►
From YouTube: IETF103-GAIA-20181106-1350
Description
GAIA meeting session at IETF103
2018/11/06 1350
https://datatracker.ietf.org/meeting/103/proceedings/
B
A
Hello,
everyone
welcome
to
Gaia.
We
have
an
interesting
experiment
on
today
with
most
of
our
participants
providing
presentations
remotely
so
we're
gonna,
give
it
a
go
and
see
how
it
works.
The
agenda
does
have
Lakshmi
going
first,
but
we
don't
see
them
yet
in
the
room,
so
we're
going
to
start
with
a
dose
or
nin
a
sec.
I've
got
to
do
a
little
housekeeping,
so
there
are
blue
sheets
circulating
in
the
room.
Please
sign
in
if
anyone's
willing
to
help
us
take
notes.
A
As
Matt's
just
asked,
we
have
someone
raised
their
hand
to
help
us,
maybe
add
a
thermal.
Thank
you
yeah.
Mostly,
the
Q&A
will
help
you
capture
or
capture
that
thank
you,
pardon
yeah,
there's
a
Java
room
open
and
my
name
is
Jane
coffin
and
one
of
the
co-chairs,
my
other
co-chair,
is
Leandro
Navarro
and
Leandra's
in
Spain
right
now,
Matt
Ford
who's
sitting
next
to
me
from
the
internet
city,
used
to
also
be
a
co-chair
of
Gaia,
so
Matt's
giving
me
a
hand
today,
so
we're
gonna
get
started
quickly
with
the
just
quick
note.
A
Most
of
you
know
the
IRT
F
IPR
policy,
but
we
just
want
to
make
sure
you
have
the
opportunity.
Sorry,
it's
a
little
bulky
the
opportunity
to
see
what
that
is,
and
it's
online
as
well,
so
there
it
is,
and
here's
our
agenda
and
time
plan.
So,
given
that
we
don't
see
Lakshmi
I
just
want
to
think
we're,
gonna
ask
you
to
go
first
and
do
note
that
we
are
working
on
a
mash-up
of
a
draft
on
different
connectivity
models
and
we'll
be
talking
to
you
about
that
toward
the
end
of
the
session.
C
Okay,
while
we
waiting
so
I
mean
to
do
myself
good
afternoon,
everyone,
my
name
is
a
tea
soreness
in
subtly,
so
I
am
from
I'm
a
senior
research
associate
at
Intel.
Our
nation
is
of
the
technology,
so
today,
I
will
keep
you
kind
of
the
update
of
what
we
are
doing
here
with
the
project.
Authenticity
come
the
network
that
have
tried
to
provide
the
connectivity
to
the
Royal
of
Thailand,
and
this
is
a
bliss
of
my
my
quarters
that
we're
working
together
with
any
solid
Nanta
pot
and
he
pulls
our
PT
student
data
polls.
C
C
So
this
is
what
I'm
talking
about
so
the
talk
that
is,
the
listeners
project
that
tried
to
use
utilizing
the
wireless
mesh
network
to
provide
Internet
connectivity
in
the
northern
part
of
Thailand
called
Mysore
district
I
mean,
which
is
a
border
between
Thailand
and
Burma.
We
started
project
since
2013
is
about
to
be
in
about
six
years
already,
and
we
have
already
poured
about
seventeen
remote
village
already.
C
The
target
is
run
based
on
the
volunteer,
so
we
have
the
student
and
volunteer
from
the
local
people
helping
out
for
the
deployments
and
maintenance
the
network.
So
we
could
not
go
so
far.
We
have
we
can't
deploy
almost
one
village
year
so
since
recently
interest
in
17th,
we
ship,
we
Shane
the
business
model
with
the
hell
from
th
nice
Foundation,
and
we
heard
that
the
new
BC,
the
motorcar
technique.
C
So
the
detail,
probably
at
the
body
else,
I
ornament-
we
already
have
documented
and
then
gonna
be
puppies
in
GLS
wash
so
that
18
we're
going
to
be
released
next
next
week.
So
from
then
we
could
be
able
to
deploy
the
network
like,
for
example,
from
last
year,
we
can
deploy
almost
six
village
one
year,
and
this
year
we
released
about
seven
village
already.
C
Ok,
so
in
terms
of
technology
actually
target,
we
is
the
attention
of
the
overall
project
called
Dumbo,
which
is
about
two
using
the
1s
mesh
network
to
to
support
the
post
disaster
scenario.
So
in
DUMBO
we
use
the
OSR
nothing
protocol
to
create
a
violation,
and
that
is
the
the
project
that
we
have
and
then
we
try
to
attend
to
use
in
the
HMDA
network
context.
C
You
can
see
that
from
the
equipment
that
we
use
this
most
of
them
from
the
outer
shell
equipment,
like
the
tp-link
I,
could
mention
that
this
more
one
that
I
showed
on
right
hand
side
that
the
lai-wan,
which
is
quite
possible
so
because
the
purpose
of
that,
because
we,
the
villager,
can
benefit
from
that.
But
when
yeah
it's
kind
of
disaster
happening
like
a
earthquake
of
writing
the
book
the
religious
can
took
out
the
router
and
then
they
can
th.
The
one
is
made
immediately.
C
We
have
any
people
Asian
or
regulation,
and
that's
the
purpose
of
that-
and
we
also
using
another
equipments
as
well
to
provide
a
better
coverage
area
like
this
one.
The
background
at
the
white
one
is
gone
and,
of
course,
we
use
a
mass
Philippe
I
to
run
our
own
local
services
inside
the
networks
as
well.
C
This
is
the
location
or
that
relate.
They
happened
in
a
few
tight
surrounding
by
the
mountains
and
we
have
a
lot
of
high
tree.
So
this
is
a
complete
non
laughs
I.
So
so
we're
looking
around
what
could
be
the
next
solution
for
our
two?
So
this
kind
of
issue
and
every
father
TTY
space-
will
be
a
good
option
for
us,
so
it
was
played
basically,
is
the
TV
channel
to
respect
them
that
we
know
that
this
is
a
low
frequency
band.
Four
hundred
we
actually
have
USF.
C
D
D
C
Up
to
early
next
year
for
the
nineteen,
so
the
proposal
project,
please
try
to
do
kind
of
exporting
the
benefit
of
the
tea
white
space.
So
we
we
hold
right
colony
with
co-holder
license
from
the
whole
band
for
the
TV
channel,
four
hundred,
seventy
two
seventy
to
ninety,
and
so
it
is
going
to
be
the
first
u.s.
trial
in
silence
as
well.
C
Apart
from
that,
we
need
to
tell
you
a
lot
of
measurement
to
identify
what
kind
of
whitespace
in
the
euro
area.
This
is
the
main
purpose
of
the
project,
so
to
start
with
to
start
a
project,
so
we
first
have
to
do
a
lot
of
measurement
to
identify
the
ID
the
respect
my
space
and
those
is
equipment
that
we
use
at
the
moment.
C
So
we
use
the
low-cost
spectrum,
my
so-called
basic
Cola,
the
benefit
of
that,
because
it's
cheap,
we
can
buy
a
lot
of
them
and
then
we
can
keep
them
to
a
minimal
interior
like
student
to
run
the
measurement
in
their
home
and
in
the
house,
and
we
also
using
the
small
equipment
like
a
laptop
or
an
a3
pie,
and
we
also
get
help
from
IC
IC
DP
in
TST.
Italy
is
our
partner
to
develop
the
software
tool.
According
the
Spector
measurement
form
data
support,
and
then
they
call
them
in
digital
format.
C
C
Yeah,
so
in
terms
of
location,
we
are
considering
interesting
into
different
location
like
up
inside
the
building
indoor
and
outside
the
building
out
of,
and
we
also
considering
the
high
of
antennas
as
well
like
at
the
carrier
waves
or
at
the
rooftop,
like
15
meters,
10
meters
and
so
on.
So
what's
the
problem.
So
now
we
have
a
bunch
of
data,
so
the
next
thing
that
we
had
to
do
identify
what
kind
of
TV
white
space
so
the
the
common
method
used
in
the
little
Asian.
C
Mostly
we
are
using
the
fixed,
a
show
where
we
set
up
the
threshold
for
to
to
identify
the
signal,
for
example,
if
the
signal
is
all
larger
than
threshold
we
consider
this
is
the
signal
the
channel
is
being
used.
Fcc
recommended
the
100.
Minus
114
is
the
recommendation
from
there
from
the
FCC,
but
in
the
literature
people
say
that
it
is
quite
conservative,
4
and
then
people
prefer
to
use
the
100.
C
This
is
kind
of
blue
of
some
number
and
then
but
being
used
that
one,
and
we
also
have
another
threshold
Hawk
visit-
a
show,
for
example,
if
two
and
if
the
signal
is
not
being
used
more
than
80%,
we
consider
that
this
is
a
widespread.
So
what
happened
with
this
net
thought
this
is
showing
the
measurement
from
the
same
location
with
the
same
with
the
same
equipment.
This
is
showing
the
channel
4
channel
26.
So
if
now
60,
you
can
see
by
eyes,
we
had
both
of
them
had
the
same
pattern.
C
So
the
the
dark
blue,
showing
that
that
spectrum
is
analysis
free,
is
not
being
used,
but
the
problem
is
from
these
two
measurement.
We
found
that,
from
with
the
fix
ho
method,
the
first
measurement
we
ended
by
the
10:00
news,
10
channels
is
being
used,
but
the
second
strain
measure
ensuring
that
14
channels
is
easy
and
easy.
So
so
this
is
not
quite
accurate.
Why?
C
C
We
can
you
see
if
they
have
many
values
in
terms
of
noise
and
there
were
CDs
as
well.
So
adaptive
Center
should
be
better
solution
for
that.
So
we
are
literally,
we
just
published
the
work
that
we
done
here.
A
value
tested
in
you
can
look
into
the
paper
and
what
we
were
just
a
publisher
is
income
passed
early
this
year,
so
so,
in
terms
of
comparison,
this
is
we
look
into
the
parameter,
panic
or
false
alarm
detection
and
miss
appropriate
false
alarm
and
copy
of
miss
detection.
C
So
from
the
mist,
the
false
alarm
in
the
false
alarm
in
the
channel
is
free,
but
we
decided
the
system
show
is
not
free,
that's
awesome
and
on
a
white
horse,
I
was
hard
to
miss
detection
when
the
channel
is
not
free,
but
we
mentioned
that
is
free.
