CS 134 Winter 2018 LECTURE 15 Privacy and Anonymitysconce.ics.uci.edu/134-W18/slides/LEC15.pdf ·...

31
LECTURE 15 Privacy and Anonymity CS 134 Winter 2018 1

Transcript of CS 134 Winter 2018 LECTURE 15 Privacy and Anonymitysconce.ics.uci.edu/134-W18/slides/LEC15.pdf ·...

Page 1: CS 134 Winter 2018 LECTURE 15 Privacy and Anonymitysconce.ics.uci.edu/134-W18/slides/LEC15.pdf · • Encryption (e.g., SSL or IPSec) does not hide identities ... • 1.5 million

LECTURE 15 Privacy and Anonymity

CS 134 Winter 2018

1

Page 2: CS 134 Winter 2018 LECTURE 15 Privacy and Anonymitysconce.ics.uci.edu/134-W18/slides/LEC15.pdf · • Encryption (e.g., SSL or IPSec) does not hide identities ... • 1.5 million

Privacy

• Privacy and Society •  Basic individual right & desire •  Also relevant to corporations & government

agencies •  Recently increased awareness

•  But, public’s perception of privacy is fickle

• Privacy and Technology in Recent Years •  >>Information disclosed on the Internet •  >>Handling and transfer of sensitive

information •  <<Privacy and accountability

2

(Image from geekologie.com)

Page 3: CS 134 Winter 2018 LECTURE 15 Privacy and Anonymitysconce.ics.uci.edu/134-W18/slides/LEC15.pdf · • Encryption (e.g., SSL or IPSec) does not hide identities ... • 1.5 million

Privacy on Public Networks

•  TheInternetisdesignedasapublicnetwork•  AnyoneonyourLAN(wiredorwireless)canseeyourtraffic•  Networkroutersseealltrafficthatpassesthroughthem

• Routinginformationispublic•  IPpacketheadersidentifysourceanddestinationaddresses•  Apassiveobservercaneasilyfigureoutwhoistalkingtowhom

•  Encryption(e.g.,SSLorIPSec)doesnothideidentities•  Encryptionhidespayload,notrouting+addressinginformation•  EvenIP-levelencryption(tunnel-modeIPsec/ESP)revealsIPaddressesofIPsecgateways

3

Page 4: CS 134 Winter 2018 LECTURE 15 Privacy and Anonymitysconce.ics.uci.edu/134-W18/slides/LEC15.pdf · • Encryption (e.g., SSL or IPSec) does not hide identities ... • 1.5 million

Applications of Anonymity (1)

• Privacy• Hideonlinetransactions,Webbrowsing,etc.fromintrusivegovernments,marketers,archival/searchentities(e.g.,Google)aswellasfromcriminalsandsnoops

• UntraceableElectronicMail•  Corporatewhistle-blowers•  Politicaldissidentsinoppressivesocieties•  Sociallysensitivecommunications(onlineAAorSTDmeeting)•  Confidentialbusinessnegotiations

• LawEnforcementandIntelligence•  Stingoperationsandhoneypots•  Secretcommunicationsonapublicnetwork

•  Informers,secretagents,etc.4

Page 5: CS 134 Winter 2018 LECTURE 15 Privacy and Anonymitysconce.ics.uci.edu/134-W18/slides/LEC15.pdf · • Encryption (e.g., SSL or IPSec) does not hide identities ... • 1.5 million

Applications of Anonymity (2)

• Digital/ElectronicCash•  Electroniccurrencywithpropertiesofpapermoney(onlinepurchasesunlinkabletobuyer’sidentity)

• AnonymousElectronicVoting

• Censorship-ResistantPublishing

• Crypto-Anarchy•  “Somepeoplesaythat“anarchywon'twork.”That'snotanargumentagainstanarchy;that'sanargumentagainstwork.”–BobBlackJ

5

Page 6: CS 134 Winter 2018 LECTURE 15 Privacy and Anonymitysconce.ics.uci.edu/134-W18/slides/LEC15.pdf · • Encryption (e.g., SSL or IPSec) does not hide identities ... • 1.5 million

