Controlling the what and where of declassification in language-based security

Publikationstyp
Conference Paper
Date Issued
2007-03
Sprache
English
Author(s)
Mantel, Heiko 
Reinhard, Alexander  
TORE-URI
http://hdl.handle.net/11420/13919
First published in
Lecture notes in computer science  
Number in series
4421 LNCS
Start Page
141
End Page
156
Citation
Lecture Notes in Computer Science 4421 LNCS: 141-156 (2007)
Contribution to Conference
16th European Symposium on Programming, ESOP 2007  
Publisher DOI
10.1007/978-3-540-71316-6_11
Scopus ID
2-s2.0-37149055443
Publisher
Springer
While a rigorous information flow analysis is a key step in obtaining meaningful end-to-end confidentiality guarantees, one must also permit possibilities for declassification. Sabelfeld and Sands categorized the existing approaches to controlling declassification in their overview along four dimensions and according to four prudent principles [16]. In this article, we propose three novel security conditions for controlling the dimensions where and what, and we explain why these conditions constitute improvements over prior approaches. Moreover, we present a type-based security analysis and, as another novelty, prove a soundness result that considers more than one dimension of declassification.
DDC Class
004: Informatik