Ok,
ok,
improvement
is
in
is
doing
quite
well
like
in
terms
of
false
alarm,
with
adaptive
threshold.
C
So
so
we
have
compared
the
data
measurement
from
the
eighty
compass
because
in
eighty
compass
we
have
a
good
car
truth
from
the
from
the
the
liquidator
because
they
have
the
report
from
the
specimen
location
in
us
and
it
is
quite
accurate
and
quite
reliable.
So
we
compare
the
results
and
be
showing
a
benchmarking
that
we
showing
that
the
our
adaptive
threshold
could
be
good
to
to
do
more
benchmarking
to
another
occasion.
So
next
one
so
from
that
technique.
C
C
C
Have
Internet
can
be
over
there?
Thus,
the
distance
between
the
base
station
to
the
village
is
rarely
from
500-meter
to
up
to
10
kilometers,
so
it
could
be
fit
to
the
device
play
the
technology
we
use.
We're
gonna
use
a
standardization
estilo
2.11
AF,
which
is
called
super
Wi-Fi
as
well.
At
the
second
solution,
we
look
into
the
access
layer
inside
I
realize
scholarly.
We
have
only
one.
C
But
we
also
expand
using
the
towards
paid
for
the
LTE
small
cell
base
station
inside
the
units
as
well
so
Gophers.
So
we
decided
to
choose
the
equipment
from
operation
equipment
available
in
the
market
called
cousin
gentry
and
which
is
the
support,
a
zero
to
11
a.m.
and
we
just
got
equipment
easily
last
month,
and
we
now
try
to
take
them
up
inside
the
campus
to
checking
that
ok,
the
Commish
working
properly
and
what
kind
of
at
the
to
Center
we
can
make.
So-
and
this
is
the
diagram
architecture
that
we
we
set
up.
C
Ok,
we
look
quite
quick
over
there.
Ok
and
so
the
maximum
is
10.
We
test
I
think
we
just
made
it
last
week.
It's
about
one
one,
one
point
two
kilometers-
and
this
is
the
the
pro
finder
because,
like
we
got
a
pretty
good,
receive
signal
power.
The
to
put
is
about
15
10
megabit
per
second
for
applicant
darling
pack,
a
lot
we
are
see
from
almost
zero
percent
packet
loss.
Itt
is
quite
low.
C
Okay
at
a
second
solution,
as
I
mentioned,
we're
also
working
on
in
power
or
working
with
Microsoft
Research
suitcase
as
well
to
do
the
trial
for
the
LTE
using
over
at
ey
space,
and
this
is
equipment
that
we
use
and
donate
is
just
given
by
the
Microsoft
Azure
UK.
So
that's
this
our
base
station,
that's
antennas,
SIM
card
and
also
the
juice
inside.
We
had
the
knife
I
like
a
pocket
Wi-Fi.
So
this
is
never
configuration.
We
have
the
fab
with
me
in
the
base
station.
C
C
Unfortunately,
since
this
is
working
with
Microsoft
right
and
then
the
cloud
that
we
use
in
Microsoft
cloud-
and
luckily
we
don't
have
them
the
Microsoft
cloud
in
Bangkok
the
concept
one,
it
could
be
in
Singapore
and
Hong
Kong.
So
we
already
do
the
trials
of
almost
two
months
already,
and
so
that
is
the
base
station.
We
put
the
antenna
in
the
base
station
in
IT.
C
C
So-
and
this
is
a
combination
that
we
can
far,
we
found
that
we
get
the
commercially
about
500
meters,
which
is
the
cover
the
part
of
the
campus
already,
and
then
we
do
some
testings
as
well
like
we
testing
inside
the
lab
in
the
lab
in
my
lab
and
the
distant
from
our
lab
to
that
position
is
about
200
meter,
and
this
is
head
off.
We
do
just
a
simple
speed
test
to
see
that,
okay,
what
kind
of
the
bad
we
can
see
that
we
have,
and
also
we
mentioned
exam
as
well.
C
We
did
two
types
of
testing
like
retesting
the
server
in
Hong
Kong,
because
at
the
time
were
for
testing
the
EPC
is
located
in
Hong
Kong
choose
a
Hong
Kong
for
the
for
the
primary
EPC,
and
we
do
the
testing
compare
with
a
bkk
we
found
out
for
the
Hong
Kong
server.
The
Kong
Hong
Kong
case
is
pretty
much
slightly
better
compared
to
the
big
here
because,
as
we
know
as
the
cars
in
Hong
Kong
right,
so
you
should
get
a
better
performance
as
well.
C
But
this
is
what
quite
interesting,
because
we
want
to
deploy
the
system
equipment
in
net,
which
is
far
away
from
Bangkok,
so
we
might
get
the
problem
in
terms
of
connectivity
as
well,
because
the
EPC
sees
in
the
cloud
and
in
the
car
is
outside
the
country.
So
this
is
might
be
another
new
research
dimension
that
we
are
we
are
focusing
with
as
well,
so
the
next
step,
actually
the
mini
step.
C
The
first
thing
we
want
to
define
the
system
that
we
have
like
the
to
Isis
backhaul
and
also
the
Audi,
is
more
sail
in
the
path
net,
which
is
a
northern
part
of
Thailand.
So
we
plan
to
do
that.
You
said
maybe
next
month
at
earliest,
and
we
also
try
to
explore
another
method
for
the
idea
for
the
spectrum
measurements
as
well,
because
we
want
to
build
by
introducing,
like
hydro,
machine
learning,
neural
network,
to
to
better
build
the
accurate
data
for
thus
the
white
space.
And,
of
course
we
keep.
C
We
still
keep
expanding
that
net
to
in
order
to
get
connect
people
as
far
as
we
can
so
that's
it
I
think,
okay,
so
yeah!
This
is
our
team
for
the
tag
native
look
I
have
to
mention
that
with
the
Professor
K
tonight
is
the
whole
team
that
we
working
together,
and
we
were
I,
would
like
to
thank
th
Nix
as
well.
C
Is
our
sponsor
our
sponsor
so
far
as
in
the
beginning
of
the
project,
and
we
have
n
BTC
to
give
us
the
research
grant
and
also
the
licen,
and
we
thank
you,
I
CDP,
to
helping
out
for
this
measurement,
and
you
guys
on
this
campus
and
Microsoft
Research
UK
for
for
cooperating
it
of
LTE
solution.
So
this
is
it.
Thank
you.
C
A
B
B
B
C
That's
correct,
actually,
the
critical
parameter
it
could
be
here
are
the
myse
detection,
because
if
you
miss
detected
Ismene
you
try
to
interfere,
interfere,
the
primary
user.
So
far,
our
statistic
is
not
it's
just
the
the
best,
the
better
way
we
get
about
30
percent,
which
is
still
quite
high,
and
we
keep
an
eye
on
that
as
well,
but
because
we
try
to
introduce
another
new
technique
like
machine
learnings
or
another
kind
of
neural
networks
as
well.
C
C
E
E
E
C
Okay,
yeah
actually
at
this
one,
that
the
product
we
have
here
just
one
particular
question
that
we
have
a
bunch
of
them.
Not
only
one
location.
We
have
at
least
Eva
at
at
the
end
size
as
well.
Yeah,
both
sigh
but
I,
first
mentioned
that
because
we
choose
the
location
inside
a
campus,
because
we
have
four
good
data
from
the
in
beta
from
the
liquidator,
because
they
have
accurate
data
which
respect
is
use
or
not,
and
we
use
a
mass
account
which
compared
to
our
measurement.
G
G
C
Yeah
that
could
be
I
think
what
a
human
is,
not
not
the
big
problem,
because
we
can
give
them
to
run,
and
we
can.
We
still
have
the
funding
because
we
bought
equipment
anyway,
but
the
main
problem
is
gonna,
be
the
license,
because
now
our
license
is
gonna
be
end
by
next
year,
and
hopefully
we
could
get
attention
and
we
don't
do
more
trial
to
to
support
those
people.
But
we
try
to
do
these
projects
as
a
final
project
and
then
try
to
convince
our
liquidator
okay.
C
D
F
G
A
A
C
A
You
have
questions
for
either
sort
and
find
him
in
the
halls,
and
one
thing
that
he
was
mentioning
with
respect
to
TV
white
space.
Is
that
there's
a
big
regulatory
push
on
for
some
folks
in
5g?
They
also
want
some
of
that
TV
white
space
space,
so
the
community
networks
that
we
work,
where
they're
having
a
bit
of
a
difficult
time
trying
to
convince
regulators
to
let
them
use
some
other
TV
white
space.
So
it's
been
an
interesting
endeavor
Laxmi
you're
up.
Can
we
give
you
a?
H
Fantastic
hello,
everybody,
sorry
I.
If
I
got
a
little
bit
delayed
to
bootstrap
the
session,
I
was
still
trying
to
wade
through
the
system
to
get
the
exact
link
up
so
I'm,
very,
very
happy
to
press
and
so
Jane
do
I.
Have
my
slides
up.
You
do
okay,
awesome.
How
do
I
you
to
go
to
the
next
slide?
Should
I
just
say
next
slide.
Absolutely.
A
H
Thank
you.
So
it's
a
great
pleasure
to
introduce
to
you
all
Gaius
networks.
This
is
the
very
first
time
I
believe
we
are
giving
this
talk
and
we
all
a
lot
of
gratitude
to
the
gaya
community
for
the
creation
of
our
new
startup
gaius
networks
and
it's
been
in
stock
mode
and
today's
the
first
day
we
are
declaring
it
open.
H
Gaius
comes
from
the
male
version
of
Gaia
and
we
have
a
very
shared
vision
with
respect
to
all
the
goals
that
I
was
trying
to
achieve,
and
we
hope
to
work
with
the
Gaia
community
to
make
both
Gaia
and
Gaius
very
successful
slide.
Number
two:
here's
a
brief
description
of
the
guys
networks
team.
As
you
can
see,
the
person
in
charge
running
the
show
is
going
to
be
a
Juna
gonna
play
the
CEO
role
and
we
have
a
collection
of
six
PhDs
one.
H
H
There
are
many
places
around
the
world
where
servers
are
sitting
in
a
developed
country
and
there
is
very
little
local
content,
generation,
distribution
and
consumption,
and
that's
one
of
the
core
problems.
