Hac k- a- Vote : Studying Se c ur ity Issue s with E - Voting Da - - PowerPoint PPT Presentation

▶

Apr 27, 2023 274 likes •546 views

Hac k- a- Vote : Studying Se c ur ity Issue s with E - Voting Da n Wa lla c h Ric e Unive rsity Collabor ator s: Jo na tha n Ba nne t Da vid W. Pric e Alg is Rudys Justin Sing e r Pe r c e ption vs. r e ality Pe r c e ption vs. r

SLIDE 1

Hac k- a- Vote : Studying Se c ur ity Issue s with E

Voting

Da n Wa lla c h Ric e Unive rsity

Collabor ator s:

Jo na tha n Ba nne t Da vid W. Pric e Alg is Rudys Justin Sing e r

SLIDE 2

Pe r c e ption vs. r e ality

Vo te r fe e ls tha t

Vo te wa s c o unte d Vo te wa s priva te No b o dy e lse c a n vo te mo re tha n o nc e No b o dy c a n a lte r o the rs’ vo te s

Pe o ple b e lie ve tha t the ma c hine

wo rks c o rre c tly

T

he se ha ve to do with pe rc e ptio n

It is also impor tant that the se pe r c e ptions ar e tr ue .

Pe r c e ption vs. r e ality

SLIDE 3

Re lianc e on c e r tific ation

I nde pe nde nt T e sting Autho ritie s

Allo we d to se e the c o de

No b o dy e lse lo o ks

Ce rtify sa tisfa c tio n o f F

E C sta nda rds

Re q uire d b y ma ny sta te s

Re sult: “F aith- base d voting”

SLIDE 4

Inspir ation

Ha ve a n e -vo ting syste m to “de mo nstra te ” inside r fla ws

Orig ina l ide a fro m Da vid Dill Orig ina l c o de b y Da vid W. Pric e Writte n summe r 2003 Ab o ut 2000 line s o f Ja va

Unne c e ssa ry a fte r Die b o ld finding s

SLIDE 5

Se c ond applic ation?

Ho w a b o ut in-c la ss use ?
Old pro je c t: “sma rt c a rd so da

ma c hine ”

1) de sig n & fo rma lly mo de l c rypto pro to c o l 2) swa p with o the r g ro ups 3) imple me nt with re a l c a rds

Re a l sma rt c a rds a re pa inful

SLIDE 6

Hac k- a- Vote pr

je c t

Re mo ve “c he a ting ” c o de

~150 line s, mo stly in o ne file

T hre e pha se a ssig nme nt

1) Be e vil (2 we e ks) 2) Be a n I T A (1 we e k) 3) De sig n / fo rma lly mo de l b e tte r ve rsio n o f Die b o ld sma rtc a rd (2.5 we e ks)

SLIDE 7

Be e vil?

Stude nts’ ro le : c o rrupt de ve lo pe r

inside ve ndo r

Co de must still pa ss te sts “Minima l” c o de c ha ng e s

Multiple ha c ks e nc o ura g e d

Co de sho uld a ppe a r “no rma l”

De live rable s: Co de + Writte n Re po rt

SLIDE 8

Be an IT A?

Swa p c o de fro m g ro ups E

ve ry g ro up a udits two ve rsio ns

Ho no r c o de : no running diff

I

mpe rfe c t simula tio n o f re a l I T As

Stude nt fa milia rity with c o de Sma lle r c o de b a se

De live rable s: Writte n Re po rt

SLIDE 9

Be tte r smar tc ar d pr

toc ols?

L

e c ture s ha ve pre pa re d stude nts

c r

yptyc fo r pro to c o l mo de lling

(Re la tive ly) usa b le type c he c ke r

c ryptyc .c s.de pa ul.e du

De live rable s: Mo de l + Writte n Re po rt

SLIDE 10

Die bold’s smar t c ar d pr

toc ol

My passwor d is (8 byte s)

T e rmina l Ca rd

“Okay” Ar e you valid? “Yup” Canc e l your se lf, ple ase . “Okay”

SLIDE 11

Hac k- a- Vote softwar e

I nspira tio n: Ha rt I nte rCivic e Sla te

SLIDE 12

e Slate pr

toc ol (hope fully)

Base station Voting mac hine

…

ne twor k

Pin: 1234 Pin: 1234 Valid? 1234

SLIDE 13

Hac k- a- Vote live de mo

SLIDE 14

Hac k- a- Vote de sign

SLIDE 15

Hac k- a- Vote de sign

SLIDE 16

Hac k- a- Vote de sign

SLIDE 17

Hac k- a- Vote de sign

SLIDE 18

Wide gamut of attac ks

Ma nipula te e le c tio n re sults Vio la te vo te r a no nymity Cra sh / Do S vo ting ma c hine

SLIDE 19

Cle ve r hac ks

Ove rlo a d equals() / hashCode() Va ria b le with sa me na me a s c la ss

Unusua l c o ntro l flo ws

Re use c o nsta nts in the c o de

Ne two rk po rt: 1776 Use a s b a c kdo o r PI

N

“Sta rt o ve r” a lso sub mits a vo te

SLIDE 20

De e pe r hac ks

We a k ra ndo m numb e r g e ne ra to r

E

a sie r to g ue ss va lid PI Ns

RNG fo r vo te shuffle se e de d with

te rmina l I D

Atta c ke r c a n undo shuffle

Only c he a t if te rmina l I

D > 2

L

e ss like ly to o c c ur in te sting

SLIDE 21

Did the IT As c atc h the hac ks?

Ha c k Atte mpts F

und
nc e

twic e

Mo dify a lre a dy- c a st vo te s

6 7 4 4

Ca st multiple vo te s

5 6 7

Vio la te vo te r a no nymity

3 6 2 3 2

De nia l o f se rvic e

SLIDE 22

Implic ations for r e al IT As

Ca n re a l I

T As do b e tte r?

+ T he y c a n run diff + T he y c a n pe rfo rm “pa ra lle l te sting ” – Co de b a se s a re muc h la rg e r – Are the y e xpe c ting T ro ja n Ho rse s? – Ho w c lo se ly do the y re a d the c o de ?

Ve ry little suppo rt fro m to o ls

SLIDE 23

Uglie r issue s for c e r tific ation

T

o lc ha in ta mpe ring (T

ho mpso n)

T

a mpe ring with “e mb e dde d” OS

Audite d c o de = a c tua l c o de in

ma c hine ?

SLIDE 24

Public ity

I E E E S e c urity & Privac y, Jan/ F e b 2004

Re printe d in Co mpute r Use r Sto ry o n lo c a l T

V ne ws

I

mpa c t o n ve ndo rs / I T As?

SLIDE 25