Declarative cartography under fine-grained access control

TY - GEN

T1 - Declarative cartography under fine-grained access control

AU - Jensen, Thomas

AU - Vaz Salles, Marcos Antonio

AU - Bang, Michael Vindahl

PY - 2018

Y1 - 2018

N2 - Visualization of spatial data is of increasing importance in science and society, but opens up justified concerns about data privacy and security. A classic methodology for cartography through generalization is data selection; however, data selection can be challenging under security constraints for two main reasons. First, individual records are kept in the visualization, so a data security approach such as access control needs to be put in place to avoid leakage of information about protected records to unauthorized parties. Second, it can be computationally hard to pick out records from a large spatial dataset so as to create an aesthetically pleasing visualization respecting user constraints and optimization goals. The latter expense can get compounded by the need to additionally respect access control restrictions. This paper presents a way to integrate label-based access control into an existing technique for declarative cartography termed global selection. Through a set of theorems and new algorithms, we demonstrate that we can reuse derivation and resolution of record conflicts when computing global selections across access roles in a security hierarchy. In experiments with realistic datasets, the runtime of the best among these new methods achieves an improvement of up to 2x-5x compared with repeatedly computing the global selection in medium-to-large security hierarchies.

AB - Visualization of spatial data is of increasing importance in science and society, but opens up justified concerns about data privacy and security. A classic methodology for cartography through generalization is data selection; however, data selection can be challenging under security constraints for two main reasons. First, individual records are kept in the visualization, so a data security approach such as access control needs to be put in place to avoid leakage of information about protected records to unauthorized parties. Second, it can be computationally hard to pick out records from a large spatial dataset so as to create an aesthetically pleasing visualization respecting user constraints and optimization goals. The latter expense can get compounded by the need to additionally respect access control restrictions. This paper presents a way to integrate label-based access control into an existing technique for declarative cartography termed global selection. Through a set of theorems and new algorithms, we demonstrate that we can reuse derivation and resolution of record conflicts when computing global selections across access roles in a security hierarchy. In experiments with realistic datasets, the runtime of the best among these new methods achieves an improvement of up to 2x-5x compared with repeatedly computing the global selection in medium-to-large security hierarchies.

UR - http://www.scopus.com/inward/record.url?scp=85054936352&partnerID=8YFLogxK

U2 - 10.1145/3221269.3232012

DO - 10.1145/3221269.3232012

M3 - Article in proceedings

AN - SCOPUS:85054936352

BT - Proceedings of the 30th International Conference on Scientific and Statistical Database Management (SSDBM)

PB - Association for Computing Machinery

T2 - 30th International Conference on Scientific and Statistical Database Management, SSDBM 2018

Y2 - 9 July 2018 through 11 July 2018

ER -