Gaius
aims
to
fix,
and
if
you
look
at
mobile
web
in
challenged
environments
in
the
next
slide
slide
number
four,
it's
actually
not
surprising.
To
note
server
latency
is
like
two
to
three
seconds.
H
We
have
three
scene
scenarios
where
DNS
sometimes
takes
15
seconds
to
resolve
through
the
curse
of
DNS
queries-
and
you
know,
web
pages,
with
hundreds
of
objects
and
30
to
40
connections
can
take
over
a
minute,
sometimes
to
load,
and
sometimes
it's
not
just
a
problem
of
network
connectivity.
It's
just
that
the
web
is
a
mess
right.
So
if
you
want
to
really
fix
this
problem
in
gaius
in
the
next
slide
slide
number
five,
we
do
two
completely
different
things.
First,
we
want
to
design
a
web
where
the
content
is
really
ultra
lightweight.
H
You
know:
can
you
design
content
that
is
20
times
smaller
30
times
smaller
50
times
smaller
than
the
standard
web
page
that
you
face
today?
We
think
javascript
can
be
a
curse.
Can
you
actually
design
web
pages
with
no
JavaScript,
and
so
one
of
the
things
we
have
done
here
is
we
have
designed
the
new
mobile
app
specification
language
and
a
lightweight
content
ad
platform
which
I'll
talk
about
which
enables
really
ultra
lightweight
content.
The
second
aspect,
from
the
networking
side
is:
can
we
make
content
ubiquitous
and
really
fast?
H
H
H
So
what
we
want
to
do
is
we
want
to
have
a
gaius
app
that
people
use
right,
which
is
basically
a
gaius
browser,
and
we
want
to
have
intermixing
of
content
from
local
content
providers
and
local
ad
providers.
So
content
is
going
to
be
viewed
in
the
concept
of
channels
and
is
going
to
automatically
get
intermixed
with
ads
on
the
fly
to
present
content
as
channels
to
the
end
user
using
a
gaius
app.
H
H
There
you
have
content
providers
having
small
cloudlets
right
and
the
cloudlets,
essentially
hosting
content
off
a
content
provider
right
and
when
somebody
clicks
a
particular
content
on
a
mobile
phone
right
through
the
app
each
content
request
is
going
to
in
essence,
create
and
auction
a
content
ad
option
where
the
ad
auction
is
going
to
run
on
the
edge
and
the
right
ad
is
going
to
fill
in
into
the
right
contact.
Now,
unlike
a
traditional
ad
of
Ad
Exchange,
here,
the
ad
exchange
is
controlled
by
the
edge
and
what
we
want
to.
H
Essentially,
look
for
is
local
ad
providers.
Let's
say
I
am
producing
milk
and
I'm.
Distributing
milk
in
Rwanda
right
do
I
ever
even
get
an
opportunity
to
advertise
in
any
place.
I
barely
get
any
opportunities,
like
maybe
local
radio.
If
you
have
a
platform
like
this
I
want
to
enable
that
milk
producer
to
be
a
potential
content,
add
provider
who
could
be
a
small
amount
of
money
and
get
advertising
as
part
of
local
channels.
A
B
H
So
I'm
slight
I'm
in
slide
number
seven,
so
I
talked
about
the
pentad
exchange.
That's
the
core
product
offering!
So
it's
a
mobile
app
with
a
content
exchange
running
on
the
edge
so
slide
number
eight
I'll
go
through
it
a
little
bit
briefly.
Essentially,
what
runs
on
an
app
is
this
layered
stack
there?
We
can
take
a
traditional
TCP
TLS
encrypted
framework
and
run
our
own
congestion
control
called
al-sisi
on
top,
and
we
have
the
specification
language
called
mamo,
and
this
browser
can
basically
handle
Mammal
content
and
show
mamo
content
on
a
phone.
H
So
it's
almost
like
a
mammal
browser
where
mammal
stands
for
mobile,
app,
markup
language
and
al-sisi
is
an
efficient
protocol.
That
I
will
I'll,
probably
very
briefly,
get
to
at
the
very
end,
so
slide
number
nine.
A
brief
description
of
mammal,
so
manual,
isn't
is
a
relatively
trivial
concept,
but
it
can
appear
a
little
bit
complicated.
Okay,
so
think
about
a
web
page
right.
Think
about
a
web
page
is
a
rectangle
or
a
collection
of
several
rectangles
where
each
rectangle
is
basically
an
object
right.
H
So,
given
a
web
page
can
I
basically
precompile
the
whole
web
page
such
that
I,
completely
flatten
out
a
Dom
structure
right
and
create
a
new
web
page,
which
has
no
recursive
request.
No
JavaScript
and
each
object
has
a
very
basic
set
of
primitives,
including
the
ability
to
refresh
right.
So
my
claim
is,
you
can
create
a
web
page
that
is
basically
having
the
same
look
and
feel
as
the
original
page
without
any
JavaScript,
but
basically
gives
you
the
same,
look
and
feel
and
is
twenty
times
smaller,
so
slide
number
10.
H
Here's
an
example
of
two
such
pages.
So
if
you
look
at
the
two
pages
on
the
right,
they
are
the
pictures
of
the
NIH
web
page
and
in
the
left
you
have
a
picture
of
one
of
the
new
sites
right
and
the
left.
One
is
actually
a
mammal
page,
and
if
you
take
the
nih
pages,
the
one
in
the
middle
is
the
actual
nih
page
and
the
one
on
the
right
is
the
mammal
page
that
we
create.
H
H
Now,
how
do
people
think
about
this
right?
So
we
compared
mammal
with
Opera,
Mini
and
chrome,
and
we
showed
that
it
loads
much
faster
and
its
size
is
significantly
smaller
compared
to
even
Oprah
mini
right
and
we
compared
it
with
different
fertility
levels.
So
our
app
has
the
ability
to
change
the
fertility
level
where
mammal
low
means
very
low
fertility
and
mammal
high
means
full
fertility
level
right.
H
But
mammal
is
a
straight
one,
shoot
one
shot
download
of
a
webpage,
and
the
second
thing
is:
we
ran
a
user
study
across
hundreds
of
users
and
we
showed
that
80
to
90%
of
people
who
looked
at
mammal
pages
found
that
it
was
very
similar
to
the
very
to
the
original
page,
while
10
to
20%
felt
it
was
similar.
Very
few
people
said
it
had
little
similarity
to
the
original
page
right
and
going
on
to
slide
number
12
talking
a
little
bit
of
the
content
ad
platform.
H
Essentially
it's
as
I
mentioned
it's
running
an
ad
option
and
we
can
take
different
types
of
ads.
We
can
take
video
ads,
audio
ads
image,
ads
and
so
on,
and
we
can
change
the
fidelity
of
the
ads
and
we
can
also
have
ads
that
can
the
same
ad
that
can
be
represented
over
an
IVR
call
or
an
SMS
are
the
same
ad
that
is
represented
as
a
full
video
ad
for
a
particular
webpage.
H
So
we
specifically
asked
ad
providers
to
give
ads
at
different
fidelity
levels
and
we
can
tune
the
type
of
ad
that
gets
displayed
on
the
fly
as
a
function
of
the
use
of
bandwidth
and
preferences
and
the
content
providers
preferences
okay,
so
map.
So
the
ad
platform
is
compatible
across
different
types
of
content.
H
We
wrote
a
second
paper
about
it,
and
then
we
have
a
follow-up
paper
at
use
makes
a
DC
as
a
short
paper,
and
we
are
building
upon
it
to
make
it
into
like
a
full-fledged
application
layer.
Congestion
control
stack,
so
our
mobile
platform
between
the
mobile
app
and
the
edge
can
basically
run
al-sisi
and
can
guarantee
you
very
high
performance
and
with
very
low
delay.
Characteristics
like
a
15
times.
H
H
We
also
have
a
distributed
content,
diffusion
framework
called
X
cache,
which
allows
you
to
smartly
move
content
to
the
edges
in
in
a
sort
of
a
decentralized
manner
and
manage
these
edge
caches
in
an
interesting
manner,
so
that
we
can
have
live
content
close
to
the
edges
as
quickly
as
possible,
and
this
distributed.
Caching
solution
performs
better
than
traditional
caching,
solutions
that
have
been
proposed
in
the
architecture
and
we
have
compared
it
with
legacy
caching
solutions
and
have
shown
between
a
2x
and
a
5x
gain
because
of
this
distribution.
H
Caching
system,
so
gaius
is
trying
to
put
all
these
pieces
together
to
design
a
really
fast
lightweight
web
and
it's
specifically
targeting
emerging
markets,
but
the
core
crux
of
the
proposition
is
to
enable
localized
content
and
localized
ads
to
create
a
sustainable
ecosystem
in
each
city
to
be
able
to
run
a
decentralized
web
with
that.
Let
me
stop
here
and
take
questions.
Thank
you
very
much.
Thank.
A
C
H
H
The
second
part
of
the
answer
is
people
who
really
feel
that
they
want
to
maintain
the
look
and
feel
of
a
traditional
web
page
right.
We
have
gone
through
the
extra
mile
to
say
that
we've
played
around
with
5200
websites
and
we
have
been
able
to
show
that
we
can
automatically
convert
a
variety
of
existing
web
pages
to
have
the
same,
look
and
feel
you
using
mammal,
and
we
have
been
able
to
support
dynamic
objects.
H
We
have
been
able
to
support
forms
and
we
have
being
able
to
support
refreshing
of
content,
and
so
on,
so
we've
been
able
to
support
most
of
the
basic
primitives,
including
audio
video
and
so
on,
as
part
of
mammal,
and
it's
it's
mostly
centered
around.
The
idea
of
pre
compiling
a
web
page
rather
than
having
recursion
and
JavaScript
to
be
so
mammal,
does
not
allow
a
very
detailed
analytics
type
of
request
that
some
traditional
web
pages
might
have
there.
You
can
move
your
cursor
and
you
know
you
can
track
your
cursor
about
certain
things.
H
C
H
H
So
in
some
sense
the
content
provider
has
to
you
know
opt-in,
to
be
part
of
gaius
to
make
that
type
of
content
visible
within
the
gaius
ecosystem
and
Gaius
is
primarily
designed
for
localized
content,
so
any
user
can
start
creating
content.
So
you
have
a
chicken-and-egg
problem
of
how
you
bootstrap
the
system
right.
So
one
way
to
bootstrap
the
system
is
to
get
local
popular
content
channels
to
participate
as
part
of
the
guys
ecosystem.
