Visible to the public Confidentiality and Authenticity in a Platform Based on Network-on-Chip

TitleConfidentiality and Authenticity in a Platform Based on Network-on-Chip
Publication TypeConference Paper
Year of Publication2017
AuthorsSilva, M. R., Zeferino, C. A.
Conference Name2017 VII Brazilian Symposium on Computing Systems Engineering (SBESC)
ISBN Number 978-1-5386-3590-2
Keywordsadvanced encryption standard, AES model, authenticity, Computer architecture, confidentiality, cryptography, Encryption, low-cost interconnect architecture, Many-core systems, Metrics, Multicore Computing, multicore computing security, multiprocessing systems, Network interfaces, network level, network performance, network-on-chip, Networks-on-Chip, processing elements, Program processors, pubcrawl, reference multicore platform, resilience, Resiliency, Scalability, security, security aspects, security mechanisms, security properties, silicon overhead, SoCIN-based systems
Abstract

In many-core systems, the processing elements are interconnected using Networks-on-Chip. An example of on-chip network is SoCIN, a low-cost interconnect architecture whose original design did not take into account security aspects. This network is vulnerable to eavesdropping and spoofing attacks, what limits its use in systems that require security. This work addresses this issue and aims to ensure the security properties of confidentiality and authenticity of SoCIN-based systems. For this, we propose the use of security mechanisms based on symmetric encryption at the network level using the AES (Advanced Encryption Standard) model. A reference multi-core platform was implemented and prototyped in programmable logic aiming at performing experiments to evaluate the implemented mechanisms. Results demonstrate the effectiveness of the proposed solution in protecting the system against the target attacks. The impact on the network performance is acceptable and the silicon overhead is equivalent to other solutions found in the literature.

URLhttp://ieeexplore.ieee.org/document/8116584/
DOI10.1109/SBESC.2017.38
Citation Keysilva_confidentiality_2017