There
were
in
communities
that
have
no
internet.
D
We
were
working
on
some
some
strategic
partnerships
in
Asia,
Pacific,
actually
trained
thousands
or
hundreds
of
thousands
of
individuals
and
that's
making
very
good
progress
with
the
expected
work
to
run
a
pilot
by
the
end
of
the
year.
We're
also
looking
at
translation
of
the
course
into
other
languages,
so
French
and
Arabic,
and
we
want
to
make
some
substantive
of
this
to
the
content
in
2019.
It's
substantive
from
the
perspective.
D
The
training
is
12
12
modules,
and
you
usually
takes
about
six
weeks.
They
give
people
six
takes
a
week.
We
can
do
a
fast,
fast,
rap
version
where
we
can
bring
people
to
a
classroom
as
well
and-
and
that
would
take
about
four
will
be
about
two
to
two
weeks.
E
F
Interested
in
learning
more
about
the
training
of
trainers
do
t-piece.
How
did
you
select
the
trainers?
How
did
you
do
outreach
for
the
people
to
be
trained,
at
least
at
the
very
first
round
and
after
the
tea
OTPs?
Was
there
also
some
sort
of
mentorship
that
was
going
on
to
accompany
the
new
trainees
into
the
knowledge
they
acquired?
Okay,.
D
So
in
terms
of
how
we
train,
we
would
have
worked
with
some
of
our
partners
to
promote
the
training
we
promoted
fire
I
thought
with
the
different
channels
like
social
media.
We
also
work
with
our
different
partners
in
the
region,
so
the
nogs
and
their
research
and
education
in
the
network.
So
we
get
a
very
diverse
group
of
participants.
D
Terms
of
the
mentorship
piece
part
of
the
trainer
training
course
includes
how
to
effectively
deliver
online
done
training,
that's
actually
an
additional
module
that
we
got
so
teach
people
how
to
build,
how
to
build
online
learning
communities,
how
to
develop
social
learning
or
the
key
people
engage
how
to
manage
people
who
are
not
confident
and
and
that's
part
of
helping
them
to
effectively
deliver
training.
There
was
a
third
third
question:
how
did.
D
The
first
trainers
came
from
or
different
strategic
partners,
so
across
various
region,
the
first
trainees.
It
was
a
combination.
We
we
had
a.
We
had
a
cook,
a
call
for
applications
and
people.
We
had
probably
had
400
something
applications
for
about
what
I
would
say
in
the
90
spots,
and
then
we
also
would
have
had
chapters
and
knog
who
who
were
interested
in
delivering
training
to
their
members.
D
F
G
D
G
B
D
D
India
and
the
part
of
that
partnership
will
be
translated
into
local
languages,
so
probably
Purdue
or
some
other
language
or
or
Hindi,
and
we're
also
looking
right
now,
actually
I
translated
it
into
Arabic
for
2018.
A
If
anybody
is
interested
in
us
translating
it
into
another
language,
maybe
Thai
we'd
be
happy
to
talk
to
you
about
that.
Talk
to
Niall
Leandro
from
the
UPC
University
in
Barcelona
is
a
partner
in
crime
and
also
someone
who's
part
of
grief
II,
which
is
one
of
the
biggest
community
networks
in
the
world.
H
H
So
we
came
up
with
a
mechanism
which,
which
has
to
do
with
a
kind
of
mandatory
infrastructure
sharing
and-
and
this
has
been
developed
in
terms
of
of
an
ordinance.
It
is
a
kind
of
local
regulation
by
the
municipalities
in
Catalonia.
We
have
about
700
and
then
many
of
them
are
really
tiny
and
they
need
help
to
come
up
with
a
ordinance
which
which
complies
with
the
Catalan
Spanish
European
legislation,
which
is
not
really
easy
task.
So
well,
when
we
talk
about
universal
connectivity.
H
In
the
in
the
least
developed
areas
in
the
ones
that
there
is
a
negative
profit,
let's
say
and
and
well
I
mean
there
are
there's
a
hard
time
in
many
parts
of
the
world
to
achieve
that
and
and
the
result
is
but
typically
the
European
Commission
call
it
the
market
failure
and
they
never
know
if
it's
dark
black
or
white
zones
without
infrastructure,
and
then
there
is
need,
obviously
for
for
not
technical
mechanisms
which
are
the
technology.
Is
there
and
is
sometimes
very
an
expensive,
quite
an
expensive,
but
also
policy
of
regulation
mechanisms?
H
There
is
a
restaurant
model
that
we
know
it's
regulated
predictable,
but
it's
pensive,
sometimes
or
the
other
model,
which
is
self
provision
Wi-Fi
in
a
ways.
