honey onions: exposing snooping tor hsdir relays · • targeting drupal and ruby on rails 12....

15
Honey Onions: Exposing Snooping Tor HSDir Relays Guevara Noubir & Amirali Sanatinia {noubir,amirali}@ccs.neu.edu Northeastern University 1

Upload: others

Post on 25-Aug-2020

3 views

Category:

Documents


0 download

TRANSCRIPT

Page 1: Honey Onions: Exposing Snooping Tor HSDir Relays · • Targeting Drupal and Ruby on Rails 12. Snoopers’ Most Likely Geolocation 13. Snoopers’ Identity • Hard to identify the

Honey Onions: Exposing Snooping Tor HSDir Relays

GuevaraNoubir &Amirali Sanatinia{noubir,amirali}@ccs.neu.edu

NortheasternUniversity

1

Page 2: Honey Onions: Exposing Snooping Tor HSDir Relays · • Targeting Drupal and Ruby on Rails 12. Snoopers’ Most Likely Geolocation 13. Snoopers’ Identity • Hard to identify the

Motivations

• Previousresearchstudiedthemaliciousnessoftherelays• KnownbadExitnodes• Otherworklookedatthenatureofhiddenservicescontent• NopriorworkontheHiddenServiceDirectories(HSDirs)• IndexinghiddenservicesrequiresmodificationtoTor,whichcanbeanindicatorofsomeeffortandpotentiallymoremaliciousactivities

2

Page 3: Honey Onions: Exposing Snooping Tor HSDir Relays · • Targeting Drupal and Ruby on Rails 12. Snoopers’ Most Likely Geolocation 13. Snoopers’ Identity • Hard to identify the

Tor & Hidden Services

• Awidelyusedpracticalanonymityinfrastructure• Providesanonymityforboththeclientsandtheserverthroughhiddenservices• Dependsonthehonestbehaviorofthevolunteeringrelays• Itisknownthatsomerelaysaremisbehaving(BadExitnodes)• SomeExitnodesactivelytrytoperformManintheMiddleAttack(MITM)• NotmuchisknownabouttheHSDirs orHiddenServicesingeneral

3

Page 4: Honey Onions: Exposing Snooping Tor HSDir Relays · • Targeting Drupal and Ruby on Rails 12. Snoopers’ Most Likely Geolocation 13. Snoopers’ Identity • Hard to identify the

Hidden Service Directories (HSDir)

Client

Hidden Service

IP

RP

HSDir

(1)

(2)(3)

(4)

(5)

(6)

(7)

4

Page 5: Honey Onions: Exposing Snooping Tor HSDir Relays · • Targeting Drupal and Ruby on Rails 12. Snoopers’ Most Likely Geolocation 13. Snoopers’ Identity • Hard to identify the

Ring of Responsible HSDirs

5

Page 6: Honey Onions: Exposing Snooping Tor HSDir Relays · • Targeting Drupal and Ruby on Rails 12. Snoopers’ Most Likely Geolocation 13. Snoopers’ Identity • Hard to identify the

Honey Onions (HOnions)

• EachHOnion correspondstoaserver/process• RunonlocalIPaddress(HiddenService)• AccessibleonlythroughTorandnotsharedanywhere• Threeschedules• Daily• Weekly• Monthly

• Logtherequestsforfurtherinvestigation

6

Page 7: Honey Onions: Exposing Snooping Tor HSDir Relays · • Targeting Drupal and Ruby on Rails 12. Snoopers’ Most Likely Geolocation 13. Snoopers’ Identity • Hard to identify the

HOnions Architecture

1. Generate honions

hoi

hoj

2. Place honions on HSDirs3. Build bipartite graph

On visit, mark potential HSDirs

hoj

di

di+2

di+1

di

di+1

di+2

On visit, add to bipartite graph

7

Page 8: Honey Onions: Exposing Snooping Tor HSDir Relays · • Targeting Drupal and Ruby on Rails 12. Snoopers’ Most Likely Geolocation 13. Snoopers’ Identity • Hard to identify the

