user centric security models
play

User-centric Security Models Manuel Ziegler Advisor: Heiko - PowerPoint PPT Presentation

User-centric Security Models Manuel Ziegler Advisor: Heiko Niedermayer Seminar Future Internet SS 2015 Chair for Network Architectures and Services Department for Computer Science Technische Universitt Mnchen April 10, 2015 Technische


  1. User-centric Security Models Manuel Ziegler Advisor: Heiko Niedermayer Seminar Future Internet SS 2015 Chair for Network Architectures and Services Department for Computer Science Technische Universität München April 10, 2015 Technische Universität München – Chair for Network Architectures and Services 1

  2. Outline Introduction 1) Direct trust and well known decentralized trust models 2) Trust Establishment 3) Trust Decisions 4) Trust Metrics 5) Conclusion 6) Technische Universität München – Chair for Network Architectures and Services 2

  3. Introduction Centralized CA models:  TLS, S/MIME  Tree based certificate chain  Certificate Authority tells you, who to trust  CAs are paid for signing certificates root certificate (CA) TLS certificate TLS certificate S/MIME certificate example.com example.de John Doe Technische Universität München – Chair for Network Architectures and Services 3

  4. Introduction The user-centric approach:  Not just tree alike trust chains  User decides by his own, who to trust  Difficult decision making for less experienced users  When used correctly: Better trustworthiness than CA based Trust Alice Bob Chuck Carol Dan Technische Universität München – Chair for Network Architectures and Services 4

  5. Direct and well known decentralized Trust Models Direct trust model:  Personal meeting between users required  Does not scale  Can be combined with decentralized trust models  Some applications provide the direct trust model retrospectively: – TextSecure – PGP Technische Universität München – Chair for Network Architectures and Services 5

  6. Direct and well known decentralized Trust Models OpenPGP / GnuPG:  Public key servers – everybody can modify entries  Trust Signatures  Privacy Problem: Social environment can be determined from trust signatures  Often Trust is established with the help of CA models (Fingerprint on TLS encrypted Website) SSH:  Local key database  Trust on first connection  Similar Concept: Certificate Pinning in Web browsers Technische Universität München – Chair for Network Architectures and Services 6

  7. Trust Establishment Trust Chains:  Usually given as graph structures  Rating of trust relationships possible  Trust Relationships can be derived from a trust chain Technische Universität München – Chair for Network Architectures and Services 7

  8. Trust Establishment Trust-Rated Authentication in domain structured networks:  Trust Token with meta information is obtained from a Domain Authentication Server (Holz, et al. 2008 [1]) – Key exchange information – Known frauds – Previous contacts Technische Universität München – Chair for Network Architectures and Services 8

  9. Trust Establishment Trust Establishment with the help of a councelor:  PGP – Alice obtains the certificate from Bob through Carol (signed) – Bob obtains the certificate from Alice through Carol (signed) Technische Universität München – Chair for Network Architectures and Services 9

  10. Trust Establishment Trust Establishment with the help of a councelor:  Trust Exchange Protocol (Kinkelin, et al. 2014 [2]) Technische Universität München – Chair for Network Architectures and Services 10

  11. Trust Decissions Trustworthiness and reputation:  Reputation = Users Assessment of another user's ability to validate the identity of other users (Kinkelin, et al. 2014 [2])  Often Coupled with trustworthiness Supporting the decision making process:  Collect Information  Categorize  Evaluate  Display  Examples: Web of Trust Browser Plugin, OTO – Online Trust Oracle (Kim, et al. 2012 [3]) Technische Universität München – Chair for Network Architectures and Services 11

  12. Trust Decissions Source: Kim, et al. 2012 [3] Technische Universität München – Chair for Network Architectures and Services 12

  13. P 1 = A → B → E Trust Metrics P 2 = A → C → E P 3 = A → D → F → E P = e 1 ⋅ ... ⋅ e k n k ∑ ∏ e i, j j = 1 i = 1 T = n T = 0,8 ⋅ 0,3 + 0,6 ⋅ 0,7 + 0,9 ⋅ 0,8 ⋅ 0,9 3 T = 0,24 + 0,42 + 0,648 3 T = 0,436 Technische Universität München – Chair for Network Architectures and Services 13

  14. Trust Metrics IL = Identification Level RL = Reputation Level IL CI ⋅ RL RC ⋅ IL RC ∑ RL max IL max C ∈ Councelors IL RI = ⋅ d ∣ Counselors ∣ Kinkelin, et al. 2014 [2] Technische Universität München – Chair for Network Architectures and Services 14

  15. Conclusion  Many different decentralized trust models  Few user supporting trust decision systems  Decentralized Trust models require expert knowledge  Most global infrastructure based on CA based models → Still a lot work to do Technische Universität München – Chair for Network Architectures and Services 15

  16. References [1] R. Holz, H. Niedermayer, P. Hauck, and G. Carle: Trust rated authentication for domain structured distributed systems , Public Key Infrastructure – 5th European PKI Workshop: Theory and Practice, 2008 [2] H. Kinkelin, H. Niedermayer, S. Mittelberger, and G. Carle: Human-centric and privacy protecting trust establishment over the internet , CENTRIC: The Seventh International Conference on Advances in Human-oriented and Personalized Mechanisms, Technologies, and Services, 2014 [3] T. H.-J. Kim, P. Gupta, J. Han, E. Owusu, J. Hong, A. Perrig, and Debin: OTO: Online Trust Oracle for user-centric trust establishment , ACM conference on Computer and communications security, 2012 Technische Universität München – Chair for Network Architectures and Services 16

  17. Questions Technische Universität München – Chair for Network Architectures and Services 17

Recommend


More recommend