Many
times
is
used
for
sale
provision
and
then
particularly,
the
public
authorities
have
the
role
of
arbitrating,
the
use
of
the
public
space,
which
is
required
to
deploy
infrastructures.
Less
obviously
I
mean
towers
in
the
case
of
wireless,
but
that's
in
the
case
of
of
municipal
land
or
even
like
undersea
infrastructures
in
the
case
of
long-distance
connections.
H
So
the
idea
is
to
to
come
up
with
a
mechanism
where
that
belongs
to
everyone
creates
a
return
to
everyone.
While
there
is
some
private
interest
in
in
using
that
space.
So
it's
about
this
idea
of
open
access
bands
for
fiber,
let's
say
so.
You
know
this
is
like
a
diagram
that
shows
the
huge
amount
of
infrastructure
that
is
already
deployed
and
it
doesn't
cover
everything.
It's
just
one
of
the
examples
of
you
see
the
the
brown
color.
It
relates
to
the
density
of
population,
so
the
less
populated
areas
as
less
infrastructure.
H
So
basically,
the
idea
is
when,
when
a
company
comes
to
the
municipality,
these
mechanism
promotes
the
the
deployment
of
the
a
of
the
those
privative
cables,
but
the
d1
the
promoter
becomes
responsible
for
the
course
of
management,
and
maintenance
has
to
do
it
anyway,
because
it's
is
going
to
get
that
cable
and
then
the
City
Council
is
exempt
from
the
maintenance
cost
of
of
the
remaining
fibers.
That
say
so.
One
charge
takes
the
cost
of
one
part
that
is
going
to
be
used
for
private
use.
H
The
municipality
gets,
let's
have
free
access
to
the
fibers
in
exchange
of
allowing
this
plan
at
the
private
deployment,
and
then
the
there
is
a
part
of
the
fiber.
Let's
say
which
is
under
the
model
of
a
commons
and
then
the
cost
of
the
usage
or
the
maintenance
of
that
part
of
the
of
the
of
the
cable
is
shared
proportionally
to
the
to
the
users.
It's
according
to
a
cost
oriented
model
and,
and
then
there
is
the
they
have
to
create
an
entity
like
which
which
makes
sure
that
this
cost
sharing
is
working.
H
H
Then
there
is
the
City
Council
use
for
whatever
they
need
and
third,
it's
a
the
deployment
creates
some
fibers
which
are
open
for
the
use
of
anyone
else,
any
other
operator,
any
other
company
in
the
in
the
village,
any
other
school
whatever
they
want
to
they're
interested
in
using
that
infrastructure
that
can
be
implemented
differently
depending
on
what
is
the
type
of
cable.
Where
is
the
capacity
the
number
of
how
its
organized
so
basic?
H
But
you
can
see
in
the
diagram
that
usually
every
like
a
tube
is
divided
into
different
fibers,
which
are
some
of
them
are
for
for
private
use.
Some
of
them
are
shot
shares.
Some
of
them
are
used
by
the
municipality.
In
fact
they
talk
about.
They
call
it
minimal
structural
units,
which
is
how
this
is
implemented,
is
sharing.
This
mandatory
sharing
is
implemented
depending
on
the
characteristics
of
cables,
so
I
think
you
can
see
here
is
well.
H
This
is
a
bit
of
detail,
but,
and
the
interesting
part
is
that
this
model
has
been
developed
for
a
few
years,
and
now
today
we
are
in
version.
28
of
the
document
is
quite
long
if
you're
interested
I
can
share
with
you
question
14
that
that's
the
one
that
we
translated
to
English
the
person
20
wait
to
Nate's
in
Catalan.
H
If
you
want
to
learn
a
new
language,
so
I
say
I
mean
it's
about
what
we
are
doing
there
it's
about
how
to
promote
the
expansion
of
an
infrastructure
across
public
land
in
a
way
that
benefits
the
whole
community.
In
addition
to
the
promoter,
this
is
the
return
of
minimal
cost
fiber
for
everyone
else.
So
a
public
space
keeps
being
a
public
space
it
can.
We
have
applied
that
into
a
small
municipalities
in
Catalonia
for
for
land
infrastructure,
how
about
extending
it
to,
let's
say
national
and
international
cables,
which
are
over
land
or
under
seas?.
H
H
So
that's
that's
my
presentation.
If
you
have
any
questions,
I
will
be
happy
and
by
the
way
the
objective
is
to
come
up
with
a
document
that
we
can
maybe
that
contribute
and
discuss
in
the
in
the
working
group
I
think
many
locations,
many
places
where
people
could
benefit
from
that.
Thank
you
very
much.
H
I'm
potato
I'm
particularly
loose
if
there
is
any
similar
regulation
in
another
region
that,
because
we
in
the
you,
for
instance,
there
is
a
directive
course
called
cost-sharing
which
promotes
that
the
sharing
of
deployment
of
infrastructures,
but
what
we
do
is
is
making
it
more.
