Abstract
The objective of this paper is to improve the cost effectiveness of privacy impact assessments through (1) a more systematic approach, (2) a better integration with privacy by design and (3) enhanced reusability. We present a three-tier process including a generic privacy risk analysis depending on the specifications of the system and two refinements based on the architecture and the deployment context respectively. We illustrate our approach with the design of a biometric access control system.
Access this chapter
Tax calculation will be finalised at checkout
Purchases are for personal use only
Notes
- 1.
Incentives should be taken in a general sense here, including lack of awareness in the case of unintentional breach.
- 2.
And are independent of the architecture and the context.
- 3.
The user’s identity \(ID_i\) is used to fetch his enrolled template \(br_{i}\).
- 4.
For example, different locations correspond to different applicable laws (the motivation of a risk source may vary depending on the existence of data protection regulations and how strongly they are enforced), the strength (e.g., technical resources) or motivation of the local state to interfere [33], etc.
- 5.
This assumption should be valid at least for large scale attacks. However, one could argue that casinos may possess background information about certain frequent customers. Similarly, the state would be considered as having potentially a lot of background information but it is a more relevant risk source for surveillance than for identity theft. In any case, the assumptions made in this paper are for illustrative purposes only: different assumptions about background information could be made within the same framework.
- 6.
In order to err on the safe side in terms of privacy protection, we consider dependent nodes such that one node may imply the other nodes.
- 7.
In order to err on the safe side in terms of privacy protection, we consider dependent nodes such that each node may exclude the other nodes.
- 8.
Keys are assumed to be protected by techniques which are not discussed here (e.g. obfuscation).
- 9.
- 10.
References
Article 29 Data Protection Working Party: Guidelines on Data Protection Impact Assessment (DPIA) and determining whether processing is “likely to result in a high risk” for the purposes of Regulation 2016/679 (2017)
BBC Technology: Millions of Fingerprints Stolen in US Government Hack (2015)
Bringer, J., Chabanne, H., Métayer, D., Lescuyer, R.: Privacy by design in practice: reasoning about privacy properties of biometric system architectures. In: Bjørner, N., de Boer, F. (eds.) FM 2015. LNCS, vol. 9109, pp. 90–107. Springer, Cham (2015). doi:10.1007/978-3-319-19249-9_7
Cavoukian, A.: Privacy by Design: The 7 Foundational Principles Implementation and Mapping of Fair Information Practices. Office of the Information and Privacy Commissioner, Ontario, Canada Standards (2010)
Cavoukian, A., Chibba, M., Stoianov, A.: Advances in biometric encryption: taking privacy by design from academic research to deployment. Rev. Policy Res. 29(1), 37–61 (2012)
Cavoukian, A., Stoianov, A.: Privacy by Design Solutions for Biometric One-to-Many Identification Systems (2014)
Colesky, M., Hoepman, J., Hillen, C.: A critical analysis of privacy design strategies. In: 2016 IEEE Security and Privacy Workshops, SP Workshops 2016, San Jose, CA, USA, 22–26 May 2016, pp. 33–40 (2016)
Commission Nationale de l’Informatique et des Libertes (CNIL): Methodology for Privacy Risk Management - How to Implement the Data Protection Act (2012)
Commission Nationale de l’Informatique et des Libertes (CNIL): Privacy Impact Assessment (PIA) Methodology (How to Carry Out a PIA) (2015)
Commission Nationale de l’Informatique et des Libertes (CNIL): Privacy Impact Assessment (PIA) Tools (templates and knowledge bases) (2015)
Dantcheva, A., Elia, P., Ross, A.: What Else Does Your Biometric Data Reveal? A Survey on Soft Biometrics (2015)
De, S.J., Métayer, D.: PRIAM: a privacy risk analysis methodology. In: Livraga, G., Torra, V., Aldini, A., Martinelli, F., Suri, N. (eds.) DPM/QASA -2016. LNCS, vol. 9963, pp. 221–229. Springer, Cham (2016). doi:10.1007/978-3-319-47072-6_15
