Digitale Ausweise für physische Identifikation?

Transcription

1 Digitale Ausweise für physische Identifikation? Univ.-Prof. Dr. René Mayrhofer und Michael Hölzl, MSc Institut für Netzwerke und Sicherheit, Johannes Kepler Universität Linz Vortrag zur IKT-Sicherheitskonferenz :15, St. Johann im Pongau

2 Motivating Scenario: Convergence of Security-Critical Services 2

3 Digital Identity: State of the Art OpenID: some (large) providers, many (small) consumers Facebook Google... FIDO U2F UAF Österreichische Bürgerkarte all optimized for web page login, not physical identification 3

4 Motivating Scenario: Convergence of Security-Critical Services 4

5 Digital (Photo-) ID for Physical Identification Online solution: MIA (My Identity App) by Österreichische Staatsdruckerei App for smart phones for using virtual identity cards requires online connectivity for verifying these documents Current project in JRZ u smile: AmDL (Austrian mobile Driving License) with partners A1 Telekom Drei-Banken-EDV LG Nexera NXP Semiconductors Österreichische Staatsdruckerei SBA Research 5

6 AmDL Use Case 1: Identity Verification by Police All relevant attributes need to be presented, e.g. ID number: Name Date of birth Full-resolution photo (optional) Biometric identifiers Vehicle classes Restrictions/limitations Surname: EINSTEIN Givenname: Albert Date of Birth: Sex: M Place of Birth: Only accessible to officially Ulm, Germany certified readers GENUINE Citizenship: USA, Switzerland Offline ID attributes transfer Signature: and offline verification Should also work when mobile phone battery is empty! 6

7 AmDL Use Case 2: Age Verification Age verification by e.g. Automated vending machines Bouncers at clubs Entrance staff for birthday rebate promotions ID number: Surname: EINSTEIN Givenname: Only age attribute should be Albert transferred in privacysensitive manner Date of Birth: > 16 years Not the full date of birth! Support binary yes/no answers for specific use case Sex: M Place of Birth: Ulm, Germany GENUINE Citizenship: USA, Switzerland Signature: 7

8 AmDL Use Case 3: Time-based Ticket for Public Transport Typical for public transport Monthly Yearly (any time period) No identifiers should to be transferred (no name, date of birth, etc.) for privacy reasons Location traces are highly sensitive personal data Still need to support typical operations Unforgeability Revocation (loss, theft, non- payment, etc.) 8

9 Requirements Functional Mobility Real-world identification Offline One-to-many Power-off Revocation Scalability Security Privacy Key confidentiality Unlinkability Unforgeability User control Communication protection Privacy-preserving attribute State-of-the-art cryptography queries 9

10 Extensible and Privacy-preserving Mobile eid 10

11 Technologies NFC Secure Element (SE) Protect identity keys Integrity assurances Code isolation Group signatures Members can sign on behalf of the group Anonymity in the group Unlinkability 11

12 Privacy-preserving Identification ID number: ID number: Surname: Surname: EINSTEIN EINSTEIN Givenname: Givenname: Albert Date of Birth: Sex: M Place of Birth: Ulm, Germany Albert Date of Birth: > 16 years GENUINE Citizenship: Sex: M Place of Birth: Ulm, Germany GENUINE Citizenship: USA, Switzerland USA, Switzerland Signature: Signature: Group signature of national eid 12

13 Privacy-preserving Identification Privacy-preserving attribute queries Attribute inequality query ID number: E.g. above 16 years old? Surname: EINSTEIN Attribute equality query E.g. Austrian citizen? Givenname: Albert Group membership query E.g. enrolled to public transport system? E.g. allowed to drive this vehicle class (for renting a car)? Date of Birth: > 16 years Sex: M Place of Birth: Ulm, Germany GENUINE Citizenship: Austrian USA,==Switzerland Signature: 13

14 Group/Division Membership ID number: Surname: EINSTEIN Givenname: Albert Date of Birth: Sex: M Place of Birth: Ulm, Germany Citizenship: USA, Switzerland Signature: Extensibility of eid for the use with numerous groups/divisions e.g. loyalty card, public transport ticket, students card, etc. Challenge: unlinkability, untraceability 14

15 Group/Division Membership ID number: Surname: EINSTEIN Givenname: Approach Pseudonym concept of Austrian Bürgerkarte Each division has an identifier eid generates pseudonym for division id d Albert id u, d =H (id u id d ) derive Divisions can add data to eid Protected with TOFU database C=E(pk d, id u,d n D) where n is a random nonce and D are data attributes 15

16 Extensible and Privacy-preserving Mobile eid 16

17 Revocation Scenarios User lost her phone (the eid) Revocation done by user eid has been withdrawn (e.g. holder deceased) Done by central authorities Additional threat: identity theft 17

18 Revocation Usual approach: revocation list Problem: no ID in group signature Additional challenges: 1. Items on the revocation list might loose anonymity 2. Additional computation effort 3. Could weaken unlinkability 4. Growing revocation list 18

19 Preserving Privacy beyond Revocation New revocation concept based on Offline revocation token generation on SE Bloom filter Features of our approach Scalability through efficient revocation check (O(1)) Backwards unlinkability Verifier-local revocation check Offline verification 19

20 Preserving Privacy beyond Revocation Revocation list based on bloom filter Probabilistic data structure Provides scalability Revocation token difficult to restore {x, {y,x,z }y, z } w 20

21 Preserving Privacy beyond Revocation Revocation filter {x, y, z } Verifier Prover Random identifier id v and challenge ch Prover Verifier Generated revocation token rt i, v =H (id v H (id u, se gpk c i )) Validation of group membership σ= sign (gsk i, ch rt i,v )

22 Preserving Privacy beyond Revocation Computation times 22

23 Open Issue: Backup of Digital Identity 23

24 Next Step: Digital Identity in the Cloud Friends with Faces LIKE ME! Trust us, we have a Privacy Shield! 24

25 You can search for everybody And we have assistants Next Step: Digital Identity in the Cloud Data stored for future improvement of all our services 25

26 We are looking for interested partners: - technology - use cases - funding... Next Step: Digital Identity in the Cloud 1 [globally universal] [open market] certify identity PERSONAL AGENT TPM 5 digital service discovery, distributed network setup 7 VERIFIER Digital world 4 2 [locally different] 6 BIOMETRIC SENSOR TPM 3 provide unique ID 8 trigger reaction Physical world Sicherheit in der Digitalisierung 26

27 Danke für Ihre Aufmerksamkeit! Fragen? Univ.-Prof. Dr. René Mayrhofer Vorstand des Instituts für Netzwerke und Sicherheit Michael Hölzl, MSc Universitätsassistent JOHANNES KEPLER UNIVERSITÄT LINZ Altenberger Str Linz, Österreich

28 Credits Icons SIM Card by Arthur Shlain from the Noun Project People pattern by Eliricon from the Noun Project Agreement by Chameleon Design from the Noun Project Smart Phone by Emily Haasch from the Noun Project Airplane Ticket by Creative Stall from the Noun Project Bus tickets by Iulia Ardeleanu from the Noun Project Loyalty Card by icon 54 from the Noun Project Business card by Karthik Aathis from the Noun Project Pictures Broken phone from Faces from Eye with Google logo from 28