Let's
say,
kind
of
mandatory.
A
Just
an
observation:
in
the
village
we
were
in
in
northern
Greece
below
Mount
Olympus,
you
could
see
the
fibroid
trenches
in
the
road
coming
right
up
into
the
village.
The
fiber
was
dark.
You
all
know
how
how
much
capacity
there
is
on
one
little
strand
of
fiber
dark.
So
that's
why
some
of
the
projects
the
community
networks
are
coming
around
because
they
can't
afford
some
of
that
municipal
fiber.
So
this
this
potential
question
that
Leandre
is
asking
about.
Can
we
get
some
shared
use
public
use?
A
There
is
a
market
failure
here
that
we're
seeing,
obviously
right
if
regulators
are
supposed
to
help
encourage
connectivity
by
enforcing
universal
service
enforcing
the
different
rules,
but
people
don't
have
affordable
connectivity
and/or
connectivity
at
all.
Then
we've
got
a
conundrum
right.
It's
put
in
if-then
kind
of
issue,
but
Spencer's
up
he's
from
Washington
University
of
Washington
yeah
so
Spencer
over
to
you.
I
Hi
so,
like
Jenny
mentioned,
my
name
is
Spencer
I'm,
a
postdoc
at
the
University
of
Washington,
where
I
work
with
Curtis
hi
Merle,
who
many
of
you
maybe
know-
and
this
talk
is
about
our
upcoming
community
LTE
project,
which
were
calling
culty
so
I
started
with
some
background
on
community
networking.
I'm
actually
gonna
skip
it,
because
we've
just
heard
about
several
that
I
think
illustrate
a
lot
of
the
points
but
I'm
gonna
talk
to
you.
I
If
you
want
a
little
bit
about
community
cell
networks,
which
is
that
is
a
bit
less
common,
our
labs
background,
we've
started
with
a
lot
of
work
on
smaller
2g
networks
that
are
doing
just
voice
and
text
the
typically
smaller
scale
than
some
of
the
flagship
community
networks
out
there
we're
talking
hundreds
of
users,
one
to
three
towers.
The
network
architecture
of
cellular
is
obviously
a
bit
more
centralized,
which
is
good
and
bad.
It's
nice
to
kind
of
have
a
single
point
of
failure
that
we'll
set
up
we'll
get
a
tower
going
somewhere.
I
These
paths,
community
networks
typically
have
low
backhaul
requirements,
but
they
still
the
biggest
things
that
they're
still
a
community
network.
They're
still
community
owned
and
managed
I'm
mission,
dour
pass
workers
and
2g
networks,
and
so
we're
doing
voice
and
text
and
just
using
Twilio.
Typically,
some
networks
use
Skype,
but
ours
were
using
Twilio
for
a
PLM,
an
interconnect.
This
is
a
couple
of
problems.
We've
had
some
problems,
keeping
Ozma
calm,
keeping
the
build
and
development
system
kind
of
sustainable.
I
Why
not
LTE
there's
a
lot
of
really
good
stories
here
that
have
been
evolving
over
the
last
couple
of
years
when
compared
to
a
2g
network
architecture,
we
found
that
an
LTE
network
is
kind
of
a
lot
simpler,
there's
commodity
Hardware
is
coming
down
in
price
compared
to
some
of
our
previous
experience
with
2g
base
stations,
we're
going
from
8k
down
to
two.
At
the
same
time,
over
the
last
couple
of
years,
handset
support
has
been
increasing
worldwide.
I
What
sets
it
apart
from
prior
generations
of
cellular
networking
is
that
it's
all
an
IP
substrate.
This
is
a
huge
benefit.
This
is
my
background,
is
more
just
straight-up,
IP,
IP,
switching
and
routing
there's,
probably
a
lot
of
other
people
in
this
room
that
share
that
background,
and
it's
really
nice
to
come
in
and
great.
I
We
can
set
this
up
and
some
of
the
other
network
functions
like
voice
and
text
so
on
and
so
forth
are
actually
still
just
provided
over
an
IP
substrate,
which
kind
of
means
that
they
just
they're
still
just
run
over
IP
you
can.
You
can
support
a
network
with
or
without
these
functions.
It's
the
simpler
version
is
actually
just
to
build
a
data,
only
network,
which
is
what
we
did.
So
we've
got
kind
of
a
simple
diagram
here
that
I
think
was
supposed
to
be
an
animation,
but
not
that's,
not
an
issue.
I
You've
got
kind
of
you've
got
the
core.
You've
got
the
tower,
which
is
an
e
node
B,
the
handsets
off.