Applications of Anonymity (3)

• Porn

• HumanTrafficking

•  Libel

• Disinformation=FakeNews/Propaganda

•  SaleofIllegalSubstances(e.g.,SilkRoad)

•  TaxAvoidance(viaUntraceablePayments)

•  IncitementtoCriminalActivity(e.g.,Murder,Rioting,Genocide,Terrorism)6

Page 7: CS 134 Winter 2018 LECTURE 15 Privacy and Anonymitysconce.ics.uci.edu/134-W18/slides/LEC15.pdf · • Encryption (e.g., SSL or IPSec) does not hide identities ... • 1.5 million

What is Anonymity?

• Anonymity:inabilitytoidentifysomeonewithinasetofsubjects(sizevaries)•  DifferentfromPRIVACY–righttobeleftalone•  Tobeanonymous,needtohideyouractivitiesamongsimilaractivitiesbyothers•  Onecannotbeanonymousalone!

•  Bigdifferencebetweenanonymityandconfidentiality

• Unlinkability:separationofactionandidentityperformingthataction•  Forexample,senderandhisemailarenomorerelatedafterobservingcommunicationthantheywerebefore

• Unobservability:inabilitytotellwhetheracertainactiontookplace•  veryhardtoachieve

7

Page 8: CS 134 Winter 2018 LECTURE 15 Privacy and Anonymitysconce.ics.uci.edu/134-W18/slides/LEC15.pdf · • Encryption (e.g., SSL or IPSec) does not hide identities ... • 1.5 million

Attacks on Anonymity

• PassiveTrafficAnalysis•  Inferfromnetworktrafficwhoistalkingtowhom•  Tohideyourtraffic,mustcarryotherpeople’straffic!

• ActiveTrafficAnalysis•  Injectpacketsorputatimingsignatureonapacketflow

• CompromiseofNetworkNodes(suchasRouters)•  Notobviouswhichnodeshavebeencompromised

•  Attackermaybepassivelyloggingtraffic•  Donotfullytrustanyindividualnode

•  Assumethatsomefractionofnodesisgood,butdonotknowwhich

8

Page 9: CS 134 Winter 2018 LECTURE 15 Privacy and Anonymitysconce.ics.uci.edu/134-W18/slides/LEC15.pdf · • Encryption (e.g., SSL or IPSec) does not hide identities ... • 1.5 million

Chaum’s Mix (David Chaum, ca. 1980-1981)

•  Earliestproposalforanonymousemail:•  DavidChaum,“Untraceableelectronicmail,returnaddresses,anddigitalpseudonyms”,CommunicationsoftheACM,February1981.

• Public-keycrypto+trustedre-mailer(Mix)•  Untrustedcommunicationmedium•  Public-keysusedaspersistentpseudonyms

• ModernanonymitysystemsuseMixasthebasicbuildingblock

9

Before spam, people thought anonymous email was a good idea J

Page 10: CS 134 Winter 2018 LECTURE 15 Privacy and Anonymitysconce.ics.uci.edu/134-W18/slides/LEC15.pdf · • Encryption (e.g., SSL or IPSec) does not hide identities ... • 1.5 million

Basic Mix Design

10

A

C

D

E

B

Mix

{r1,{r0,M}pk(B),B}pk(mix) {r0,M}pk(B),B

{r2,{r3,M’}pk(E),E}pk(mix)

{r4,{r5,M’’}pk(B),B}pk(mix)

{r5,M’’}pk(B),B

{r3,M’}pk(E),E

Adversary knows all senders and all receivers, but cannot link a sent message with a received message

Page 11: CS 134 Winter 2018 LECTURE 15 Privacy and Anonymitysconce.ics.uci.edu/134-W18/slides/LEC15.pdf · • Encryption (e.g., SSL or IPSec) does not hide identities ... • 1.5 million

Anonymous Return Addresses

• 11

A

B MIX

{r1,{r0,M}pk(B),B}pk(mix) {r0,M}pk(B),B

M includes {K1,A}pk(mix’), K2 where K2 is a fresh public key and MIX’ is possibly different from MIX