De, S.J., Le Métayer, D.: Privacy harm analysis: a case study on smart grids. In: International Workshop on Privacy Engineering (IWPE). IEEE (2016)
De, S.J., Le Métayer, D.: Privacy risk analysis. In: Synthesis Series. Morgan & Claypool Publishers (2016)
De, S.J., Le Métayer, D.: A Risk-based Approach to Privacy by Design (Extended Version). No. RR-9001, December 2016
De, S.J., Le Métayer, D.: PRIAM: A Privacy Risk Analysis Methodology. INRIA Research Report (RR-8876), July 2016
Deng, M., Wuyts, K., Scandariato, R., Preneel, B., Joosen, W.: A privacy threat analysis framework: supporting the elicitation and fulfilment of privacy requirements. Requirements Eng. 16(1), 3–32 (2011)
European Commission: Regulation (EU) 2016/679 of the European Parliament and of the Council of 27 on the protection of natural persons with regard to the processing of personal data and on the free movement of such data, and repealing Directive 95/46/EC (General Data Protection Regulation), April 2016
Expert Group. 2 of Smart Grid Task Force: Data Protection Impact Assessment Template for Smart Grid and Smart Metering Systems (2014)
Frakes, W.B., Kang, K.: Software reuse research: status and future. IEEE Trans. Softw. Eng. 31(7), 529–536 (2005)
Friginal, J., Guiochet, J., Killijian, M.O.: A privacy risk assessment methodology for location-based systems. http://homepages.laas.fr/guiochet/telecharge/MOBIQUITOUS2013.pdf. Accessed 13 July 2016
Garcia, M., Lefkovitz, N., Lightman, S.: Privacy Risk Management for Federal Information Systems (NISTIR 8062 (Draft)). National Institute of Standards and Technology (2015)
Gartland, C.: Biometrics Are a Grave Threat to Privacy (2016). The New York Times
Gürses, S., Troncoso, C., Diaz, C.: Engineering privacy by design. Comput. Priv. Data Prot. 14(3) (2011)
Gürses, S., Troncoso, C., Diaz, C.: Engineering Privacy by Design Reloaded. Amsterdam Privacy Conference (2015)
Hoepman, J.-H.: Privacy design strategies. In: Cuppens-Boulahia, N., Cuppens, F., Jajodia, S., Abou El Kalam, A., Sans, T. (eds.) SEC 2014. IFIP AICT, vol. 428, pp. 446–459. Springer, Heidelberg (2014). doi:10.1007/978-3-642-55415-5_38
Kobie, N.: Surveillance State: Fingerprinting Pupils Raises Safety and Privacy Concerns (2016). The Guardian
Mcilroy, M.: Mass produced software components (1969)
Miglani, S., Kumar, M.: India’s Billion-member Biometric Database Raises Privacy Fears (2016). Reuters
Mili, A., Chmiel, S.F., Gottumukkala, R., Zhang, L.: An integrated cost model for software reuse. In: Proceedings of the 2000 International Conference on Software Engineering, pp. 157–166. IEEE (2000)
Oetzel, M.C., Spiekermann, S.: A systematic methodology for privacy impact assessments: a design science approach. Eur. J. Inform. Syst. 23(2), 126–150 (2014)
Oetzel, M.C., Spiekermann, S., Grüning, I., Kelter, H., Mull, S.: Privacy Impact Assessment Guideline for RFID Applications (2011)
Oppenheim, C.: Big Brother Spying is Reaching Scary Levels (2013). http://edition.cnn.com/2013/12/10/opinion/oppenheim-privacy-reform/
Pearson, S., Benameur, A.: A decision support system for design for privacy. In: Fischer-Hübner, S., Duquenoy, P., Hansen, M., Leenes, R., Zhang, G. (eds.) Privacy and Identity 2010. IFIP AICT, vol. 352, pp. 283–296. Springer, Heidelberg (2011). doi:10.1007/978-3-642-20769-3_23
Prabhakar, S., Pankanti, S., Jain, A.K.: Biometric recognition: security and privacy concerns. IEEE Secur. Priv. 1(2), 33–42 (2003)
del Prado, N., Cortez, M., Friginal, J.: Geo-location inference attacks: from modelling to privacy risk assessment. In: Tenth European Dependable Computing Conference (EDCC), pp. 222–225. IEEE (2014)
Prieto-Díaz, R.: Status report: software reusability. IEEE Softw. 10(3), 61–66 (1993)
Spiekermann, S., Cranor, L.F.: Engineering privacy. IEEE Trans. Softw. Eng. 35(1), 67–82 (2009)
Standish, T.A.: An essay on software reuse. IEEE Trans. Softw. Eng. 10(5), 494–497 (1984)
Tillman, G.: Opinion: Stolen Fingers: The Case Against Biometric Identity Theft Protection (2009). Computer World
Woodward, J.D.: Biometrics: privacy’s foe or privacy’s friend? Proc. IEEE 85(9), 1480–1492 (1997)
Wright, D., De Hert, P.: Privacy Impact Assessment. Springer, Netherlands (2012)
Acknowledgments
This work has been partially funded by the French ANR-12-INSE-0013 project BIOPRIV and Inria Project Lab CAPPRIS.