It
just
has
a
tunnel
to
the
core,
and
that's
it
you
have
IP
addresses
and
then
that's,
obviously
what
it's
using
the
core
is
doing
that
to
get
out
to
the
internet
and
load
web
pages,
but
to
add
voice
and
text.
You
just
throw
another
server
on
there,
give
it
an
IP
address
and
use
DHCP
to
tell
the
handset
oh
yeah.
I
This
is
who
you
should
talk
to
so
when
we
take
out
the
voice
and
text.
What
we're
left
with
is
an
all
data
network
that
we
started
looking
at
LTE
as
just
an
access
technology,
it
kind
of
you
can
look
at
it.
The
same
way
we're
talking
about
Wi-Fi
and
it
actually
is
a
really
great
tool
in
this
space
for
rural
access,
which
is
the
focus
of
our
lab,
we're
talking,
high-speed
Internet
access
can
I
help
you
yeah.
J
I
J
The
UE
supports
LT
issue,
but
it
also
supports
a
number
of
3gpp
interfaces.
Today.
Anybody
who
builds
LT
UE
conforms
to
the
3gpp
specs,
but
means
it
supports
our
mass
signaling
that
nasa
signaling
goes
to
you
know
to
be,
you
know,
to
be
to
Mme
and
there's
a
you
know,
degree
to
s,
gateway,
P
gateway.
What
I'm
missing
is
when
you
allocate
an
IP
address.
All
of
these
functions
are
in
the
picture
right,
but
here,
when
you
say,
I'm
allocating
an
IP
or
you
assuming
those
functions
are
existing.
It's
a
really
new
signalling.
That's
no.
I
I
A
J
I
Cool,
thank
you
so
talking
about
when
we
started
looking
at
it
as
an
internet
access
technology,
we've
got
good
internet
access.
We've
got
long
range
of
coverage
way
more
than
a
typical
Wi-Fi
access
point.
Something
that
I
think
also
makes
deployments
kind
of
easier.
Is
that
by
reducing
the
number
of
towers
and
subsequent
the
number
of
boxes
involved
were
easiest.
We're
easy
routing
we're
getting
away
from
some
of
the
challenges
that
do
pop
up
in
like
mesh
situations
and
we're
also
reducing
the
number
of
failure
points.
I
Obviously,
we
can't
get
there
because
you
do
have
to
figure
out
issues
like
bands
antenna
configuration
essentially
buying
buying
SIM
cards,
but
all
these
tasks
when
we
can't
just
automate
the
process
we've
been
creating
step
guides
documentation
go
here.
These
are
the
important
parts
that
you
need
to
look
into.
The
goal
is
trying
to
empower
people
that
maybe
aren't
incredibly
experts
in
this
field
to
set
up
and
do
LTE
on
their
own
in
the
small
scale.
I
Here's
our
network's
basic
architecture,
a
lot
of
these
rural
places
that
are
target
deployment
right
now
has
a
pretty
constrained
internet
back
home.
Let's
just
run
over
a
satellite
link,
and
so
we
had
to
keep
everything
essentially
downstream
of
that
there's
a
lot
of
networks
out
there
that
say
buy
our
tower.
I
Our
you
know
to
be
plug
it
in
and
use
our
cloud
ep,
see
or
use
our
centralized
solution,
and
that
really
we
found
wasn't
gonna
fly
in
a
situation,
but
we've
got
an
intermittent
backhaul
that
at
most
is
one
megabit
per
second,
and
so
this
brought
us
to
this.
Brings
us
to
our
hardware
stack.
We've
got
a
phone
provides
kind
of
a
nice
for
scale.
There,
we've
got
regular
LTE
supported
handsets
the
Ino.
B
is
a
piece
of
essentially
commodity
commercial
hardware.
I
This
is
bought
by
buy
cells,
but
you
can
there's
a
number
of
you
know:
B
manufacturers
that
essentially
sell
these
and
then
the
EPC
is
the
open-air
interface
project.
It's
all
exists
essentially
in
software,
and
we
run
it
just
fine
on
this
relatively
small
one
hundred
and
fifty
dollars
so
tag
C
box.
I
Kind
of
riffing
on
that
a
minute,
there's
an
interesting
dynamic
that
we've
found
much
more
to
be
the
case
in
l2,
even
2g
in
that
there's
a
standard
interface,
XS
6,
a
4e
mobis
to
connect
to
core
core
networks,
and
in
the
past
this
has
not
been
as
respected
as
it
is
in
LTE.
So
we've
been
able
to
find
tremendous
plug-and-play
success
in
that
we've
tried
a
handful
of
different
commercial.
You
know
B's,
they
all
just
plug
right
in
work
with
the
core.