Response MIX’

{K1,A}pk(mix’), {r2,M’}K2 A,{{r2,M’}K2}K1

Secrecy without authentication (good for an online confession service J)

Page 12: CS 134 Winter 2018 LECTURE 15 Privacy and Anonymitysconce.ics.uci.edu/134-W18/slides/LEC15.pdf · • Encryption (e.g., SSL or IPSec) does not hide identities ... • 1.5 million

Mix Cascade

• Messagesaresentthroughasequenceofmixes•  Canalsoformanarbitrarynetworkofmixes(“mixnet”)

•  Somemixesmaybecontrolledbyattacker,butevenasinglegoodmixguaranteessomeanonymity

• Padandbuffertraffictofoilcorrelationattacks12

Page 13: CS 134 Winter 2018 LECTURE 15 Privacy and Anonymitysconce.ics.uci.edu/134-W18/slides/LEC15.pdf · • Encryption (e.g., SSL or IPSec) does not hide identities ... • 1.5 million

Disadvantages of Basic Mixnets

• Public-keyencryptionanddecryptionateachmixarecomputationallyexpensive

• Basicmixnetshavehighlatency•  Okforemail,butnotforanonymousWebbrowsing

• Challenge:low-latencyanonymitynetwork•  Usepublic-keycryptographytoestablisha“circuit”withpairwisesymmetrickeysbetweenhopsonthecircuit

•  Thenusesymmetricdecryptionandre-encryptiontomovedatamessagesalongtheestablishedcircuits

•  Eachnodebehaveslikeamix;anonymityispreservedevenifsomenodesarecompromised

13

Page 14: CS 134 Winter 2018 LECTURE 15 Privacy and Anonymitysconce.ics.uci.edu/134-W18/slides/LEC15.pdf · • Encryption (e.g., SSL or IPSec) does not hide identities ... • 1.5 million

Another Idea: Randomized Routing

• Hidesourcesbyroutingmessagesrandomly•  Populartechnique:Crowds,Freenet,Onionrouting

• Routersdonotknowiftheapparentsourceofamessageisthetruesenderoranotherrouter

14

Page 15: CS 134 Winter 2018 LECTURE 15 Privacy and Anonymitysconce.ics.uci.edu/134-W18/slides/LEC15.pdf · • Encryption (e.g., SSL or IPSec) does not hide identities ... • 1.5 million

Onion Routing

15

R R4

R1 R2

R

R R3

Bob

R

R

R

•  Sender chooses a random sequence of routers •  Some routers are honest, some are not •  Sender controls path length

Alice

[Reed, Syverson, Goldschlag 1997]

Page 16: CS 134 Winter 2018 LECTURE 15 Privacy and Anonymitysconce.ics.uci.edu/134-W18/slides/LEC15.pdf · • Encryption (e.g., SSL or IPSec) does not hide identities ... • 1.5 million

Route Establishment

16

R4

R1

R2 R3 Bob Alice

{R2,k1}pk(R1),{ }k1 {R3,k2}pk(R2),{ }k2

{R4,k3}pk(R3),{ }k3 {B,k4}pk(R4),{ }k4

{M}pk(B)

•  Routing info for each link encrypted with router’s public key •  Each router learns only the identity of the next router

Page 17: CS 134 Winter 2018 LECTURE 15 Privacy and Anonymitysconce.ics.uci.edu/134-W18/slides/LEC15.pdf · • Encryption (e.g., SSL or IPSec) does not hide identities ... • 1.5 million

The Onion Router (Tor)

•  Second-generationonionroutingnetwork•  http://tor.eff.org•  Specificallydesignedforlow-latencyanonymousInternetcommunications(e.g.,Webbrowsing)

•  RunningsinceOctober2003

• Hundredsofnodesonallcontinents

•  1.5millionusersasof2016

•  “Easy-to-use”clientproxy•  Freelyavailable,canuseitforanonymousbrowsing•  Availableforsmartphonesandtabletstoo