Author information
Authors and Affiliations
Corresponding author
Editor information
Editors and Affiliations
Appendices
A Description of Phase 2 for Arch.1 and Arch.3
1.1 A.1 Arch.1: Use of an Encrypted Database
Description of Arch.1. In the simple biometric access control architecture pictured in Fig. 8, the server S stores the database of encrypted reference templates \(\overline{ebr}\) and the access control rules ac. When the user presents his identity \(ID_i\) and fresh biometric \(rd_i\) to the terminal T, T fetches the encrypted reference template \(\overline{ebr_{i}}\) from S, decrypts it using the key \(k_{br}\) and compares \(br_i\) with \(bs_i\) produced from \(rd_i\) by T (taking into account thr). The access control decision \(dec_i\) is used to allow or deny access. The access logs \(\overline{at}\) of different users are encrypted into \(\overline{eat}\) and sent back by the terminal T at regular intervals to be stored in the server S. The access log \(\overline{at}\) is updated after each access control.
The keysFootnote 8 \(k_{at}\) and \(k_{br}\), the threshold thr and access control rules ac are persistently stored in the terminal T Footnote 9. In contrast, \(\overline{at}\) is stored in T only for short time intervals. \(dec_i\), \(rd_i\), \(bs_i\), \(\overline{br_{i}}\), \(ts_i\), \(\overline{at}\), \(\overline{eat}\), \(\overline{ebr_{i}}\), \(ID_i\) are deleted from the terminal T as soon as their use is overFootnote 10.
The components in this architecture are therefore: the terminal T (C.1) and the server S (C.2).
Risk Sources for Arch.1. Since the architecture does not include any security components, we assume that no security operator is involved. The risk sources are therefore: the owner (A.1), cybercriminals (A.3), the state (A.4) and third parties (A.5). The owner (A.1) controls both the server S and the terminal T.
Personal Data for Arch.1 and Their Exploitability. At this stage, the privacy analyst presents each data element stored in each system component and its exploitability (see Table 7). As explained in Sect. 2, by “transient exploitation” of a component we mean exploitation for a short period of time or infrequent exploitation, (e.g., once in several months), whereas “persistent exploitation” means the exploitation of a component for a long period of time (e.g., for several days or months). For example, \(dec_i\) provides the result of one access control for user i, whereas \(\overline{at}\) provides the access log of all users for all previous days. So to know the access log of all users over t days, the risk source must either access all \(dec_i\) for all users for each of the t days (persistent exploitation) or access \(\overline{at}\) at the end of the t days (transient exploitation).
Refinement of Generic Harm Trees for Arch.1. In this phase, we consider the harm identity theft (H.2). Figure 9 shows the harm tree corresponding to this harm. Figure 12 in Appendix B shows how the generic harm tree (Fig. 2) for identity theft is pruned to obtain the architecture specific harm tree in Fig. 9. From Sect. A.1, we know that the risk sources for Arch.1 do not include A.2. Therefore, all branches of the generic harm tree for identity theft (H.2) that contain A.2 are pruned (pruned branches are marked by a red cross in Fig. 12). The definition of the architecture also makes it possible to instantiate the generic components \(C_i\), \(C_j\), \(C_k\), \(C_l\), \(C_m\) and \(C_n\).