I
We
started
with
the
open
air
interface
project,
a
lot
of
the
work
that
we
put
in
initially
was
kind
of
fixing
some
build
the
script
errors,
getting
it
up
running
reliably,
starting
at
restarting
it
when
it
crashes
fixed.
A
couple
of
seg
faults,
standard
open
source
work,
but
our
goal
is
to
kind
of
wrap
wrap
these
up
into
a
nice
package
that
we
can
release
soon,
and
we
currently
have
some
I
wish
I
had
pictures
of
this
yet,
but
it's
not
yet
built
we
have
our
summer.
I
Interns
are
really
working
right
now
on
this,
the
first
draft
of
a
web
configuration
kind
of
a
status
tool
to
tell
us
what's
running
right
now:
what's
the
status
of
it?
How
can
we
do
this
right?
Now?
It's
a
bit
shell
script
T,
but
the
goal
is
to
move
past
that
and
kind
of
start
moving
towards
a
product
that
is
a
bit
less
cryptic
and
a
bit
more
usable
for
for
people
in
this
space.
I
That's
overly
complicated.
The
margins
are
better
for
billing
in
that
context,
for
people
running
commercial
networks
in
this
way,
but
our
goal
is
community
owned
and
run
stuff,
so
we're
just
billing
ourselves,
so
we
reduced
it
just
IP
billing,
which
actually
makes
life
really
simple.
There's
a
lot
of
great
tools
that
monitor-
and
you
know
in
log
traffic
like
this
and
leads
us
to
our
big.
Our
second
big
decision,
which
is
when
we
started
these
discussions
with
this
community.
The
big
theme
that
we
kept
hearing
was
oh,
but
can
you
support?
I
Whatsapp
and
I
was
like
well,
yes,
we
have
internet
access
and
we
were
trying
to
figure
out
how
to
set
up
voice
and
text.
Anyone
said
well,
you
know
yeah,
that's
nice,
but
really
we
want.
What's
that
so
eventually
said
well.
Actually
this
makes
our
life
easier,
so
we
don't
actually
provide
voice
or
text
at
all
right
now,
in
this
initial
deployment,
it's
all
Skype
whatsapp
other
over-the-top
services
that
we
found
the
community
wanted
and
preferred
to
use
anyways.
I
So
we
wrote
a
web-based
portal,
it's
it's
got.
You
know
an
engineer's
design
touch
being
myself,
but
it
works.
You
know,
we've
got
phone
number
information.
Current
balance,
you
can
buy
data
packages,
you
can
send
data,
you
can
send
data
packages,
you
can
top-up
other
accounts
or
send
money
from
one
user
to
another,
which
is
actually
a
huge
service
in
these
communities.
I
So
we
got
this
billing
package.
We've
got
this
network
that
essentially
runs,
and
the
last
thing
that
we
put
on
top
was
local
services,
and
by
that
I
mentioned
that
our
target
deployment
is
very
backhaul
constrained.
It's
a
one
megabyte
second
satellite
link
when
it's
up,
and
so
the
question
was:
how
do
we
provide
network
use?
Even
in
these
situations,
he
went
even
when
we
have
a
best
saturated
at
worst
offline
backhaul
link,
so
we
started
locally
hosting
some
web
services.
Wikipedia
is
a
classic
one.
Media
server
is
great
because
we're
also
asked
well
hi.
I
Can
we
stream
our
own
content?
Well,
yeah?
Of
course
we
can
do
that,
so
we
built
those.
It's
really
nice
for
just
running
him
on
the
EPC
and
again,
because
we
have
this
kind
of
simple
network
architecture
of
IP
connectivity,
it's
as
easy
as
just
running
a
DNS
caching,
server
post,
pointing
everyone
to
the
service
name,
dot,
beau
Contini,
which
is
the
community
that
we're
working
in
and
the
last
thing
that
we're
kind
of
iterating
on
right.
I
Now
that
is
working
but
we're
still
kind
of
building
out
is
OpenStreetMaps
also
offers
an
offline
kind
of
you
know
run
this
on
vagrant,
it's
a
mapping
service
in
a
box,
and
we
said
well.
This
is
actually
something
also
really
useful
and
one
of
the
main
reasons
people
are
one
of
main
things.
People
are
doing
with
the
Internet.
We
can
also
locally
host
this
so
I
mentioned.
We've
got
our
first
upcoming
deployment
is
in
Bochum
Deenie
that
previously
had
a
2g
community
cell
network.
I
I
meant
already
mentioned
the
backhaul
and
that
we're
doing
data
only
and
I'm
literate
heading
straight
out
from
IDF
I'm
gonna
have
a
stopover
in
Seattle
to
pick
up
all
this
equipment
and
then
we're
going
straight
out.
We're
gonna
be
there
until
around
mid-august,
at
which
point
we
should
have
a
lot
more
concrete
information
as
to
kind
of
how
this
went
and
what
that
narrative
looks
like
once
we
get
back,
we
don't
slow
down.