So
you
want
to
start
with
a
city
granularity
where
the
content
Reuters
first
decide.
H
A
A
A
C
A
A
I
Thank
You
Jane.
Thank
you.
Thank
you.
Everyone
Thank
You
Gaea
community
for
this
opportunity
to
be
a
I,
the
Alliance
for
affordable
Internet,
which
is
hosted
by
the
death
foundation
but
partnered
by
a
number
of
entities,
including
folks,
like
the
Swedish
government
or
the
US
government,
the
UK
government,
and
also
our
host
governments
such
as
Myanmar,
Bangladesh,
Nigeria,
Ghana,
Kenia
and
so
on.
I
Something
Lakshmi
just
now
elicited
very
beautifully
I'm
presenting
this
very
brief
ten-minute
presentation
without
a
PowerPoint
presentation,
as
well
as
a
video
stream,
and
also
I,
must
admit
here
that
I,
my
presentation
is
not
about
any
technological
or
engineering
solution
such
as
the
last
two
presentations,
the
TV
white
space
1
by
AI
key
and
the
gaius
solution
for
local
content
and
local
content.
I
had
streaming
bye-bye
Lakshmi,
whereas
I'm
going
to
focus
more
on
the
policy
interventions
that
are
required
to
make
web
a
web.
I
That
is
universally
accessible,
also
by
the
60%
bottom
percentile
income
percentile
groups
within
within
countries,
but
also
the
regulatory
measures
that
could
be
undertaken
to
to
ensure
that
this.
This
could
happen
and
in
this
I'm
going
to
also
touch
base
upon
this.
The
age-old
solution,
which
is
the
Universal
Service
Access
Fund,
which,
which
was
introduced
a
few
decades
ago,
a
globally
tried
and
tested
mechanism
to
spur
ICT
investment
in
underserved
and
unserved
areas.
So
I'm
going
to
talk
about
that
as
well.
I
Now
on
contrary
to
many
people
who
don't
like
to
have
this
levy,
the
universal
service,
access
fees,
which
is
contributed
largely
by
the
course
mainly
due
to
extend
coverage
in
rural
areas
where
the
the
the
market
market
fails.
Many
people
have
done
the
research
and
come
to
a
conclusion
that
universal
access
funds
don't
really
work
well.
Well,
we
take
note
of
that.
I
But
if
you
look
at
the
latest
affordability
report,
take,
for
instance,
the
Asia
Pacific
continent
has
actually
achieved
one
for
two
target.
Would
that
mean
that
a
sea
of
Pacific
we
don't
have
to
have
any
local
content
or
connectivity
solutions?
The
answer
is
no,
because
although
the
1
for
2
target
is
met
when
we
dig
deeper
and
analyze
the
data
as
to
who
access
the
web,
we
do
come
to
recognize
that
the
top
40
percentile
income
group
only
have
access
and
the
internet
is
not
affordable
to
the
bottom
60%
rail.
I
In
many
cases,
and
in
the
case
of
Myanmar,
for
instance,
a
country
which,
which
is
long
ago,
achieved
that
with
the
4G
connectivity
at
the
fore,
we
still
do
find
that
people
not
taking
advantage
of
broadband.
So
that
said,
I
want
to
also
mention
a
recent
report
that
we
be
released
in
Africa.
That
was
in
May
this
year,
which
categorically
shows
that,
amongst
the
countries
that
we
surveyed
37
African
countries
do
have
a
UCF
setup,
but
62
only
62
percentage
of
these
funds
are
active.
I
That
means
a
good
number
of
countries
are
not
utilizing
this
fund
and-
and
there
is
an
unspent
front
of
total
estimated
off
a
billion
available
in
in
Africa,
which
could
be
spent
in
multiple
solutions,
including
the
solutions
that
the
two
solutions
that
will
demonstrated
just
now,
the
TB
white
space
and
the
local
content
solution.
But
we
also
want
to
take
note
of
the
need
for
spending
on
women
bringing
women
online,
so
the
408
million,
which
is
unspent
in
Africa,
we
believe,
can
be
good
enough
to
bring
six
million
women
online
and
provide
a
6
million.
I
A
16
million
women
with
the
digital
skills,
women
and
girls,
with
digital
skills,
and
also
a
good
mechanism
for
us
to
spend
the
money
that
has
been
collected
sitting
idle.
Although
in
many
countries
we
have
seen
certain
steps
undertaken,
including
countries
such
as
India,
Colombia,
Benina,
Ghana
and
so
on.
We
do
see
that
there
is
an
ample
opportunity
for
us
to
use
utilize
these
funds
to
bring
women
online.
We
also
take
note
of
another
policy
opportunity
which
is
increasing
public
access
amongst
the
underserved
and
unserved
communities.
I
Delivery
platform
that
we
built
when
the
web
was
a
very
primitive
stage,
using
world
space,
radio
and
other
technologies
to
push
and
pull
content.
But
many
of
these
remained
as
experiments
until
now
because
of
the
lack
of
policy
and
financial
support
rendered
to
these
kinds
of
networks
that
have
come
into
play.
So,
in
conclusion,
what
I
want
to
highlight
here
is
if
the
universal
service
access
funds
are
spent
properly,
with
the
right
people
on
the
board
in
a
transparent
manner,
the
more
and
more
multi-stakeholder
involved
in
the
administration
of
the
fund.
I
As
women
in
Benin.
We
have
a
BAC
ocap
program
that
supports
women,
rural
women,
entrepreneurs
with
mobile,
based
a
mobile
phone
based
system
that
provides
a
price
of
local
agriculture,
goods
such
as
corn,
millet,
soya
bean
and
peanuts,
and
so
on,
because
these
are
all
sailors
who
need
to
know
the
latest
price
so
that
they
can
bar
gain
a
good
deal
in
Costa
Rica.
We
have
seen
good
examples
of
women
actually
in
the
female-headed
households,
Behram
mobile
access
is
promoted
and
you
know
they
become
really
entrepreneurs
in
distributing
services.
I
So
I
would
like
to
stop
here
by
also
highlighting
Botswana
and
Thailand
there's
a
good
practice
when
it
comes
to
public
Wi-Fi
hotspots
public
access
internet
in
India,
you
have
more
than
200,000
common
service,
centers
UCF
utilized
in
expanding
the
broadband
network.
Through
the
broadband
limited,
the
straight
fronted
company
in
Botswana.
We
we
see
a
number
of
sites
connected
through
Wi-Fi
hotspots,
such
as
hospitals,
bus
stops,
taxi
ranks
and
so
on.
I
So
I
want
to
utilize
this
opportunity
to
emphasize
that,
in
addition
to
the
local
connectivity
content,
technological
solutions,
we
also
need
to
advocate
for
policy
and
regulated
interventions
that
enable
us
to
utilize
the
windows
of
opportunities
such
as
the
the
universal
service
obligation
funds
to
not
only
provide
connectivity
solutions
but
to
grow
the
ecosystem
in
in
rural
areas,
so
that
we
do
not
have
failures
of
our
pilots
but
successes
of
our
scallops.
Thank
you.
Thank.
A
As
shadi
saying
in
Africa
alone,
you've
got
over
400
million
dollars
sitting
in
different
regulatory
accounts
that
could
help
support
internet
exchange
points,
community
networks
and
other,
and
so
folks,
like
a
for
AI,
we're
taking
a
small
look
at
it,
but
really
a
for
AI
has
done
a
great
job
on
this,
as
has
the
Association
for
progressive
communication,
so
we
keep
promoting
with
people.
We
work
with,
not
only
in
the
technical
side
that
you
need.
This
policy
related
aspect
pushed
so
shadi.
Thank
you
very
much.
We
appreciate
it.
A
This
was
a
much
appreciated
that
you
did
this
on
short
notice
as
well.
So
any
other
questions
for
shadi.
If
not
we're
gonna
move
over
to
char
Bonnie
who's
coming
from
Mumbai
and
she
works
at
a
university.
There
she'll
briefly
introduce
herself
she's
another
partner
of
many
of
ours
on
connectivity,
local
connectivity
in
villages
and
also
a
tech
options
that
they're
they're
taking
a
look
at
social
bunny.
Could
you
quickly
introduce
yourself
and
we'll
pull
up
your
presentation.
K
J
K
Hello,
I'm
Shivani
from
granmarg
at
the
Indian
Institute
of
Technology,
Bombay
and
I
work
here
as
a
senior
project
research
scientist
in
one
of
the
eight
ideas
I
have
actually
presented
about
what
granmarg
is
working
ahead
on
the
projects
that
we
are
taking
forth:
rural
connectivity
to
the
remote
and
unserved
villages
of
India.
We
have
we
are
now
today,
I
am
going
to
present
to
you
about
connecting
villages,
the
role
of
village
administration.
This
is
not.
K
K
If
you
look
at
the,
if
you
look
at
the
objectives,
the
objective
is
that
till
now,
75
percent
of
rural
India
is
still
unconnected.
Some
of
it
doesn't
even
have
voice
connectivity,
and
the
Government
of
India
is
really
trying
hard
to
connect
these
villages
through
the
valid
broadband
nigam
limited.
That
is
a
b
BN
l,
which
is
laying
optical
fiber
at
for
connecting
the
gram
panchayats.
That
is
the
that
is
that
that
is
one
of
the
locations
where
the
connectivity
goes
from
the
from
the
telephone
exchanges.
That's
where
it
goes.
K
Gram
panchayat
is
sort
of
the
head
village
and
it
has
three
to
four
villages
along
with
it.
The
government
also
has
agenda
that,
if
you
look
into
the
agenda
of
the
parrot
net
is
that
the
government
is
agenda
only
to
only
connect
the
gram
panchayats,
not
the
villages
and
that's
a
that's
a
really
sad
part
of
the
story
that
eventually,
even
if
the
gram
panchayats
are
connected,
the
villages
in
India
can
never
get
connected
and
what
is
the
status
of
these
villages?
K
Some
of
these
villages
are
remote
and
unserved
villages,
and
these
remote
and
Unsoeld
villages
actually
do
not
will
not
get
connectivity
for
a
longer
duration
of
time.
They
even
are
not
connected
through
private
telecom
operators
or
or
through
fibre.
So
these
the
the
probability
that
these
villages
will
remain
unconnected
for
a
longer
duration
of
time
is
even
higher
than
even
the
villages
that
are
very
close
by
to
the
highway
or
close
by
to
the
cities.
