Evaluating the Code Encryption Effects on Memory Fault Resilience

Cantoro, Riccardo; Deligiannis, Nikolaos I.; Reorda, M. Sonza; Traiola, Marcello; Valea, Emanuele

doi:10.1109/lats49555.2020.9093670

Cited by 7 publications

(5 citation statements)

References 8 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Specifically, for fault multiplicity higher than 200 the block ciphers' detection rates drop. These results show a contrasting trend compared to our previous work [6], where we found that the block ciphers have a high detection capability for faults occurring in the code memory space (see Section II). Indeed, a fault may transform an instruction into an illegal one, causing a failure in the code execution.…”

Section: A No Pad: Experiments Without Padding Checkscontrasting

confidence: 96%

“…Block ciphers perform better than the stream ciphers. In this scenario, we obtained a trend similar to the one observed in our previous study [6]. Finally, regardless of the multiplicity of the injected faults, the PCBC configuration achieves ≈ 100% detection rate.…”

Section: B Pad: Experiments With Padding Checkssupporting

confidence: 86%

“…Such an effect opens some opportunities for fault tolerance mechanisms. In a previous study [6], we showed that augmenting the corruption of an application code leads to increasing the fault detection probability for single faults and especially for multiple faults. This interesting property can be useful to optimize the utilization of costly ad-hoc mechanisms to detect faulty conditions.…”

Section: Introductionmentioning

confidence: 93%

“…We previously evaluated the error detection capabilities that specific encryption mechanisms can provide when applied to the application code, stored in NVMs [6]. In that context, a fault may turn an instruction into an illegal one, causing a failure in the code execution.…”

Section: Preliminaries and Specificationsmentioning

confidence: 99%

See 3 more Smart Citations

Evaluating Data Encryption Effects on the Resilience of an Artificial Neural Network

Cantoro

Deligiannis

Reorda

et al. 2020

2020 IEEE International Symposium on Defect and Fault Tolerance in VLSI and Nanotechnology Systems (DFT)

View full text Add to dashboard Cite

Nowadays, many electronic systems store valuable Intellectual Property (IP) information inside Non-Volatile Memories (NVMs). Therefore, encryption mechanisms are widely used in order to protect such information from being stolen or modified by human attacks. Encryption techniques can be used for protecting the application code, or sensitive sets of data in the NVM. In particular, in machine-learning applications, the weights of an Artificial Neural Network (ANN) represent a highly valuable IP stemming from long time invested in training the system along the development phase. On the other side, systems implementing ANN applications are increasingly used in safety-critical domains (e.g., autonomous driving), where a high reliability level is required. In a previous paper, we have shown that encryption techniques, applied to the application code of generic systems, provide a significantly higher error detection rate. In this paper, we focus on an ANN application and we evaluate the detection rate induced by encryption mechanisms for transient faults possibly impacting the ANN weights. We performed experiments on a pre-trained ANN, whose weights represent the sensitive IP of our system. We executed fault injection campaigns to evaluate the ANN resilience when different encryption methods are used. Experimental results showed that the presence of specific encryption mechanisms alone induces high fault detection rates in such applications. This may allow the designer to consider security and safety mechanisms together, achieving the same results with lower costs.

show abstract

Section: A No Pad: Experiments Without Padding Checkscontrasting

confidence: 96%

Section: B Pad: Experiments With Padding Checkssupporting

confidence: 86%

Section: Introductionmentioning

confidence: 93%

Section: Preliminaries and Specificationsmentioning

confidence: 99%

See 2 more Smart Citations

Evaluating Data Encryption Effects on the Resilience of an Artificial Neural Network

Cantoro

Deligiannis

Reorda

et al. 2020

2020 IEEE International Symposium on Defect and Fault Tolerance in VLSI and Nanotechnology Systems (DFT)

View full text Add to dashboard Cite

show abstract

“…We perform fault injection campaigns only on the ML application weights and not on other ML data or application code. Concerning the application code, in a previous work [40] we have shown that encryption enables high fault detection rates. We classify faults as follows:…”

Section: A Fault Models Fault Classification and Fault Effectsmentioning

confidence: 97%

Towards the Integration of Reliability and Security Mechanisms to Enhance the Fault Resilience of Neural Networks

Deligiannis¹,

Cantoro

Reorda

et al. 2021

IEEE Access

Self Cite

View full text Add to dashboard Cite

Nowadays, many electronic systems store valuable Intellectual Property (IP) information inside Non-Volatile Memories (NVMs). Designers widely use encryption mechanisms to enhance the integrity of such IPs and protect them from any unauthorized access or modification. At the same time, often such IPs are critical from a reliability standpoint. Thus, dedicated techniques are employed to detect possible reliability threats (e.g., transient faults affecting the NVM content). The weights of a neural network (NN) model (e.g., integrated into an object detection system for autonomous driving or robotics) are typical examples of precious IP items. Indeed, NN weights often constitute proprietary data, stemming from an extensive and costly training process; moreover, their correctness is key for the NN to work reliably. In this article, we explore the capability of encryption mechanisms to ensure protection from both reliability threats. In particular, we assess, via extensive fault injection campaigns, the capability of different memory encryption schemes -usually used only for security purposes -to detect faults and thus, enhance the reliability of the system. Experimental results show that by cleverly choosing the proper encryption scheme, it is possible to achieve very high fault detection rates (greater than 99%) with respect to Multiple Bit Upsets. The gathered results pave the way to the integration of reliability and security mechanisms to achieve better results with lower costs.

show abstract