OB-PWS: Obfuscation-Based Private Web Search Ero Balsa , Carmela - PowerPoint PPT Presentation

OB-PWS: Obfuscation-Based Private Web Search Ero Balsa , Carmela Troncoso and Claudia Diaz ESAT/COSIC, IBBT - KU Leuven Wednesday, 23 May 2012

Introduction Modelling OB-PWS The Privacy Problem Existing OB-PWS Systems Our contribution Summary, future work and conclussions The Privacy Problem art sports music E. Balsa, C. Troncoso and C. Diaz Obfuscation-Based Private Web Search 2/13

Introduction Modelling OB-PWS The Privacy Problem Existing OB-PWS Systems Our contribution Summary, future work and conclussions The Privacy Problem restaurants in Chicago art quit sports music bio smoking products E. Balsa, C. Troncoso and C. Diaz Obfuscation-Based Private Web Search 2/13

Introduction Modelling OB-PWS The Privacy Problem Existing OB-PWS Systems Our contribution Summary, future work and conclussions The Privacy Problem restaurants in Chicago cross-dressing art quit sports music Eco bio HIV smoking Activism products treatment E. Balsa, C. Troncoso and C. Diaz Obfuscation-Based Private Web Search 2/13

Introduction Modelling OB-PWS The Privacy Problem Existing OB-PWS Systems Our contribution Summary, future work and conclussions The Privacy Problem restaurants in Chicago cross-dressing art quit sports music Eco bio HIV smoking Activism products treatment PRIVACY PROBLEM: Individual search queries and/or profiling may reveal sensitive information. E. Balsa, C. Troncoso and C. Diaz Obfuscation-Based Private Web Search 2/13

Introduction Modelling OB-PWS The Privacy Problem Existing OB-PWS Systems Our contribution Summary, future work and conclussions The Privacy Problem restaurants in Chicago cross-dressing art quit sports music Eco bio HIV smoking Activism products treatment PRIVACY PROBLEM: Individual search queries and/or profiling may reveal sensitive information. Some solutions: Anonymous communications PIR OB-PWS ⇒ Prevent profiling and provide query deniability . E. Balsa, C. Troncoso and C. Diaz Obfuscation-Based Private Web Search 2/13

Introduction Modelling OB-PWS The Privacy Problem Existing OB-PWS Systems Our contribution Summary, future work and conclussions Our contribution General model. Evaluation framework ⇒ with relevant privacy properties (details in the paper). Analysis of 6 existing systems (4 in this talk). E. Balsa, C. Troncoso and C. Diaz Obfuscation-Based Private Web Search 3/13

Introduction Modelling OB-PWS Abstract model Existing OB-PWS Systems Evaluation framework Summary, future work and conclussions An abstract model for OB-PWS real queries the user E. Balsa, C. Troncoso and C. Diaz Obfuscation-Based Private Web Search 4/13

Introduction Modelling OB-PWS Abstract model Existing OB-PWS Systems Evaluation framework Summary, future work and conclussions An abstract model for OB-PWS real queries the user semantic classification algorithm real profile E. Balsa, C. Troncoso and C. Diaz Obfuscation-Based Private Web Search 4/13

Introduction Modelling OB-PWS Abstract model Existing OB-PWS Systems Evaluation framework Summary, future work and conclussions An abstract model for OB-PWS dummy queries real queries the user dummy semantic generation classification strategy algorithm real profile E. Balsa, C. Troncoso and C. Diaz Obfuscation-Based Private Web Search 4/13

Introduction Modelling OB-PWS Abstract model Existing OB-PWS Systems Evaluation framework Summary, future work and conclussions An abstract model for OB-PWS dummy queries unclassified queries real queries the user dummy semantic adversarial semantic generation classification classification algorithm strategy algorithm real profile observed profile E. Balsa, C. Troncoso and C. Diaz Obfuscation-Based Private Web Search 4/13

Introduction Modelling OB-PWS Abstract model Existing OB-PWS Systems Evaluation framework Summary, future work and conclussions An abstract model for OB-PWS profile filtering algorithm dummy queries unclassified queries filtered real queries profile the user queries classified as real queries dummy classified semantic dummy adversarial semantic generation as dummies classification classification classification algorithm strategy algorithm algorithm real profile observed profile E. Balsa, C. Troncoso and C. Diaz Obfuscation-Based Private Web Search 4/13

Introduction Modelling OB-PWS Abstract model Existing OB-PWS Systems Evaluation framework Summary, future work and conclussions An Evaluation framework for DGS A dual analysis is required: E. Balsa, C. Troncoso and C. Diaz Obfuscation-Based Private Web Search 5/13

Introduction Modelling OB-PWS Abstract model Existing OB-PWS Systems Evaluation framework Summary, future work and conclussions An Evaluation framework for DGS A dual analysis is required: Query-Based Analysis Exploit vulnerabilities in the DGS to distinguish real from dummy queries. E. Balsa, C. Troncoso and C. Diaz Obfuscation-Based Private Web Search 5/13