K
The
project
is
going
very
slow
and
this
the
first
phase
of
the
project
is
over,
where
they,
where
they
have
connected
almost
100
125,000
GPS
the
gram
panchayats,
but
there
is
still
a
lot
that
needs
to
be
connected,
so
it
isn't
total
to
two
hundred
and
fifty
thousand
gram
panchayats
that
need
to
be
connected.
So
there
is
a
decision
now
taken
by
the
government
that
125,000
remaining
GPS
can
be
connected
on
an
optimal
mix
of
solutions
like
fiber,
radio
and
satellite
and
in
the
as
an
interim
measure,
while
the
fiber
is
being
laid.
K
So
it's
like
radio
and
satellite
other
options
at
the
government
has
taken
up
the
agenda
is
that
to
two
hundred
and
fifty
thousand
gram
panchayats
need
to
be
connected
before
the
next
election,
so
it's
like
2019,
2019
March
is
going
to
has
being
set
as
a
cutoff
date
for
connecting
entire
rural
India.
So
next
slide
so
going
to
this,
what
is
we
have
done?
K
We
have
looked
into
what
is
the
current
connectivity
requirements
in
rural
India,
and
we
have
done
this
that
that
we
have
seen
that
less
than
so
it
is,
it
is
only
if
you
look
into
the
pie
chart
we
have
done
the
calculation
based
on
the
population,
size
of
the
villages
and
contention.
Ratio
of
1
is
to
25
throughput
requirement
of
the
gram.
K
Panchayats
can
be
calculated
as
such,
and
you
can
see
that
that
many
of
the
gram
panchayats
most
of
it
is
like
they
require
less
than
40,
plus
40
Mbps
throughput,
and
we
are
taking
only
those
villagers.
We
we,
the
village
administration,
should
look
into
those
villagers
where
it
is
less
than
20
Mbps
bandwidth,
because
that
that
comes
into
the
that
comes
into
the
financial
scheme
of
the
of
the
village
administration
and
explain
it
to
you
a
little
later.
K
So
we
are
concentrating
only
on
those
villages
that
have
less
than
20
Mbps
bandwidth
requirement
and
the
throughput
requirement,
and
actually
these
these
20.
These
villagers
are
completely
remote
villagers.
Some
of
these
are
like
tribal
blocks
or
tribal
villages;
they
they
don't
even
have
a
road
to
travel,
to
reach
there.
The
next
slide,
so
so.
The
architecture
that
is,
that
is
that
we
have
that
we
are
using
here,
is
that
the
the
we
take
two
options
here.
K
If
you
look
into
the
left
hand
side,
we
look
into
an
option
of
telecom
towers
that
are
at
a
at
a
location
of
5
kilometers
in
the
vicinity
of
the
villages
or
a
connected
GP.
So,
as
you
see
that
the
government
has
already
enabled
the
first
phase
of
connecting
the
gram
panchayats,
so
we
take
two
options.
K
A
K
A
We
can,
can
you
just
try
and
speak
up
a
little
yeah.
K
So
why
do
these
remote
villages
need
the
connectivity?
These
villages
are
completely
unserved
there
are,
there
are
no.
So
there
is
a
presence
of
telecom
tower
private
telecom
tower
or
private
telecom
operators
over
there,
like,
like
Reliance,
Geo,
Vodafone
idea
at
L
and
others,
but
they
don't
serve
the
population
over
there.
This
is
mainly
because
they
don't
because
they,
because
they
don't
get
the
annual
revenue
per
user
back
from
the
subscribers.
K
So
what
happens?
Is
that
the
official
work?
All
the
official
work
is
taken
to
the
cyber
cafe
in
the
city
and
which
which
which
a
lot
of
time
and
money
is
spend
because
because
they
take
you
to
the
city
and
they
don't
get
a
reimbursement
of
it,
so
the
village
administration
actually
spends
the
money
out
of
their
own
pocket,
so
the
quality
of
work
gets
affected.
Now
when
the
quality
of
work
gets
affected.
K
So
some
of
the
villagers-
don't
even
have
you
know,
don't
even
have
their
I
cards
like
the
other
card,
which
is
a
which
is
a
mandatory
thing.
They
don't
even
have
a
voter
card
and
they
cannot
so
they
themselves
go
to
the
villages.
So
the
governance
services
that
the
government
has
actually
framed
for
the
villagers
made
for
the
villagers
over
there
in
these
remote
areas
cannot
be
accessed
by
them.
The
next
slide.
K
So
so
why
the
village
administration?
Now
the
village
administration
is
a
necessity,
so
is
a
very
necessary
partner
in
this
model
for
taking
the
connectivity
to
the
remote
and
unserved
locations,
and
it
is
an
integral
part
of
seeding,
the
growth
of
community
network
in
remote
in
these
remote
villages,
so
the
growth,
otherwise,
the
growth
of
community
network
in
these
villages
is
not
is
not
going
to
is
not
going
to
be
that
easy.
K
If
the
village
administration
doesn't
take
responsibility
of
the
network
and
to
take
responsibility
of
the
network,
they
really
need
to
now
there
and
that
they
really
need
to
own
the
network.
So
this
is
one
of
the
reasons
why
the
village
administration
needs
to
be
an
integral
or
important
stakeholder
in
this.
K
In
this
approach,
the
also
the
sustainability
of
the
network,
the
sustainability
of
the
network,
can
only
be
managed
because,
if
the
village
administration
take
the
responsibility
because
the
village
administration
has
the
authority
and
the
finance
they
have
the
capability
to
finance
for
the
network
in
these
villages.
So
so
we
have
worked
on
work
done
with
them
and
the
community.
K
Also,
when
the
village
administration
takes
a
responsibility,
the
community
also
feels
very
secured,
and
they
feel
that,
yes,
they
should
also
be
a
part
of
the
of
the
community
of
the
network
and,
of
course,
the
security
and
longevity
of
the
devices
in
these
villages
where
the
administration
takes
over
the
network
is,
is
for
a
longer
duration.
It's
long
I
mean
there
is.
There
are
no
devices
that
are
stolen,
people,
take
care
of
the
devices
and
things
of
that
sort.
The
next
slide,
so
how
the
village
administration
can
be
involved
in
the
connectivity
process.
K
So
we
work
together
with
the
village
administration.
We
are
currently
in
the
process
of
developing
this
and
developing.
This
is
a
policy
intervention
that
connectivity
is
needed
both
by
the
villagers
as
well
as
the
gram
panchayat
office.
The
village
administration
there
is
a
need
for
low
high
tech
towers
at
the
gram
panchayat
office,
the
village
administration
15
metres
less
than
15
metres
and
use
of
alternate
power
supplies
such
as
solar
panel
should
be
employed.
So
these
are.
These
are
places
where
the
village
administration
can
fund
for
these
from
their
from
their
fund.
K
That
comes
from
the
state
from
the
state
government
and
pay
for
2
Mbps
bandwidth
monthly
from
the
local
ISP.
The
local
ISP
is
ready.
We
have
channelized
the
local
ISP
and
we
have
got
them
in
touch
with
the
village
administration
where
they
can
pay
for
the
the
village
administration
can
pay
for
the
bandwidth,
so
the
next
slide.
K
This
is
a
partnership
model
that
has
been
developed
by
us.
We
have
identified
the
important
partners
and
panchayat-
that
is
a
village
administration,
is
an
important
place,
an
important
role
in
this
partnership
model,
and
this
is
the
reason
why,
based
on
this
model,
we
have
actually
discussed
with
the
village
administration
to
start
funding
for
the
connectivity
in
their
in
their
own
villages.
So
this
is
the
public-private
panchayat
partnership
model
that
we
have
develop,
and
this
is
a
sustainable
model
on
field
which
has
been
validated
on
field
and
it
is
a
sustainable
model.
K
The
next
slide
now
each
of
these
villages
that
each
of
the
villages
in
India
have
a
5-year
financial
plan
that
they
need
to
put
it
up
to
the
village,
to
the
district
administration
or
the
state
administration
and
fines
start
flowing
from
the
state
administration.
In
this
we
have
identified
that
there
is
a
flaw
that
there
are
streetlights
putting
up
streetlights
water
taps
road
connectivity,
everything,
but
there
is
more
internet
for
development
now
internet
for
development
if
it
is
included
in
the
in
this
financial
plan.
Most
of
this
is
a
five
year
financial
plan.
K
Most
of
the
most
of
the
connectivity
can
be
taken
up
by
the
village
administration
and
it
can
be
funded
by
the
village
administration
and
owned
by
the
village
administration.
The
next
slide,
so
internet
for
development
needs
to
be
included,
and
we
have
worked
out
on
our
government
expenditure
because
the
government
to
show
it
to
the
government
we
really
want
to
to
include
so
only
the
first
year.
There
is
going
to
be
a
capex
involved
with
the
bandages
for
six
months
and
from
next
year
from
the
next
year
onward.
K
It
is
going
to
be
very
minuscule
amount
of
optics
that
needs
to
be
generated,
that
that
can
be
that
the
village
administration
needs
to
pay.
So
only
the
first
year
is
going
to
take
a
chunk
of
the
money
for
putting
up
the
tower
with
solar
batteries,
solar
panels
and
solar
battery
backup
and
and
based
on
this,
we
can.
K
We
have
also
worked
out
on
a
per-user
village
cost,
so
it
is
roughly
around
120
to
150
INR,
that
is
around
one
and
a
half
one
and
a
half
dollars
roughly
around
one
and
a
half
dollars
per
user
per
village
cost.
So
if,
if
the
village
wants
that
the
optics
has
to
be
generated
from
the
village
by
the
utilization
of
the
services
in
the
village,
then
they
can
charge
the
people
around
the
one
and
a
half
dollars
and
generate
the
operational
expenses.
We
are
also
using.
K
Now
we
have,
we
have
done,
we
have
deployed,
we
have
deployed
our
connectivity
in
in
twenty-five
villages,
and
we
have
tested
the
sustainable
model
in
in
two
different
parts.
In
fifteen
villages
we
have
tested
the
model
that
the
connectivity
reaches,
the
village
administration
and
from
there
the
local
ISP
uses
a
marketing
strategy
by
which
he
sells
the
bandwidth
inside
the
village,
and
we
have
done
a
cost-benefit
analysis
of
it,
and
you
can
see
that
there,
in
the
cost-benefit
analysis
that
actually
it
is
it
is
it
is
it
is.