17

Page 18: CS 134 Winter 2018 LECTURE 15 Privacy and Anonymitysconce.ics.uci.edu/134-W18/slides/LEC15.pdf · • Encryption (e.g., SSL or IPSec) does not hide identities ... • 1.5 million

Tor Circuit Setup (1)

• ClientproxyestablishesasymmetricsessionkeyandcircuitwithOnionRouter#1

18

Page 19: CS 134 Winter 2018 LECTURE 15 Privacy and Anonymitysconce.ics.uci.edu/134-W18/slides/LEC15.pdf · • Encryption (e.g., SSL or IPSec) does not hide identities ... • 1.5 million

Tor Circuit Setup (2)

• ClientproxyextendsthecircuitbyestablishingasymmetricsessionkeywithOnionRouter#2

•  TunnelthroughOnionRouter#1

19

Page 20: CS 134 Winter 2018 LECTURE 15 Privacy and Anonymitysconce.ics.uci.edu/134-W18/slides/LEC15.pdf · • Encryption (e.g., SSL or IPSec) does not hide identities ... • 1.5 million

Tor Circuit Setup (3)

• ClientproxyextendsthecircuitbyestablishingasymmetricsessionkeywithOnionRouter#3

•  TunnelthroughOnionRouters#1and#2

20

Page 21: CS 134 Winter 2018 LECTURE 15 Privacy and Anonymitysconce.ics.uci.edu/134-W18/slides/LEC15.pdf · • Encryption (e.g., SSL or IPSec) does not hide identities ... • 1.5 million

Using a Tor Circuit

• ClientapplicationsconnectandcommunicateovertheestablishedTorcircuit(alsotomultipledst-s)

•  Datagramsaredecryptedandre-encryptedateachlink

21

Page 22: CS 134 Winter 2018 LECTURE 15 Privacy and Anonymitysconce.ics.uci.edu/134-W18/slides/LEC15.pdf · • Encryption (e.g., SSL or IPSec) does not hide identities ... • 1.5 million

Tor Management Issues

• Manyapplicationscanshareonecircuit•  MultipleTCPstreamsoveroneanonymousconnection

•  Torrouterdonotneedrootprivileges•  Encouragespeopletosetuptheirownrouters•  Moreparticipants=betteranonymityforeveryone

• Directoryservers•  Maintainlistsofactiveonionrouters,theirlocations,currentpublickeys,etc.•  Controlhownewroutersjointhenetwork

•  “Sybilattack”:attackercreatesalargenumberofrouters•  Directoryservers’keysshipwithTorcode

22

Page 23: CS 134 Winter 2018 LECTURE 15 Privacy and Anonymitysconce.ics.uci.edu/134-W18/slides/LEC15.pdf · • Encryption (e.g., SSL or IPSec) does not hide identities ... • 1.5 million

Location Hidden Servers

• Goal:deployaserverontheInternetthatanyonecanconnecttowithoutknowingwhereitisorwhorunsit

• Accessiblefromanywhere

• Resistanttocensorship

• Cansurviveafull-blownDoSattack

• Resistanttophysicalattack•  Cannotfindthephysicalserver!

23

Page 24: CS 134 Winter 2018 LECTURE 15 Privacy and Anonymitysconce.ics.uci.edu/134-W18/slides/LEC15.pdf · • Encryption (e.g., SSL or IPSec) does not hide identities ... • 1.5 million

Creating a Location Hidden Server

24

Server creates circuits to “introduction points”

Server gives intro points’ descriptors and addresses to service lookup directory

Client obtains service descriptor and intro point address from directory

Page 25: CS 134 Winter 2018 LECTURE 15 Privacy and Anonymitysconce.ics.uci.edu/134-W18/slides/LEC15.pdf · • Encryption (e.g., SSL or IPSec) does not hide identities ... • 1.5 million

Using a Location Hidden Server

25

Client creates a circuit to a “rendezvous point”

Client sends address of the rendezvous point and any authorization, if needed, to server through intro point

