Please use this identifier to cite or link to this item: https://www.um.edu.mt/library/oar/handle/123456789/58998
Title: Separation kernel robustness testing : the xtratum case study
Authors: Grixti, Stephen
Sammut, Nicholas
Hernek, Maria
Carrascosa, Elena
Masmano, Miguel
Crespo, Alfons
Keywords: Data types (Computer science)
Electric fault location
Computer architecture
Application program interfaces (Computer software)
Robust control
Soft errors (Computer science)
Reliability (Engineering)
Issue Date: 2016
Publisher: Institute of Electrical and Electronics Engineers Inc.
Citation: Grixti, S., Sammut, N., Hernek, M., Carrascosa, E., Masmano, M., & Crespo, A. (2016). Separation kernel robustness testing : the xtratum case study. IEEE International Conference on Cluster Computing, Taipei. 524-531.
Abstract: With time and space partitioned architectures becoming increasingly appealing to the European space sector, the dependability of separation kernel technology is a key factor to its applicability in European Space Agency projects. This paper explores the potential of the data type fault model, which injects faults through the Application Program Interface, in separation kernel robustness testing. This fault injection methodology has been tailored to investigate its relevance in uncovering vulnerabilities within separation kernels and potentially contributing towards fault removal campaigns within this domain. This is demonstrated through a robustness testing case study of the XtratuM separation kernel for SPARC LEON3 processors. The robustness campaign exposed a number of vulnerabilities in XtratuM, exhibiting the potential benefits of using such a methodology for the robustness assessment of separation kernels.
URI: https://www.um.edu.mt/library/oar/handle/123456789/58998
Appears in Collections:Scholarly Works - FacICTMN

Files in This Item:
File Description SizeFormat 
Separation_kernel_robustness_testing.pdf1.09 MBAdobe PDFView/Open


Items in OAR@UM are protected by copyright, with all rights reserved, unless otherwise indicated.