I
But
the
flipside
is
that
there's
a
nearby
fiber
patch
that
we
apparently
can
get
into
somehow
so
I'm,
not
quite
sure
what
these
details
are,
but
those
are
the
second,
those
are
numbers,
2,
&,
3
and
then
beyond
that
we're
always
looking
for
more
people
to
kind
of
talk
with
and
communities
that
we
can
support
I'm,
gonna
kind
of
try
to
move
quickly
through
some
of
these
last
lines.
There's
a
couple
of
three
more
directions
that
our
lab
is
kind
of
trying
to
push
things.
The
first
is
community
based
repair
and
maintenance.
I
In
supporting
this,
as
like
a
community
structure,
the
second
is
in
network
services
and
billing,
and
then
the
third
is
just
bigger
questions
about
internet
architecture,
so
the
first
idea
of
community
repair
we
have
previously
published
work
on
about
this.
Where
can
we
crowdsource
via
text
messages?
Community
incentivizing?
Can
we
get
simple
repairs
going
on
one
of
the
main
reasons,
these
community
networks,
when
you
see
hardware
failures,
nine
times
out
of
ten
there's
kind
of
a
lot
of
low
hanging
fruit,
solar
panels
are
dirty
I'm
antenna
got
knocked
over
in
a
storm.
I
There's
really
great
work
that
our
grad
student
Esther
did
on.
Can
we
say
about
texts,
get
people
to
respond,
encourage
them
kind
of
wraps
into
some
of
those
previous
discussions
about
training
and
just
there's
a
lot
of
stuff
where
people
were
scared
to
get
their
hands
on
stuff,
but
that's
not
necessarily
has
to
be
the
case,
and
this
builds
to
some
of
our
future
work,
which
is
sensitive,
related
or
esther
again
is
building
out
these
kind
of
boxes.
These
little
thoughts
here
on
the
left
are
lights.
I
Second
thrust:
we've
been
talking
about
I
already
mentioned
local
services,
but
they
intersect
really
closely
with
billing,
because
there's
questions
about
doing
free
or
discounted
data
rate
for
some
of
these
local
front
hall,
that
isn't
saturated.
There's
another
question
about
here
about
the
social
benefit
of
calling
and
texting
via
whatsapp
or
whatever
we've
been
talking
about
offering
this
straight
up
free
to
the
community,
because
there's
a
tremendous
social
benefit
and
these
calls
are
so
cheap
that
they're
not
going
to
saturate
our
thing.
I
Obviously,
there's
butts
up
against
net
neutrality
concerns,
there's
an
a
huge
debate
here
that
it's
I
think
is
gonna
start
growing
a
bit
going
forward
and
the
third
ideas
we've
been
talking
about,
trying
to
support
ways
to
do
web
content.
Caching,
everyone
always
wants
to
do
this.
It's
tricky,
which
leads
us
into
some
to
my
final
point,
which
is
talking
about
some
of
the
in
and
architecture
pushes
that
are
going
on.
There's
a
huge
question
of:
are
we
a
telecom
or
an
ISP
and
who
cares?
I
What's
the
difference,
a
lot
when
we're
talking
about
title
2
or
other
types
of
regulations,
emergency
services,
regulation
so
on
and
so
forth
architecture?
What's
the
difference
between
doing
volte
to
Verizon
or
using
just
Skype
whatsapp?
Well,
it's
actually
a
big
difference.
If
we're
trying
to
do
two-factor,
authentication,
I
can't
tell
someone
just
just
text.
My
Skype
address,
but
really
when
everything's
over
IP
anyways
these
technology
underlying
technologies
are
really
similar.
I
So
there's
a
whole
bunch
of
weird
edge
points
when
we
started
kind
of
debating
and
talking
about
this,
that
kind
of
lead
into
what
I
think
is
gonna,
be
a
future
research
agenda
and
some
of
these
ideas
that
we're
trying
to
push
on
thanks
a
bunch
I
just
covered
a
lot
of
stuff
we've
got.
This
is
my
email.
If
anyone
has
other
questions
offline-
or
this
is
a
blog-
we've
been
keeping
on
the
project
and
will
continue
to
update
while
we're
abroad.
K
Well,
my
first
reaction
is
well
really
great
job.
Awesome
I
think
you
made
some
really
good
trade-offs
and
the
simplicity
is
just
beautiful.
I
have
one
question:
did
you
have
any
struggles
with
the
users
moving
from
Wi-Fi
to
LTE
and
did
you
have
restrictions
on
the
cell
phone?
It
was
either
one
or
the
other
they
had
manually
had
to
switch.