1.2 A.2 Arch.3: Match-on-Card Technology
Description of Arch.3. Arch.2 is more protective than Arch.1 as the former uses a secure component M to perform the comparison between the fresh template and the reference template. In addition, it involves a security operator (A.2) for a better separation of responsibilities. However, in Arch.2, the fresh reference template \(bs_i\) is still available in T along with \(ID_i\). Moreover, the clear template \(\overline{br_i}\) can still be accessed by the security operator (A.2) who controls M. In fact, A.2 has access to a lot of personal data. One way to overcome these difficulties is to use the match-on-card technology. In Arch.3, pictured in Fig. 10, each user possesses a smart card C that stores his identity \(ID_i\) along with his enrolled template \(br_i\) (i.e., it stores \(\overline{br_i}\)), the threshold thr and access control rules ac and performs the matching operation without disclosing \(ID_i\) or \(br_i\) to the terminal T. The owner does not store any database of reference templates.
The user inserts the card into the terminal T and submits the fresh biometric raw data \(rd_i\). T derives a fresh template \(bs_i\) from \(rd_i\) and transfers it to C. C compares \(bs_i\) with \(br_i\) using the threshold thr and transfers the result of the access control \(dec_i\) to T. T informs the user about \(dec_i\) and sends it to the physical access control mechanism. The card C does not transfer any information apart from \(dec_i\) (not even the user identity \(ID_i\)) to T. C is assumed to be completely secure (e.g., it is tamper-resistant and personalized by a certified issuer during the enrolment phase). Both \(rd_i\) and \(bs_i\) as well as \(dec_i\) are deleted from T and C as soon as their uses are over. No access log \(\overline{at}\) is recorded.
The system components in this architecture are: the terminal T (C.1) and the smart card C (C.4).
Risk Sources for Arch.3. We assume that there is no security operator (A.2) in this architecture, since the security relies only on the smart cards possessed by the users. Therefore, the risk sources to be considered include: the owner (A.1), cybercriminals (A.3), the state (A.4) and third parties (A.5). The owner (A.1) controls the terminal T.
Personal Data and Their Exploitability for Arch.3. Table 8 presents each data item stored in each system component and the corresponding exploitability values for Arch.3. A risk source must have enough technical resources to exploit T persistently to get access to \(dec_i\), \(rd_i\) or \(bs_i\). However, in contrast with Arch.1 and Arch.2, \(ID_i\) is not stored in any component in Arch.3. Thus, in order to exploit \(dec_i\) or \(rd_i\), \(bs_i\), risk sources must have \(ID_i\) as background information. Since C is considered to be secure and belongs to the user, it does not appear in Table 8.
Refinement of Generic Harm Trees for Arch.3. Figure 15 in Appendix B shows how the generic harm tree for identity theft (H.2) (presented in Fig. 2) can be pruned to derive the corresponding harm tree for Arch.3 (presented in Fig. 11). In Arch.3, \(ID_i\), \(\overline{br_i}\), \(\overline{ebr_i}\) and \(k_{br}\) are not present at any moment in any of the components that the risk sources may access (i.e., terminal T). So all branches in the generic tree corresponding to these data elements are pruned. Also, the risk source A.2 is not a part of Arch.3. So all branches concerning A.2 are pruned too. The definition of the architecture also makes it possible to instantiate the generic components \(C_i\), \(C_j\), \(C_k\), \(C_l\), \(C_m\) and \(C_n\).
B Pruning of Harm Trees and Likelihood Computation for Identity Theft (H.2)
In this appendix, we present the harm trees for identity theft, showing in detail how branches of the generic tree are pruned based on different conditions (related to the architecture and the context) discussed in the paper.
Rights and permissions
Copyright information
© 2017 Springer International Publishing AG
About this paper
Cite this paper
De, S.J., Le Métayer, D. (2017). A Refinement Approach for the Reuse of Privacy Risk Analysis Results. In: Schweighofer, E., Leitold, H., Mitrakas, A., Rannenberg, K. (eds) Privacy Technologies and Policy. APF 2017. Lecture Notes in Computer Science(), vol 10518. Springer, Cham. https://doi.org/10.1007/978-3-319-67280-9_4
Download citation
DOI: https://doi.org/10.1007/978-3-319-67280-9_4
Published:
Publisher Name: Springer, Cham
Print ISBN: 978-3-319-67279-3
Online ISBN: 978-3-319-67280-9
eBook Packages: Computer ScienceComputer Science (R0)