Computing Reviews
Today's Issue Hot Topics Search Browse Recommended My Account Log In
Review Help
Search
A framework for privacy-preserving data publishing with enhanced utility for cyber-physical systems
Sangogboye F., Jia R., Hong T., Spanos C., Kjærgaard M. ACM Transactions on Sensor Networks14 (3-4):1-22,2018.Type:Article
Date Reviewed: Feb 19 2020

I recently did research on cyber-physical systems (CPSs) based on sensors that collect data about the humans who are wearing them. The actual problem that arises in such an environment is related to privacy, because CPS deployment may provide sensitive personal information. So, I was searching for resources focused on these issues. Plenty of titles deal with the Internet of Things (IoT), embedded systems, wearable sensors, and CPSs, but they rarely focus on a privacy-preserving architecture when data publishing is incorporated into the system. Fortunately, I found this paper to be a worthwhile study on privacy challenges related to publishing data collected within actual CPSs. It is clear that the convergence of computing and physical sensing creates a complex engineering ecosystem in which sensitive data exposure with privacy attacks could jeopardize the integrity and security of these systems.

However, it should be noted that CPSs are not just two separate parts, but also the interaction of the physical and the cyber parts, thus the need for new concepts of design and frameworks that can deal with personal data and privacy issues. Security and privacy are the great concerns for CPSs in which privacy attacks target data collections that can be used to leak sensitive information. The problem arises with the need for data publishing, where a maximum form of anonymity must be provided and guaranteed.

The authors start with the premise that the distributed sensing, processing, and storage of massive amounts of data that CPSs provide impact privacy breaches when personal data is in use. Privacy-preserving data publishing frameworks can be found in recent literature. The general goal is to prevent the linking of data records and sensitive information in the publishing process, with the highest data quality possible. The authors successfully confront the “current practice in publishing CPSs’ datasets,” that is, to provide agreements for regulating data use, sharing, and retention. Hence, they are conscious of how this approach is vulnerable, for example, “datasets are often anonymized by suppressing direct identifiers,” and instead apply k-anonymity models. It is interesting how they connect k-anonymity with privacy-preserving data publishing in an elegant way through the PAD ecosystem (which is presented in their previous work [1]).

An additional bonus that readers interested in PAD may find useful is an open-source project done in Python and developed through a GitHub repository: https://github.com/PAD-Protecting-Anonymity/PAD.

The PAD framework is extended to data utility based on deep neural networks. Thus, the authors use distance metric learning in “the accurate estimation of arrival and departure time from a database containing daily occupancy profiles.” The primary task in a CPS environment is “to publish the dataset with k-anonymity guarantee as well as high quality in support of the required data analysis.” They successfully present improved data reliability “by learning how the data is intended to be used and then adjusting the data perturbation algorithm accordingly”; microaggregation as the perturbation technique (mostly used in Eurostat) provides an acceptable level of privacy protection.

The overall value of this study lies in the evaluation and results: “using various datasets collected in real-world buildings,” the authors look at, for example, “the utility of PAD with a generic distance metric” and “the utility of PAD with a customized distance metric.”

Everyone involved in big data and data mining research and utilization within CPS and IoT environments should consider the privacy issues in data publishing. This study could be a valuable resource for their work. It is also recommended as a supplementary reference for undergraduate and postgraduate courses.

Reviewer:  F. J. Ruzic Review #: CR146896 (2007-0169)
1) Jia, R.; Sangogboye, F. C.; Hong, T.; Spanos, C.; Kjargaard, M. B. PAD: protecting anonymity in publishing building related datasets. In Proceedings of the 4th ACM International Conference on Systems for Energy-Efficient Built Environments ACM, 2017, Article No. 4.
Bookmark and Share
  Featured Reviewer  
 
General (H.2.0 )
 
 
Privacy (K.4.1 ... )
 
 
Sensor Networks (C.2.1 ... )
 
Would you recommend this review?
yes
no
Other reviews under "General": Date
Design of the Mneme persistent object store
Moss J. ACM Transactions on Information Systems 8(2): 103-139, 2001. Type: Article
Jul 1 1991
Database management systems
Gorman M., QED Information Sciences, Inc., Wellesley, MA, 1991. Type: Book (9780894353239)
Dec 1 1991
Database management (3rd ed.)
McFadden F., Hoffer J., Benjamin-Cummings Publ. Co., Inc., Redwood City, CA, 1991. Type: Book (9780805360400)
Jun 1 1992
more...

E-Mail This Printer-Friendly
Send Your Comments
Contact Us
Reproduction in whole or in part without permission is prohibited.   Copyright 1999-2024 ThinkLoud®
Terms of Use
| Privacy Policy