Did
you
care
about
making
it
transparent
or
having
such
and
survivability
across
moves,
or
anything
like
that?
K
I
Is
something
we
haven't
thought
about
at
all,
so
this
target
community?
What
we're
going
in
BO
convenient
doesn't
have
internet
access,
they
had
a
2g.
No
there,
the
majority
of
Internet
access
is
people
have
phones
and
they
use
them
when
they
get
into
one
minute,
which
is
about
a
six
hour
drive
away,
but.
L
I
profile
from
grace
of
Campinas
Brazil,
thanks
for
the
work
I
mean
I,
was
looking
for
this
long
time
ago,
and
I
was
trying
to
do
exactly
this
these
next
semester.
So
it's
good
to
see
that
someone
else
is
doing
the
same.
So
I'd
like
to
share
some
knowledge
with
you,
and
my
question
is
about
you
know
about
any
commodity
hardware
project
about
involving
a
node
B
that
can
reduce
the
costs
of
I,
mean.
I
So
right
now
our
current,
we
were
surprised,
so
our
current
commercial
hardware
is
sold
I
cells,
which
is
Chinese
manufacturer.
We've
had
good
success
with
them.
We
were
surprised
how
cheap
it
was.
It
sounds
like
these
prices
have
kind
of
lowered
a
lot
in
the
last
year,
but
also
Facebook
has
this
open
cellular
project,
which
we've
been
talking
with
a
little
bit,
and
that
is
designed
to
be
fully
open,
that
you
can
like
provide
they're,
providing
hardware,
schematics
and
stuff
I.
Think
LTE
support
for
that
is
upcoming.
I
This
is
a
space
where
we're
trying
to
because
a
lot
of
my
work
is
more
in
software.
We're
trying
to
create
I'm
trying
to
create
essentially
on
the
blog
I've,
got
kind
of
this
buyer's
review
thing
where,
as
we
work
with
different,
you
know,
B's
I
want
to
say
these
guys
were
great.
This
kind
of
struggled
I'm
trying
to
be
transparent
on
that
process,
but
really
it's
it's
not
a
place
where
we've
done
a
lot
of
deep
digging
or
I
haven't
personally
in.
L
I
There's
a
little
bit
of
right
now,
it's
in
kind
of
a
more
rudimentary
stage
of
like
making
sure
that
things
reboot
when
they're
needed
to
and
so
on
and
so
forth.
The
network
right
now
is
we're
talking
about
to
a
node
bees
and
one
EPC.
So
this
is
not
a
mind-blowing
ly.
Large
network
and
I
think
our
vision
going
forward
when
we
start
working
with
more
communities
is
that
each
one
of
these
will
be
a
self-contained
core
and
one
or
two
e
node
bees.
So
the
idea
is
less
this.
I
L
I
This
is
another
one
of
these
weird
points
that
I
think
it's
hard
designing
when
you
don't
know
where
to
start
so
the
transparent
solution
is,
this
is
going
to
be
owned
and
run
by
the
community,
so
their
bills
are
the
one
megabit
per
second
VSAT
connection
is
actually
billed
monthly.
So
there's
a
trade-off
of.
We
don't
want
to
discourage
usage
because
every
second
we
don't
have
a
pin
to
the
wall
as
a
second
lost.
I
So
we
have
that
going
on.
We
have
also
we
wanted
discourage,
can
use
during
congested
time.
The
goal
is
I
mean
all
extra
money
is
gonna,
be
going
back
into
the
community.
So
there's
no
reason
to
you
know:
there's
you
know
questions
there
about.
How
much
should
we
budget
to
revention,
replace
the
hardware
and
all
this
is
gonna-
be
a
big
community
discussion
going
forward
that
we
don't
want
to
drive
too
aggressively
you're
welcome.
M
Hi
max
Paula
cable
labs
thanks
great
work,
I
had
the
privilege
and
unfortunate
to
have
to
work
with
networks
recently
I'm
more
of
an
internet
guy.
So
you
know
all
this
little
3gpp
craziness
is
really
sticking
with
me.
One
of
the
work
that
actually
I've
been
doing
for
the
past
six
months
seems
to
work
in
some
other
areas
and
CPRS,
particularly
alliances
spectrum.
M
M
The
second
thing
is
in
this
direction,
since
you're
trying
to
save
costs
right,
especially
the
point
infrastructure.
One
of
the
things
that
we
looked
at
is
to
work
with
the
companies
called
that
produces
a
products
called
open,
EPC
I,
don't
know
if
you
are,
if
you
talk
with
them
and
basically
they
came
out
from
the
front
of
our
Institute
in
Germany,
and
they
have
a
fully
software
core
implementation
that
can
skate
up
in
the
cloud
from
a
raspberry
pi
into
a
cloud
deployment
fully
containerized
deployment,
so
something
that
I
would
suggest
you
to
convert
them.