Set Cover Problem

• !"# = &': )*++,-./0123ℎ!"#2+5-.6

• !7 = ℎ*8:!792*93ℎ.31.0:2023,&

• ; = !"# ∪ !7

• = = ℎ*8, &8 ∈ !7×!"# ℎ*81.0A-.Bedon&'andwasvisited}

• ": ∀ ℎ*8, &' ∈ =, ∃&′' ∈ " ∧ ℎ*8, &′' ∈ =R⊆TRUVWXY'Z

• ThesetcoverisanNP-completeproblem• Canbecalculatedusingapproximationalgorithms• SetcovergivesthelowerboundonthenumberofsnoopingHSDirs

8

Page 9: Honey Onions: Exposing Snooping Tor HSDir Relays · • Targeting Drupal and Ruby on Rails 12. Snoopers’ Most Likely Geolocation 13. Snoopers’ Identity • Hard to identify the

Heuristic Approach

• Input:G(V,E): Bipartitie graphofHOnions toHSDirs• Output:S:Setexplainingvisits

• " ⟵ ∅

• while; ∩ !7 ≠ ∅_o• Pick& ∈ ; ∩ !"#: 123ℎℎ26ℎ,03 degree• ; ← ;\ &.9&230!792*99,26ℎb*+0

• end

9

Page 10: Honey Onions: Exposing Snooping Tor HSDir Relays · • Targeting Drupal and Ruby on Rails 12. Snoopers’ Most Likely Geolocation 13. Snoopers’ Identity • Hard to identify the

Integer Linear Programming (ILP)

• min (ef, … , eTRU)∑ e8|TRU|8jf

subjectto∀ℎ*' ∈ !7 ∑ e8 ≥ 1�∀8: nop,qr ∈s

• ProvidesalowerboundonthenumberofsnoopingHSDirs toexplainthevisits

10

Page 11: Honey Onions: Exposing Snooping Tor HSDir Relays · • Targeting Drupal and Ruby on Rails 12. Snoopers’ Most Likely Geolocation 13. Snoopers’ Identity • Hard to identify the

Connectivity Graph

11

Page 12: Honey Onions: Exposing Snooping Tor HSDir Relays · • Targeting Drupal and Ruby on Rails 12. Snoopers’ Most Likely Geolocation 13. Snoopers’ Identity • Hard to identify the

Snooping Behavior

• Widevarietyofbehavior• Automatedvsmanualprobing• Aggressive,periodicprobing• Attemptstofindvulnerabilities• SQLInjection• XSS• Pathtraversal• PHPEasterEggs• TargetingDrupalandRubyonRails

12

Page 13: Honey Onions: Exposing Snooping Tor HSDir Relays · • Targeting Drupal and Ruby on Rails 12. Snoopers’ Most Likely Geolocation 13. Snoopers’ Identity • Hard to identify the

Snoopers’ Most Likely Geolocation

13

Page 14: Honey Onions: Exposing Snooping Tor HSDir Relays · • Targeting Drupal and Ruby on Rails 12. Snoopers’ Most Likely Geolocation 13. Snoopers’ Identity • Hard to identify the

Snoopers’ Identity

• Hardtoidentifytherealentitybehindtherelays• MorethanhalfoftheHSDirs arehostedoncloudplatform• Thegeolocationscorrespondtothelocationofthehostingplatformandnotnecessarilytheentityrunningthem• Numberofcloudplatformsarelocatedincountrieswithstrongerprivacyprotectionforcostumers• Somecloudplatformacceptpaymentsoverbitcoin,makingitevenhardertoidentifytherealactors

14

Page 15: Honey Onions: Exposing Snooping Tor HSDir Relays · • Targeting Drupal and Ruby on Rails 12. Snoopers’ Most Likely Geolocation 13. Snoopers’ Identity • Hard to identify the

Conclusion

• HoneyOnions(HOnions)isaframeworktodetectsnoopingHSDirs• Providesalowerboundonsuchrelays• Torreliesonthehonestbehaviorofthevolunteeringrelays• Thedetection,identificationandmitigationofmisbehavingrelayshelpstoimprovetheprivacyandsecurityofTor• ThisworkisanadditiontothepreviousbodyofworkfocusingondetectionofmisbehavingTorrelays

15