K
It
is
sustainable
from
from
here
to
so
from
from
your
to
the,
and
you
can
see
the
blue
line
that
the
bandwidth,
the
bandwidth,
increases,
just
roaming,
the
bandwidth
increases
and
the
need
for
bandwidth
increases,
and
this
this
actually
is
being
being
monitored
by
the
by
the
local
ISP
and
the
local
ISP
starts,
sells
the
bandwidth
and
as
soon
as
the
local
ISP
starts
to
earn
a
profit.
This
is
where
the
data
that
we
already
have
from
the
field
live
data,
so
as
soon
as
the
sustainability
as
soon
as
the
local
ISP
attain
sustainability.
K
The
return
on
investment
is
there.
There
is
an
investment
again
reinvestment
again
by
the
local
ISP
and
that's
how
so.
The
blue
line
says
that
the
bandwidth
that
is
required
and
the
ROI
is
scaled
up
in
the
next
slide.
If
you
can
see
the
next
slide,
we
have
seen
that.
How
can
the
sustainability?
We
have
done
a
predictive
modeling
for
five
years,
and
we
see
that
the
sustainable
the
the
model
can
be.
The
return
on
investment
is,
is
sustainable.
K
It
keeps
on
increasing
year
after
year
and
as
the
sustainability
in
as
a
return
on
investment
increases
as
a
steady
return
on
investment.
The
bandwidth
of
the
bandwidth
available
are
made
available
by
the
by
the
local
ISP
also
increases,
so
that
he
makes
a
profit
and
then
he
makes
he
gets
in
more
bandwidth.
Now,
if
you
look
into
your
for
almost
their
stable,
the
bandwidth
is
stable
over
there,
because
most
of
the
users
in
these
villages
become
returning
users.
That
means
they
are
fixed
users,
paying
paying
paying
a
particular
amount
every
month.
K
Next
slide
in
one
of
in
the
other
model
that
we
have
looked
is
the
is
the
is
a
really
the
village
level
entrepreneur.
So
a
local
youth
in
the
village
is
nominated
as
an
entrepreneur,
and
then
the
local
youth
sells
the
bandwidth
in
the
village
and
them,
and
the
bandwidth
is
purchased
in
bulk
by
CSC
II
governance,
India
service,
that
just
now
shanty
was
talking
about
so
here
you
can
see
that
the
bulk
bandwidth
that
is
purchased
is
30
Mbps
of
bulk
bandwidth
and
how
the
ROI
is
so
again
from
year.
2
it
it
increases.
K
There
is
a
steady
return
on
investment
and
bandwidth
increases,
and
the
bandwidth
is
made
available
accordingly,
a
five
year
predictive
model
in
the
next
slide
says
that
you
can
see
that
it
is
how
the
sustainability,
the
how
the
model
is
sustainable
from
year
2,
so
it
is
actually
negative
till
year,
but
from
year
2
onwards
it
actually
catches
up,
and
there
is
the
bandwidth
requirement
and
the
availability
is
almost
at
par
with
each
other.
This
is
this
is
what
this
is.
This
is
my
all
about
my
presentation.
Thank
you.
A
A
A
Had
to
ask
for
privacy
reasons:
mom,
oh,
it's
fine,
thank
you
and
sure
Bunny's
a
good
partner
of
ours
in
the
field,
and
we
are
very
keen
to
see
how
things
are
going
with
her.
Thank
you
very
much
and,
as
you
can
see,
we've
son
we've
done
a
reverse
sunchaser
event.
We
started
off
in
Arizona
and
then
came
back
here
to
Thailand.
We've
now
hit
we're
out
for
Pakistan.
So
Arzak
are
you
with
us.
M
My
name
is
Rosa
and
I'm,
founder
and
director
at
internet
policy
of
the
military
Pakistan.
We
were
founded
in
2014
to
fulfill
the
gap
between
academia,
policy
makers,
advocacy
groups
and
the
telephone
industry.
We
are
conducting
policy.
Research
and
ICT
is
in
collaboration
with
some
of
the
leading
institutes
closely,
and
we
are
also
playing
an
active
role
on
issues
like
net
neutrality
in
Pakistan,
the
recent
cyber
crime
legislation
and
increasing
access
which
we
have
taken
up
greatly
along
with
providing
tactical
operations
during
disasters.
M
M
We
started
by
conducting
joint
policy
research
on
the
impact
of
ICTs
and
developing
countries
like
Pakistan
and,
more
precisely,
what
is
the
use
of
response
on
issues
like
net
neutrality,
cybercrime
and
recently,
cyber
security?
We
are
also
working
on
increasing
access
to
the
Internet
in
communities
that
have
been
unserved
remotely
served
by
telephones.
Can
you
move
on
to
the
next.
M
Now
it
close,
the
digital
gap
is
one
of
the
initiative
at
the
observatory,
where
we
aim
to
provide
Internet
connectivity
to
remote
communities
early
Pakistan,
as
you
are
well
aware,
that
importance
of
connected
citizen
is
very
important
for
the
information
society
as
a
tool
for
human,
developing
and
empowerment.
I
cities
have
no
equal,
and
this
is
what
is
the
aim
of
our
project
that
is
close.
The
digital
gap?
Can
you
move
on
to
next,
so
how
it
started?
M
M
This
is
the
map
of
Balochistan
with
inspark
stands
largest
province
in
the
northwest
bordering
the
areas
of
Afghanistan
Iran,
and
there
you
can
see
it.
It
is
one
of
the
most
challenging
areas
and
the
country,
but
dispersed
populations
and
making
the
telecourse
business
case
a
very
challenging
one,
especially
when
it
comes
to
deployment
of
internet
3G.
4G
solution.
Can
we
move
on
to
the
next.
M
The
terrain
is
very
challenging
as
well,
and
it's
very
difficult
to
access
without
many
highways.
The
security
situation
is
volatile,
especially
with
the
local
insurgency
and
then
sentence
he
with
the
hovering
years
of
honest
on
and
tokenization.
It
was
extremely
challenging
project
for
the
object,
victory,
especially
tackling
these
challenges
of
existing
remote
areas,
and
one
of
the
other
challenge
that
we
faced
during
this
project
was
electricity.
M
As
you
can
see,
this
is
a
graph
being
made
after
conducting
a
study
on
the
availability
of
electricity
in
most
part
of
the
province,
since
most
of
the
developing
countries
are
faced
with
this
dilemma.
Tsuki
Shortridge,
we
in
Balochistan
that
we
wanted
to
initiate
this
close.
The
develop
and
close
the
digital
cap
project
also
faced
this
issue
where
there
was
no
electricity
for
more
than
20
to
22,
always
and
say,
as
you
can
see
in
the
graph
majority
or
four
years,
have
very
limited
access
to
electricity.
M
M
Internet
guitars
are
basically
houses
made
of
straw
and
mud
and
get
down
as
a
poetry.
Word
Bellucci
is
a
local
man
which
here
and
because
we
do
have
a
lot
of
these
design
houses
and
remote
areas
and
remote
villages
where
people
come
for
tea.
They
come
forward
to
discuss
their
local
issues
that
come
to
access
issues
regarding
identity,
citizenship
and
other
e-government
services.
So
we
decided
to
come
up
with
a
different
design.
Can
you
move
on
to
the
exactly?
M
The
idea
was
to
come
up
with
a
different
way
of
thinking
and
doing
things
instead
of
going
and
falling
of
the
traditional,
you
know
going
thing
that
was
to
push
yourself
for
rolling
out
the
fiber
networks
and
punish
those
10
equals
this
one
unable
to
serve
in
their
areas.
But
we
were
witnessing
that
this
issue
was
limiting
on
for
quite
a
while,
and
there
was
very
little
change
in
the
community,
especially
when
it
came
to
internet
access,
broadband
spirit,
work
suffering
and
the
quality
of
conferred
access
was
very
poor
as
well.
M
So
it
won't
be
a
surprise
if
you
see
being
going
off
during
the
presentation
there,
so
our
team
decided
to
come
up
with
different
proposal
and
the
next
can
you
move
on
to
the
next
act?
Is
one
of
the
team
members
says
instead
of
mounting
of
easier
on
a
traditional
version,
you
might
have
seen
a
better
version
of
this
in
silence
since
the
water
yeah.
M
The
idea
was
to
roll
out
fiber
networks,
but
rolling
out
fiber
network,
as
we
have
seen,
is
very
costly
I,
given
the
right
to
accessories
and
pathways
and
insecure
information
from
to
Municipal
Committee.
So
we
were
balancing
out
various
technologies
and
tools
we
can
use
for
providing
access
to
that
answer
back.
Can
you
move
on
to
the
next.
M
It
has
to
be
accessible
since
most
of
the
population
that
we
are
dealing
do
not
have
the
computer
literacy
in
English
language,
then,
which
was
one
of
the
barriers
goals
and
the
tools
and
services
how
we
can
train
them
on
using
those
tools
and
services
and,
most
importantly,
what
is
the
community
participation?
It
was
very
important
for
us
that
the
community
that
you
are
working
for,
participated
and
interacted
and
created
content
in
the
local
language
and
ensure
that
these
were
in
sustainable.
M
Otherwise,
we
have
seen
that
from
MTV
Networks
web
committees
do
not
participate
or
take
ownership
usually
dine
out
with
the
passage
of
time.
So
we
approached
the
local
community.
In
most
of
the
areas
there
was
no
internet
and
clean.
With
the
help
of
civil
administration,
we
were
able
to
secure
buildings
that
were
left
out.
Various
government
offices
seizing
cooperation.
Those
areas
we
started
connecting
with
donors
in
Paris
comm,
who
could
donate
us
with
computer
equipments,
laptop
mobile
phones
or
any
other
connectivity
devices
that
you
could
use
for
setting
up
internet
in
those.
M
M
M
So
we
have
a
consistent
supply
of
electricity
and
we
mounted
this
buildings
with
research
collections
and
from
reset
lengths
be
furthered.
Our
network
using
outdoor
Wi-Fi
to
the
community
that
was
residence,
but
initially
the
idea
was
to
set
up
the
committee
arms
and
provide
access
and
training
to
Internet
to
these
communities.
M
But
with
the
passage
of
time,
you
observed
the
female
population
which
which
do
have
limitations
and
travel
to
public
spaces
in
these
areas,
like
it's
some
part
of
India
as
well,
and
because
of
the
split
through
traditional
tribal
culture
and
religious
norms
being
said,
they
were
unable
to
eat
Chinese
areas.