If server chooses to talk to client, connect to rendezvous point

Rendezvous point matches the circuits from client & server

Page 26: CS 134 Winter 2018 LECTURE 15 Privacy and Anonymitysconce.ics.uci.edu/134-W18/slides/LEC15.pdf · • Encryption (e.g., SSL or IPSec) does not hide identities ... • 1.5 million

Deployed Anonymity Systems

•  FreeHavenprojecthasanexcellentbibliographyonanonymity•  http://www.freehaven.net/anonbib

•  Tor(http://tor.eff.org)•  Overlaycircuit-basedanonymitynetwork•  Bestforlow-latencyapplicationssuchasanonymousWebbrowsing

• Mixminion(http://www.mixminion.net)•  Networkofmixes•  Bestforhigh-latencyapplicationssuchasanonymousemail

26

Page 27: CS 134 Winter 2018 LECTURE 15 Privacy and Anonymitysconce.ics.uci.edu/134-W18/slides/LEC15.pdf · • Encryption (e.g., SSL or IPSec) does not hide identities ... • 1.5 million

Dining Cryptographers

• Howtomakeamessagepublic,butinaperfectlyuntraceablemanner•  DavidChaum.“Thediningcryptographersproblem:unconditionalsenderandrecipientuntraceability.”JournalofCryptology,1988.

• Guaranteesinformation-theoreticanonymityformessagesenders•  VERYstrongformofanonymity:defeatsadversarywhohasunlimitedcomputationalpower

• Difficulttomakepractical•  IngroupofsizeN,needNrandombitstosend1bit

27

Page 28: CS 134 Winter 2018 LECTURE 15 Privacy and Anonymitysconce.ics.uci.edu/134-W18/slides/LEC15.pdf · • Encryption (e.g., SSL or IPSec) does not hide identities ... • 1.5 million

Three-Person DC Protocol

•  Threecryptographersarehavingdinner.

•  EitherNSAispayingforthedinner,oroneofthemispaying,butwishestoremainanonymous.

1.  Eachdinerflipsacoinandshowsittohisleftneighbor.

•  Everydinerseestwocoins:hisownandhisrightneighbor’s

2.  Eachdinerannounceswhetherthetwocoinsarethesame.Ifheisthepayer,helies(saystheopposite).

3.  IFNumberof“same”=1or3⇒NSAispayingIFNumberof“same”=0or2⇒oneofthemispaying

•  Butanon-payercannottellwhichoftheothertwoispaying!28

Page 29: CS 134 Winter 2018 LECTURE 15 Privacy and Anonymitysconce.ics.uci.edu/134-W18/slides/LEC15.pdf · • Encryption (e.g., SSL or IPSec) does not hide identities ... • 1.5 million

Non-Payer’s View: Same Coins

29

?

“same” “different”

payer payer

?

“same” “different”

Without knowing the coin toss between the other two, non-payer cannot tell which of them is lying

Page 30: CS 134 Winter 2018 LECTURE 15 Privacy and Anonymitysconce.ics.uci.edu/134-W18/slides/LEC15.pdf · • Encryption (e.g., SSL or IPSec) does not hide identities ... • 1.5 million

Non-Payer’s View: Different Coins

30

?

“same” “same”

payer payer

?

“same” “same”

Without knowing the coin toss between the other two, non-payer cannot tell which of them is lying

Page 31: CS 134 Winter 2018 LECTURE 15 Privacy and Anonymitysconce.ics.uci.edu/134-W18/slides/LEC15.pdf · • Encryption (e.g., SSL or IPSec) does not hide identities ... • 1.5 million

Super-posed Sending

•  ThisideageneralizestoanygroupofsizeN

•  Foreachbitofthemessage,everyusergenerates1randombitandsendsittoONEneighbor

•  Everyuserlearns2bits(hisownandhisneighbor’s)

•  EachuserannouncesownbitXORneighbor’sbit

•  SenderannouncesownbitXORneighbor’sbitXORmessagebit

• XORallannouncements=messagebit•  Everyrandomlygeneratedbitoccursinthissumtwice(andiscanceledbyXOR),messagebitoccursonce

31