M
They
plan
to
that
very
open
source
implementation.
They
have
an
opportunity
mutation
where
they
say
it's
very
outdated
and
they
yeah.
This
is
this
is
what
I
remember
and
they
like
3G
4G
and
the
building
5g
as
well.
So
you
might
they're
very
open
to
have
conversation
with
people,
so
I
encourage
you
to
come
to
them
and
I.
Don't.
I
M
Name
is
a
little
bit
misleading
open.
A
PC
is
not
open
source
implementation.
They
share
the
source
code
with
their
clients,
but
it's
not
an
open
source
implementation.
They
have
another
one
that
it's
not
called
open
a
PC
I,
don't
remember,
then
the
name,
but
it's
a
little
outdated.
So
the
open
source.
I
The
open
air
interface
project
is
essentially
that,
and
is
it
does
have
a
robust
community
and
it's
powered
by
the
urakami
guys
in
France,
but
we're
not
here
to
play
favorites
any
pcs,
any
PC,
and
if
one
works
better,
you
might
come
to
them
and
you'd
be
surprised.
How
much
of
the
project
that
we've
been
doing
over
the
last
year
has
been
peripheral
things
around
that,
like
once
you
get
the
EPC
up
and
running
the
rest
is
building
a
billing
package,
but,
like
all
these
other
things
that
so
we
can
swap
this
out.
I
M
N
N
I
I
Is
exactly
the
goal
we've
been
trying
to
create
I've
been
working
on
manuals
of
how
to
set
this
up
like,
conversely,
like
manuals
have
here
all
the
things
that
we
can't
just
do
for
you
like
ordering
Sims,
but
here
are
vendors
that
we've
worked
with
in
the
past.
The
goal
is
to
make
it
someone
with
ideally
IP
networking
experience,
but
potentially
not
even
that
they
want
to
get
coverage
out
somewhere.
How
can
we
help
you
every
step
of
the
way
is
very
much
the
the
the
end,
the
long-term
goal
of
this
project.
I
I
The
answer
is,
is
yes,
that's
what
I
did
it
took
me
a
while
to
figure
out
alright,
the
build
is
broken.
Why
is
the
build
broke
and
it's
supposed
to
work
here
it
with
enough
digging
and
painfully
setting
up
the
configurations
and
so
on
and
so
forth?
That
is
what
I
spent
the
last.
What
I
spend
the
first
couple
of
months
when
I
started
on
this
project?
Doing
what?
What.
N
I
N
A
I
The
answer
is
right:
now
things
are
licensed,
we're
we've
been
using
band,
8
I
think
so.
This
is
one
of
the
things
that
we've
been
trying
to
put
together
a
guide
for
which
gets
pretty
tricky
because
you
have
this
awkward
intersection
of
what
bands
are
supported
in
a
specific
country
by
the
handsets
which
goes
back
to
what
are
the
popular
bands,
which
then
goes
into
which
ones
are
allowed
and
so
on
and
so
forth.
I
A
Thank
you
very
much,
and
so
for
the
last
five
minutes,
because
we
want
to
give
the
next
team
five
minutes
to
get
ready.
We
are
keen
to
work
on
terms
of
reference
for
some
best
practices.
Bcp
I,
don't
think
we
would
call
to
be
cop,
but
in
any
event,
if
you
want
to
meet
up
with
us
tomorrow,
at
some
point,
Lando
and
I
are
gonna
sit
in
a
space
where
we
can
have
room
for
enough
people
if
you're
interested,
we'll
put
it
out
on
the
guy
list.
A
The
idea
is
to
get
those
terms
of
reference
hacked
in
the
next
two
weeks
and
then
come
up
with
a
table
of
contents
that
folks
can
live
with
and
then
use
whatever
everyone
agrees
is
an
open-source
mechanism
to
start
to
work
on
the
BCP
together.
Anyone
have
any
ideas
of
what
would
be
best
or
any
thoughts
you
want
to
share
with
us
on
what
would
be
useful
to
you
and
thank
you
from
the
community
sense.
O
A
Thank
you
for
that.
Any
other
suggestions
all
right.
Well,
thank
you
very
much
for
your
time,
thanks
for
being
here
over
to
the
crypto
guys
and
everything
we've
been
taking
notes.
Just
we've
been
frantically
typing,
both
an
ether
pad
and
I've
done
it
on
Java.
So
if
I've
missed
anything,
I'm
gonna
put
this
up
on
Gaia
for
folks
to
take
a
look
at,
and
let
me
know
if
I
got
some
of
the
acronyms
wrong
all
right.
Thanks
a
lot
Leandro
and
I
appreciate
your
coming
and
stay
tuned
and
appreciate
it.