So
we
decided
that
why
do
not?
We
donate
them
tabs
and
Internet
devices
and
provide
outdoor
Wi-Fi
coverage
to
the
villages
by
extending
the
fiber
networks
that
we'll
be
able
to
provide
them
with
internet
access
at
their
homes?
M
This
is
basically
a
TVP.
We
said
receiver
that
keep
ground
along
with
a
computer
device.
Oh
I
know
that
use
this
devices.
They
can
come
and
switch
it
on,
because
it's
not
operational
24/7
of
the
day.
We
keep
this
operational
from
9:00
in
the
morning
till
5:00
in
the
evening,
depending
upon
the
usage.
If
the
community
is
willing
to
use
like
in
some
of
the
places
they
use
it
till
date
night
as
well,
can
you
move
on
to
the
next?
M
This
is
an
outdoor
research
terminal
that
will
be
deployed
at
these
local
communities
to
connect
to
the
internet,
since
we
do
not
have
any
fiber
networks
or
access
to
telecourse,
as
you
can
see
in
the
next
life
that
most
animals
that
were
offering
voice
of
whatever
data
services
decided
to
wind
up
their
operations
in
these
parts
of
the
provinces,
and
you
all
we
can
see.
Is
these
left
out
towers
that
are
lying
there
without
any
access
worship
of
your
services?
So
we
will
try
to
contact
the
owners
of
these
star
was
maybe
this.
M
What
leads
us
this
towers?
So
we
could
deploy
radio
networks.
Some
of
the
operators
were
hyping.
Us
they
offered
the
use
of
tower
free
of
cost,
while
at
Urbana
let
them
to
give
it
to
us.
Even
though
you
can
see
the
picture
and
the
clear
to
mount
our
own
tower,
while
there
was
a
rotational
tower
antenna,
can
you
move
off
in
mixed
up.
M
In
the
past
two
years,
we
have
achieved
a
lot
in
terms
of
providing
internet
access.
We
established
26
internet
key
dhansak
originally
to
star
which
is
almost
with.
This
is
different
districts,
the
province.
We
also
provide
quantumly
internal
training
programs,
and
these
get
downs
across
forest
on
both
the
team
members
can
participate
and,
most
importantly,
access
to
Internet
is
free
for
all
the
community
users.
We
cannot
charge
them
for
getting
access
to
this
connector
blow.
M
What
we
are
doing
is
we
are
collecting
funds
from
donors
who
are
funding
us
in
the
form
of
providing
internet
access
by
providing
this
footprint,
beauty
cell,
which
is
one
of
the
largest
eco,
providing
the
Pakistan,
has
very
generously
provided
us
with
Internet
bandwidth
and
importantly
with
the
use
of
outdoor
Wi-Fi
is
connected
to
the
fiber
network.
We
have
seen
the
increasing
the
fake
of
image
review
users
in
these
communities,
which
is
very
important.
M
M
Baluchistan
is
one
of
the
most
remote
rest
areas
in
Pakistan
and
security
has
been
an
issue
for
the
past
decade
and,
lastly,
the
financial
sustainability
of
running
these
networks.
That
is
the
very
big
challenge.
Yes,
but
can
you
move
on
to
the
next
level
I
think
when
it
comes
to
community
networks,
how
it's
impacting
the
local
communities
can
be
an
important
factor.
But
how
do
you
measure
about
that?
Do
you
measure
it
by
the
streets
they
cared
or
do
you
measure
it
by
the
number
of
people
getting
connected?
M
They
are
also
using
it
for
trading
local
agriculture
products,
which
is
I,
think
very
important,
and
for
us
to
achieve
this
change
meant
that
we
had
to
loosen
our
border
metrics
on
time
spent
and
interaction.
I
think
the
change
it
was
making
in
society
both
and
I'm
very
important
for
us.
Can
you
move
on
to
the
next
act?
I.
M
M
What
we
adopted
was
three-step
principle
that
we've
got
flexible
to
the
technologies
being
used.
We
were
open
to
experimenting
and
nothing
more
and
we
were
ready
to
embrace
the
failure
as
well.
I
think
it
is
important
that
we
embrace
failure
as
well
as
it's
not
always
going
to
do
successful,
especially
in
such
hard
terrain
areas
and
challenging
environments.
And
can
you
move
on
to
the
rock
slide
in
the
next
one?
M
A
M
M
A
A
L
A
Complicated
very
small
aperture
terminals
for
those
that
may
not
know
what
he
said
is
peace
around
the
world
and
particularly
developing
coaches
Thank
You
Eric,
we're
gonna
skip
over
to
federico
Capuano,
who
is
here
to
talk
about.
We've
been
whisk,
so
federico
you're
up
next,
and
this
has
been
whisked
hackable
network
management
system
for
the
20th
century.
D
Great
it's
great
to
be
here.
Thank
you
very
much
for
inviting
me
to
present
I
will
do
a
very
fast
presentation
about
open
ways.
It's
I
think
this
is
one
of
the
right
places
to
present
up
in
whist,
because
we
we
have
a
lot
of
users
from
emerging
countries
which
are
building
Internet
infrastructure
and
they
need
to
manage
this
infrastructure
this.
They
need
to
monitor
the
system,
update
the
configuration
and
sometimes
the
stuff
is
complicated,
so
open
with
aims
to
make
this
easier.
D
Basically
Chenega
is
a
university
construction
in
Italy
and
it
runs
the
biggest
public
Wi-Fi
networks
in
Italy
run
by
the
they
are
state-funded.
This
ones
and
the
Wi-Fi
is
free,
or
sometimes
there
are
private
companies
funding.
It
I've
been
open
with
core
developers
since
2012,
because
open
whisp
is
used
by
Janaka
to
manage
these
huge
public
Wi-Fi
networks
and
then
I've
been
working
on
net
JSON,
which
some
of
you
may
know.
It's
a
specification
for
data,
interchange
format
or
networking
software
next
slide.
So,
let's,
let's
go
ahead
and
talk
about.
D
What's
what's
open,
we
open
with
with
an
open
source
network
management
system.
It
was
started
as
a
solution
to
manage
public
Wi-Fi,
but
it
evolved
into
something
that
can
manage
different
type
of
networks.
It
aims
to
make
it
easier
for
people
organizations
companies
to
manage
and
deploy
their
network,
and
one
of
them
very
good
advantages
is
that
it
helps
to
keep
the
cost
low.
That's.
D
Ok
great,
so
how
do
we
keep
the
cost
low?
It's
because
it
allows
to
automate
repetitive
actions.
Oh
I'm,
running
out
of
battery
sorry
I.
Did
it
I
forgot
to
connect
the
electricity
anyway?
It
because
automates,
repetitive
actions,
so
you
need
less
time
and
your
human
resources
will
need
less
time
to
do
things,
because
many
of
the
boring
stuff
is
automated
and
it
fully
supports
open
wrt.
So
in
this
slider
there
are
links
which
are
put
there.
So
later
you
can
look
at
some
of
these
links
if
you
don't
have
the
No.
D
Some
of
these
links.
Open
wrt
is
great
because
it
runs
on
a
lot
of
artwork
and
it
runs
also
very
cheap
hardware
and
that's
one
of
the
why
urban
wisp
is
becoming
so
used
in
emerging
countries
because
they
get
this
cheap
hardware
and
they
need
a
solution
to
manage
it.
But
if
you
want
a
more
expensive,
more
full,
fully
featured
controller,
you
need
to
look
at
Cisco
or
other
solutions
which
are
a
lot
more
expensive,
not
only
in
terms
of
licensing
but
also
in
terms
of
the
hardware
that
you
buy.
D
D
So
another
Before,
we
jump
to
do
I
jump
to
show
you
some
screenshots
and
some
of
the
features
I
want
to
underline
some
very
important
features
on
the
technical
side.
For
for
those
of
you
who
are
more
technical,
open
with
it's
really
modular
is
composed
of
different
modules
that
can
be
used
independently
or
together.
It
provides
a
default
default
installation
that
there's
a
let's
call
it
a
script
that
installs
the
system,
which
is
quite
complex.
So
it's
not
for
everyone
yet.
D
M
M
F
A
D
It
has
an
active
and
growing
community
and
is
becoming
popular
in
emerging
countries.
As
I
said.
Okay,
let's
go
ahead
and
now
I
can
show
you
what
open
we
can
do.
You
can
go
ahead
again,
yeah.
So
there's
one
of
the
features
which
is
helps
to
keep
the
cost
low.
Is
this
auto
registration
feature?
There's
a
video
link
there
there's
no
time
now.
You
can
see
this
if
you're
interested
on
YouTube
later
so
what
it?
What
it
does
is.
D
Basically,
we
can
compile
a
default
film
or
image
which
you
just
need
to
connect
to
the
internet
and
it
will
auto
register
into
the
system,
and
you
can
prepare
some
default
configuration
called
templates
which
can
be
also
differentiated
by
the
type
of
film
or
image
that
you
do.
You
could
do
a
mesh
image.
You
could
do
a
public
Wi-Fi
image
and
all
these
images,
which
will
automatically
register
and
download
the
default
configuration
which
is
available
in
the
controller
and
they
will
auto
configure
itself
you.
D
You
have
also
the
advantage
that,
by
keeping
the
configuration
in
the
in
the
controller
in
open
ways,
you
can
update
it
anytime
and
they're
all
the
devices
will,
when
you
update
a
template,
all
the
devices
that
use
that
template
will
update
themselves
so
next
slide.
This
is
how
it
looks
when
the
device
registers,
except
that
the
name
is
set
when
the
device
register
you
will
see
the
MAC
address
as
the
name.
There
is
some
information
about
the
IP,
the
management
IP,
which
is
used
by
open
ways
to
reach,
if
necessary,
the
device.
D
There
are
some
other
information
next
slide,
and
here
you
can
see
this.
For
example,
back-end.
That's
really
important,
because
with
this
solution,
we
we
are
able
to
provide
support
for
multiple
operating
systems,
even
though
now
we
only
fully
support
for
production
usage,
open
wrt,
but
later
I
will
tell
you
something
more
about
this.
Then
you
see
templates
that
are
all
the
templates
that
are
enabled
on
the
device
go
go
ahead
and
there's
also
the
device
as
its
own
configuration
that
you
can
tweak.