Introduction Modelling OB-PWS Abstract model Existing OB-PWS Systems Evaluation framework Summary, future work and conclussions An Evaluation framework for DGS A dual analysis is required: Query-Based Analysis Profile-Based Analysis Exploit vulnerabilities in the DGS to distinguish Exploit vulnerabilities in the DGS to filter real from dummy queries. observed profile and recover the real profile. E. Balsa, C. Troncoso and C. Diaz Obfuscation-Based Private Web Search 5/13

Introduction GooPIR Modelling OB-PWS PDS Existing OB-PWS Systems PRAW Summary, future work and conclussions OQF-PIR GooPIR h ( k )-Private Information Retrieval from Privacy-Uncooperative Queryable Databases [1] . A k -anonymity inspired approach. Prevents attacks based on: Timing/metadata. Popularity of queries. Statistical disclosure. However does not consider the ⇒ No dummy indistinguishability. topic of the queries. E. Balsa, C. Troncoso and C. Diaz Obfuscation-Based Private Web Search 6/13

Introduction GooPIR Modelling OB-PWS PDS Existing OB-PWS Systems PRAW Summary, future work and conclussions OQF-PIR PDS Plausibly Deniable Search [2] Lion CATS Leopard CATS Tiger CATS E. Balsa, C. Troncoso and C. Diaz Obfuscation-Based Private Web Search 7/13

Introduction GooPIR Modelling OB-PWS PDS Existing OB-PWS Systems PRAW Summary, future work and conclussions OQF-PIR PDS Plausibly Deniable Search [2] (dummy) (dummy) BATHROOM BUSINESS Lion S h o w e r S t o c k CATS Leopard CATS Tiger CATS E. Balsa, C. Troncoso and C. Diaz Obfuscation-Based Private Web Search 7/13

Introduction GooPIR Modelling OB-PWS PDS Existing OB-PWS Systems PRAW Summary, future work and conclussions OQF-PIR PDS Plausibly Deniable Search [2] (dummy) (dummy) BATHROOM BUSINESS Lion S h o w e r S t o c k CATS (dummy) (dummy) BATHROOM BUSINESS Leopard CATS S i n k I n v e s t i n g (dummy) (dummy) BATHROOM BUSINESS Tiger CATS T o i l e t S h a r e s E. Balsa, C. Troncoso and C. Diaz Obfuscation-Based Private Web Search 7/13

Introduction GooPIR Modelling OB-PWS PDS Existing OB-PWS Systems PRAW Summary, future work and conclussions OQF-PIR PDS Plausibly Deniable Search [2] (dummy) (dummy) BATHROOM BUSINESS Lion S h o w e r S t o c k S T A C (dummy) (dummy) BATHROOM BUSINESS Leopard S S i n k I n v e s t i n g T A C (dummy) (dummy) BATHROOM BUSINESS Tiger S T o i l e t S h a r e s T A C { { Disneyland Disneyland S N a p o l e o n T D Y N I O R E ) K T Disneyland M N a p o l e o n y S E m I S m H U S u Y M K d R A R ( O A T P S Justin Bieber I H Justin Bieber S E i n s t e i n D I ) K E Justin Bieber y C C E i n s t e i n m E N I m I S u S C U d C S M ( I T oy Story S Y H P T oy Story S B M W D ) I S S y K R T oy Story E B M W m A I m C V u O d S M ( R A C E. Balsa, C. Troncoso and C. Diaz Obfuscation-Based Private Web Search 7/13

Introduction GooPIR Modelling OB-PWS PDS Existing OB-PWS Systems PRAW Summary, future work and conclussions OQF-PIR PRAW (A PRivAcy model for the Web) [3] Privacy = Dissimilarity. Dissimilarity ∝ amount of dummy queries. E. Balsa, C. Troncoso and C. Diaz Obfuscation-Based Private Web Search 8/13

Introduction GooPIR Modelling OB-PWS PDS Existing OB-PWS Systems PRAW Summary, future work and conclussions OQF-PIR PRAW (A PRivAcy model for the Web) [3] Privacy = Dissimilarity. Dissimilarity ∝ amount of dummy queries. high probability region for the real profile observed profile distance between profiles (depends on dummy rate) E. Balsa, C. Troncoso and C. Diaz Obfuscation-Based Private Web Search 8/13

Introduction GooPIR Modelling OB-PWS PDS Existing OB-PWS Systems PRAW Summary, future work and conclussions OQF-PIR PRAW (A PRivAcy model for the Web) [3] Privacy = Dissimilarity. Dissimilarity ∝ amount of dummy queries. high probability region for the real profile Considering prior information Pr[ X = X ]: observed profile distance between profiles (depends on dummy rate) E. Balsa, C. Troncoso and C. Diaz Obfuscation-Based Private Web Search 8/13