01576nas a2200157   4500000000100000000000100001008004100002260001200043100001700055700001500072700002200087245008900109856007600198490000700274520113700281       2024                            d  c09/20241 aErol Gelenbe1 aMert Nakip1 aMiltiadis Siavvas00aSystem Wide Vulnerability and Trust in Multi-Component Communication System Software  uhttps://drive.google.com/file/d/1123-TES0XmyG8_l1XlmAfdWV5Jf6QPeo/view 0 v393 a<p>In the software rich environment of 6G,&nbsp;systems will be surrounded by edge devices that support distributed software&nbsp;systems which are critical to&nbsp;operations. Such systems may also be subject to frequent&nbsp;updates or uploads of individual software components.&nbsp;Trust in such systems will therefore depend on our ability&nbsp;to rapidly ensure that such software is not vulnerable to&nbsp;cyberattacks or malicious compromises. Thus this paper&nbsp;presents a novel System-Wide Vulnerability Assessment&nbsp;(SWVA) framework based on Machine Learning, that&nbsp;can be frequently activated to assess the vulnerability of&nbsp;interconnected software components over edge systems.&nbsp;The performance of the SWVA framework is illustrated&nbsp;by assessing the vulnerability of 13 versions of a real world 11 component software system, and comparingthe ARNN results against the well-known ML models&nbsp;MLP, KNN, and Lasso. The results show the superior&nbsp;performance of SWVA, offering over 85% median accuracy&nbsp;and good scalability as the number of connected software&nbsp;components increases.</p>