D
D
This
is
very
important
because
if
the
UI
doesn't
allow
you
to
configure
a
package
because
of
course
we
cannot
support
all
the
possible
configurations
with
our
limited
resources
in
the
UI,
but
we
can
allow
you
to
write
the
configuration
in
a
format
that
is
then
converted
to
the
format
that
will
be
used
on
the
open
wrt.
This
means
this
is
very
powerful
because
it
means
that
open
which
can
configure
any
package
supported
by
open
wrt.
D
So
it
is
an
network
automation
tool
for
any
software
that
runs
on
open,
wrt
and
in
fact,
it's
being
used
for
many
things,
not
just
networking
mesh
networks,
but
also
some
people
are
using
it
for
sd1
now,
and
this
is
something
that
we
didn't
foresee
coming,
but
it
is
happening
so
the
that's.
Why
I
underlined
this
feature?
This
is
not
the
only
way
to
to
configure
additional
packages
that
are.
D
There
are
also
other
ways
that
make
it
more
easy
for
users,
but
there's
no
time
to
talk
about
that
now,
but
you
need
to
know
that
it
is
possible
to
us
support
for
other
packages.
Yes,
let's
go
ahead,
so
here
you
see
different
colors,
because
it's
an
instance
for
for
a
private
company
that
wanted
a
white
label.
So
that's
why
it's
blue-
and
here
we
have
some
well
the
VPN
automation
feature.
This
is
also
another
nice
thing.
It
is
possible
to
have
the
system
automatically
create
VPN
clients
and
create
SSL
Certificates.
D
Well,
I
call
the
SSL,
it's
basically
x.509
certificates
automatically
for
clients
which
can
be
revoked
if
needed.
So
this
is
also
a
very
nice
automation
that
makes
it
easier
to
connect
clients,
so
devices
can
get
their
own
VPN
or
multiple
VPNs.
So
you
can
have
a
management
VPN.
You
can
have
a
traffic
DPN
if
you
want
to
tune
of
the
traffic
to
some
other
place
and
so
on.
Let's
go
ahead,
so
let
me
open
here
the
same
slides
on
my
screen.
So
I
see
better.
So
yes,
this
is
the
geographic
map,
nothing
very
interesting
here.
D
Well,
you
can
go
to
the
next
slide,
which
is
something
more
interesting,
probably
that
not
many
other
open-source
software
project
have
we
added
the
possibility
to
place
the
device
on
an
indoor
map.
This
is
also
very
useful
for
for
people
that
manage
a
lot
of
access
point
and
devices
in
buildings
and
nothing
else
where
there
is
a
possibility
to
to
specify
the
floor.
So
if
you
have
multiple
floors
and
in
the
future
we
we
will
be
build.
D
A
general
map
with
all
the
device
is
both
outdoor
and
indoor,
but
right
now
you
can
only
see
the
single
device
on
the
on
the
indoor
map.
Okay,
we
can
go
ahead
now
and
another
very
important
future
feature,
especially
for
mesh
networks
and
community
networks
which
use
mesh
routing
protocols.
Is
the
network
topology,
collector
and
visualizer?
So
we
have
this
module
which
collects
data
from
multiple
nodes
and
is
able
to
draw
the
graph
and
show
it.
It
is
also
possible
in
the
next
slide.
D
We
can
see
that
it
is
possible
to
go
back
in
time
and
see
how
was
the
network
at
a
previous
day,
so
we
have
daily
snapshots
which
are
collected
automatically
daily,
and
you
can
see
there
that
on
the
11th
of
September
of
this
year,
two
links
were
down,
but
instead
in
the
previous
slide,
all
the
network
was
up.
So
it's
really.
Yes,
it's
really
useful
to
keep
track
of
what
happens
in
the
network
and
we
with
the
new
feature
coming
next.
D
We
will
also
be
able
to
do
a
lot
of
more
automation
regarding
monitoring
of
these
links.
Yes,
we
can
go
ahead.
Another
important
module,
especially
for
isp
and
and
public
Wi-Fi,
it's
the
free,
reduce
module,
IR
I
added
the
logo
there
just
to
show,
but
it's
not
there
in
the
web
interface.
Let's
go
ahead
here,
just
a
few
screenshots.
If
anybody
is
familiar
with
radios,
this
module
basically
allows
you
to
configure
and
manage
the
three
values
database
or
you
can
define
the
group
checks
the
traffic
limit
for
each
profile
and
things
like
that.
D
We
can
go
ahead
and
here
another
screen
shot
in
which
we
show
accounting
data,
which
this
is
very
used,
ISPs,
to
monitor
how
much
traffic
each
device
is
consuming
or
how
much
traffic
public
Wi-Fi
user
is
consuming
any
phase
the
limit
or
not.
If
we
reduce
accounting
feature
so
the
new
free,
reduce
module
provide
all
these
features
for
free
radius,
and
then
we
have
also
more
stuff
coming
in
the
next
slides.
We
can
see
some
previews
and
some
mock-ups.
These
features
are
not
released,
yet
some
of
them
are
just
mock-ups.
D
Some
of
some
of
these
images
are
actually
some
code
running
not
perfectly,
but
we
will
get
there
so
very
important
tool
for
who
manage
network
is
being
able
to
see
how
the
network
is
performing,
how
much
traffic
is
being
consumed
and
so
on.
So
we
can
go
ahead.
The
round-trip
time
you
can
see
there
and
packet
loss
and
this
system
is
also
extensible.
It's
designed
in
a
way
that
it
is
possible
to
write
new
Python
classes
to
write
more
checks.
So
it's
really
powerful,
but
it's
not
there.
D
D
So
we
we
are
building
a
film
grade
module
which
will
allow
in
the
next
slide.
You
can
see
that
it's
possible
to
upload
firmware
images
and
we
plan
to
offer
an
automated
way
to
upload
fema
images
to
the
system
offered
after
compilation
and
in
the
next
slide.
You
can
see
that
you
can
click,
you
can
select
a
build
and
then
you
can
tell
the
system
to
upgrade
all
the
devices
that
are
related
to
that
fever
automatically.
D
So
we
are
building
basically
a
mast,
few
more
great
feature,
but
it
will
also
allow
to
upgrade
single
devices,
so
experimental
features
that
are
coming
in
the
future.
These
two.
Actually,
we
have
implemented
them
last
year,
but
the
result
is
not
production
ready.
Yet
so
we
have
developed
ubiquity
iOS
configuration
back-end,
which
allows
from
the
same
system
that
you
have
seen
to
configure
the
native
ubiquity
arrow
as
filmer.
That's
important,
because
many
communities
prefer
to
not
alter
the
firmware
of
the
antennas
for
different
reasons.
D
So
so
it
will
be
important
to
learn
this
feature
in
open
whisper
communities,
but
also
for
wireless
internet
service
providers.
Then
there's
raspberry,
which
is
already
supported
with
open
wrt
and
you
can
install
open
TT
on
raspberry
and
you
can
manage
it,
but
we
also
plan
to
support
the
native
filmer
of
the
raspberry,
which
is
also
very
used
in
other
IMT
or
type
of
networks.
So
here
I'm
getting
to
the
to
the
end
of
my
presentation.
I
try
to
keep
it
very
brief.
D
D
D
So
if
you,
if
you
think
this
is
important
for
your
organization,
let's
join
forces
come
and
talk
to
me
because
there's
a
lot
of
activity-
and
we
are
working
on
many
things
and
with
the
help
of
all
these
students
and
volunteers
that
are
helping
us
fixing
a
lot
of
problems
and
also
building
a
lot
of
features
for
the,
and
this
is
all
done
to
help
the
communities
in
the
world
that
are
not
connected
for
different
reasons.
Of
course,
is
but
I
think
it's
not
only
ISPs.
D
It
is
also
because
managing
these
networks
is
not
easy,
it's
not
easy,
they
fail,
they
they
break,
they
get
hacked
because
the
firmware
is
not
upgraded.
Sometimes
some
communities
ask
their
participants
to
participate,
but
it's
really
hard.
Sometimes
it
seems
like
you
need
a
degree
to
to
engineering
degree
to
do
some
of
the
things
we
have
to
make
it
easier
if
we
want
people
to
really
build
in
their
own
networks,
because
otherwise
it's
really
hard
I've
done
this
I've
done
this
many
times.
D
I
helped
to
build
many
nodes
here
in
Italy
network
snows
and
I
see
the
common
people
that
have
this
node
in
their
house
as
they
struggle
to
login
into
the
the
antenna
configuration
and
see
all
this.
This
information
we
ask
them
to
upgrade
their
filmer,
but
is
it's
hard
for
them?
So
we
need
to
provide
tools,
there's
a
lot
of
communities,
Libre
mesh
and
other
communities
which
are
building
these
kind
of
things.
Urban
wisp
is
another
tool
that
focuses
on
on
building
control
panel.
That's
what
it
is.
So.
Thank
you
very
much.
The
last
slide.
D
You
can
see
the
open,
wisp
logo
and
you
can
see
they
link
to
the
website
open
with
the
org.
If
you
want
to
find
more
information
and
my
presentation
and
0,
thank
you
very
much,
and
if
you
have
any
question
here
or
later
online,
you
can
find
me
on
Twitter
I'm,
very
active
there.
You
can
find
me
on
the
mailing
list
of
the
several
projects
I
participate.
So
it's
really
easy
to
reach
me.
If
you
want
Thank.
A
D
A
Great,
so
thank
you
very
much
Federico.
This
was
an
excellent
presentation,
speaks
to
the
heart
of
some
of
the
tech
work,
that's
going
on
in
community
networks.
We
thank
everyone
for
your
attention
and
sorry.
We
ran
out
of
time.
This
was
the
first
time
that
we've
done
99%
of
the
work
online.
So
thank
you
to
everyone
that
was
hugely
aware
of
that
and
synched
up
with
us
remotely.
So
this
was
kind
of
a
big
ask.
A
Thank
you
very
much
and
stay
tuned
Gaia
we're
putting
a
draft
outlined
together
that
we
hope
to
have
with
you
at
the
next
meeting
in
Prague,
but
you'll
see
us
on
than
that
where
we
are
going
to
be
running
this
all
by
you.
So
thank
you
again
and
we
appreciate
you
coming
and
thanks
so
much
the
presenters
who
are
online
with
us,
we're
going
to
sign
off
and
see
you
on
the
net
thanks
a
lot.