Embed Size (px)
Citation preview
![Page 1: E [] A [] B [] D [] E A Mobile Path to Censorship ... · Easy to shut down, censor, and spy on. Can we create a blocking-, censorship-, and surveillance-resistant network that is](https://reader036.pdfslide.us/reader036/viewer/2022071116/5ffc921b4d97a83197499cb1/html5/thumbnails/1.jpg)
A M
obile
Pat
h to
Cen
sors
hip
Res
ista
nce
and
Pri
vacy
Yue
Liu
, D
avid
Bild
, D
avid
Adr
ian,
Gul
shan
Sin
gh, R
ober
t D
ick,
Dan
Wal
lach
, an
d Z. M
orle
y M
ao
Thi
s w
ork
was
sup
por
ted,
in
part
, by
NSF
und
er a
war
d T
C-0
9645
45.
The
Int
erne
t
Ope
ned
cont
ent
acce
ss a
nd c
reat
ion,
Hie
rarc
hica
l.E
asy
to s
hut
dow
n, c
enso
r, a
nd s
py o
n.
Can
we
crea
te a
blo
ckin
g-, ce
nsor
ship
-, a
ndsu
rvei
llanc
e-re
sist
ant
netw
ork
that
is
prac
tica
l fo
r no
rmal
sm
artp
hone
use
rs?
How
wou
ld t
hat
wor
k?
The
Alt
erna
tive
Com
mod
ity
smar
tpho
nes.
App
ins
tall.
Pee
r-to
-pee
r W
iFi.
Fully
dis
trib
uted
arc
hite
ctur
e.
Wha
t ab
out de
lay,
requ
ired
ado
ptio
n ra
te, an
dim
plic
atio
ns o
f hu
man
mot
ion
patt
erns
?
Wha
t ab
out re
liabi
lity
in t
he p
rese
nce
ofde
nial
of se
rvic
e at
tack
s,Sy
bil a
ttac
ks, an
dsp
am/p
ropa
gand
a?
1am
mic
robl
oggi
ng a
ppre
leas
ed o
n ca
mpu
s.
02
44
87
29
61
20
14
41
68
19
2E
lap
sed
tim
e i
n h
ou
rs
0.0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1.0
Delivery rate
10
20304050
607
08
09
0
01
23
45
67
km
012345678 km
No
rth
10
0
10
1
10
2
10
3
85%
del
iver
y ra
te
13 h
our
late
ncy
0.3%
ado
ptio
n ra
te
Con
tact
hea
t m
ap
0.0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1.0
Ra
tio
of
rem
ov
ed
re
sou
rce
s
0.0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1.0
Delivery rate
Ra
nd
om
no
de
s
Targ
ete
d n
od
es
Targ
ete
d l
oca
tio
ns
Inte
rne
t
Wor
ks w
ith
mos
tim
port
ant
half
of p
arti
cipa
nts
disa
bled
.
Pho
ne-t
o-ph
one
cont
acts
hap
pen
thro
ugho
ut c
ampu
s.
Tw
ice
the
obse
rved
con
tact
rate
of pr
ior
user
stu
dies
.
Sybi
l at
tack
: A
sin
gle
phys
ical
ent
ity
can
pret
end
to b
em
ulti
ple
part
icip
ants
, ga
inin
g un
fair
infl
uenc
e at
low
cos
t.
A
B C
DE
[...]
A
[...]
D
[...]
E
[...]
C
[...]
B
[...]
D
[...]
E
[...]
C
[...]
A
[...]
D[..
.] E
[...]
B
[...]
A
[...]
C
[...]
D
[...]
B
[...]
A
[...]
C
[...]
E
[...]
B
A
B C
DE
B: 1
9
A: 2
5B
: 18
A: 1
2B
: 17
A: 2
0
A: 3
1B
: 27
A
B C
DE
A: 1
2
A: 2
5
A: 3
1
A: 2
0
B: 1
9
B: 1
8
B: 1
7
B: 2
7
A
I
B
S2
S1
I
A
B
MAA
S1
S2
Iden
titi
es t
rans
mit
and
rec
ord
sign
al s
tren
gths
.E
xcha
nge
obse
rvat
ions
.N
odes
for
m int
erna
lly c
onsi
sten
t w
orld
vie
ws.
Tes
ted
on H
TC
Mag
ic s
mar
tpho
nes
wit
h 11
use
rs.
Cha
lleng
es: C
an't
cha
nge
hard
war
e.N
o tr
uste
d au
thor
itie
s.
Par
tici
pant
s m
ove.
Insi
ght:
Can
det
erm
ine
true
wor
ldvi
eww
itho
ut t
rust
ing
any
part
icip
ant.
0
0.2
0.4
0.6
0.81
00.
20.
40.
60.
81
True Positive Rate (Sensitivity)
Fal
se P
ositi
ve R
ate
(1 -
Spe
cific
ity)
Offi
ce I
Offi
ce II
Caf
eter
iaO
utdo
or
0.9
0.92
0.94
0.96
0.981
00.
020.
040.
060.
080.
1
Elim
inat
es 9
1%-1
00%
of
Sybi
l id
enti
ties
,de
pend
ing
on e
nvir
onm
ent.
Wha
t ab
out sp
am/p
ropa
gand
ain
thi
s fu
lly d
ecen
tral
ized
net
wor
k?
No
cent
ral au
thor
ity.
Wha
t ca
n be
don
e w
ith
loca
l in
form
atio
n?
Use
tra
nsit
ive
rebr
oadc
ast
grap
hin
stea
d of
fol
low
ers
grap
h.
Les
s di
sass
orti
ve a
nd m
ore
clus
tere
d,lik
e no
n-di
gita
l so
cial
net
wor
ks.
Use
gra
ph s
truc
ture
to fi
nd s
pam
mer
s.
00.
10.
20.
30.
40.
50.
60.
70.
80.
91
00.
20.
40.
60.
81
True Positive Rate
Fal
se P
ositi
ve R
ate0.
0000
2
0.00
006
0.00
014
0.00
030
0.00
060
0.00
300
