JASPER: Special Session on Joint Reliability And Security Assessment of SPlit Computing for Edge Robustness
Authors: Enrico Magliano, Giuseppe Esposito, Amir Hossein Shahdadian, Rama Mounika Kodamanchili, Juan David Guerrero Balaguera, Juan Esteban Rodriguez Condia, Annachiara Ruospo, Roberta Siciliano, Stefano Di Carlo, Maksim Jenihhin, Marco Levorato, Alessandro Savino, Matteo Sonza Reorda, Christian Herglotz, Michael Hübner, Mahdi Taheri
Organizations: Dept. of Control and Computer Engineering, Politecnico di Torino, Turin, Italy · Dept. of Electrical Engineering and Information Technology, University of Naples ”Federico II”, Naples, Italy · Dept. of Computer Systems, Tallinn University of Technology, Tallinn, EE · Computer Science Department, University of California, Irvine, US · Dept. of Computer Engineering, Brandenburg University of Technology, Cottbus, DE · Dept. of Computer Science, Chair of Computer Engineering, Humboldt University of Berlin, Berlin, Germany
Split Computing (SC) enables efficient deployment of Deep Neural Networks (DNNs) by partitioning inference between edge devices and cloud servers. However, intermediate feature representations are simultaneously exposed to hardware faults and adversarial attacks, which are traditionally evaluated independently. This paper presents a unified framework for the joint assessment of reliability and security in Split Computing. First, reliability is characterized through neuron-level fault injection using the Mean Relative Accuracy Degradation (MRAD) while security through feature-map-aware adversarial attacks simulations using the Attack Success Rate (ASR). Based on these complementary analyses, the Joint Vulnerability Score (JVS) is introduced, along with a confidence-aware extension that jointly captures prediction errors and confidence degradation. The framework is evaluated on ten Split Computing configurations based on ResNet-50 trained on ILSVRC-2012. Experimental results show substantial differences across compression strategies, with MRAD ranging from 44.3% to 61.2% under fault injection, while adversarial attacks achieve up to 98.8% ASR. Furthermore, the proposed joint metrics reveal vulnerability trends that remain hidden when reliability and security are analyzed independently, providing a more comprehensive methodology for designing dependable Split Computing systems.
Figures & tables
Fig. 1 : Overview of the reliability and security threats considered in SC and their impact on safety-critical inference.
Fig. 2 : A general scheme from [ 11 ] of physical defects propagation across system components ( Left ), and an illustration of fault propagation effects on matrix multiplication operations through different levels of the tiling ( Right ).
Fig. 3 : MRAD as a function of the fault-corruption magnitude (BER).
Fig. 4 : MRAD VS Magnitude of corruption (i.e., BER) VS Confidence Gap Deviation
Fig. 5 : ASR VS magnitute of corruption (i.e., ϵ )
Fig. 6 : Normalized SΔ VS magnitute of corruption (i.e., ϵ )
Fig. 7 : JVS for different perturbation patterns under different security configurations.
Fig. 8 : Evolution of reliability ( Rtot ) and security ( Stot ) vulnerability metrics under different corruption levels and Split Computing configurations. The first two panels report reliability degradation as a function of Bit Error Rate (BER), while the last two report security degradation as a function of adversarial perturbation magnitude ( ϵ ).
Fig. 9 : Extended JVS joint reliability and security metric for different configurations and target layers. It accounts for both accuracy degradation and confidence score degradation.