I. INTRIDUCTIONThis paper presents an analysis of the effects of Single Event Upset (SEU) into the internal memory of Aeroflex Gaisler LEON3 processor which is a 32-bit synthesizable processor based on SPARC V8 architecture implemented in an FPGA. A new software methodology allowing fault injection is explored and illustrated in order to classify the faulty behaviors while executing an AES benchmark. An exhaustive fault-injection campaign was performed to test the behavior of LEON3 processor.
II. EXPERIMENTAL RESULTSThe emulation of SEU faults in data SRAM memory was done at each AES execution. The SEUs were injected only in SRAM area used by the program. It is important to note that:-Only one SEU per execution was injected.-The SRAM occupied area was exhaustively explored.It is well-known that SEUs do not permanently damage a circuit, but they can cause faulty behaviors. In the worst case, the processor will have to be reset or reconfigured to overcome the effects of a soft error.Presented table summarizes the results obtained of the exhaustive SEU injection campaign performed in the SRAM occupied area of the studied LEON3 architecture.The consequences of the injected SEUs in data SRAM memory of LEON3 can be classified as follows:-Silent fault: The algorithm provides a correct result despite the fault.-Result error: The algorithm provides a false result when the errors are injected in the data SRAM memory that contains a data or the result of AES.-Timeout: the application does not produce any result or does not terminate. The program execution duration exceeds a predefined limit, which explains in our case a power down mode of LEON3 or a Debug Support Unit (DSU) error.-Stopped mode: the application terminates abnormally with error indication. The algorithm doesn't provide any result. This can be explained by fault leading to the reset trap of LEON3, a memory access violation or an invalid instruction.The size of data SRAM occupied by the AES program is 3.6 Kb. For being exhaustive, 29858 SEU faults were injected as shown in the first column of the following table. #SEU #errors #silent-mode #timeout #stopped-mode 29858 90 29726 3 39The analysis of the obtained results can be summarized as follow: -As expected most of the injected faults, 99.6%, did not perturb the AES execution.-Even if the stack pointer (SP) and program counter (PC) were not targeted by the injected SEUs, a few number (0.14%) of timeouts and stopped-mode errors were observed. This can be explained by SEUs injected in variables used to initialize the AES's counters.-0.3% of injected SEUs lead to erroneous output of AES. This kind of error is very critical for the AES algorithm. Indeed, the faulty result does not allow retrieving the information in decryption mode.Results of fault analysis will be used in future work to propose a fault-tolerant version of LEON3 processor. To implement an SEU-tolerant processor on a non-hardened semiconductor process, several error-detection and correction techniques can be applied such as TMR (Triple Modular Redun...
Embedded system design is characterized by its complexity day by day, combined with reduced time-to-market deadlines. An embedded system integrates together a hardware and software parts on a common platform.The Advanced Encryption Standard (AES) is a cryptographic algorithm that can be used to encrypt a huge bulk of data and for embedded system security. This paper, explains the characteristics of leon3 Soft Core processor, and its implementation on FPGA board. We have implemented AES algorithm for both encryption/decryption by using C language and leon3 processor. The algorithm is implemented with different key of 128, 192 and 256 bit length. Our proposed software implementation is very efficient for encryption/ decryption functions in leon3 embedded processor. It can be used in low cost applications.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.