Application Security Enablement#

Edge AI is revolutionising multiple AI usage areas, raising critical challenges in cybersecurity. With AI systems making autonomous decisions while handling sensitive city data, facilitating production lines, monitoring human health, and so much more, robust security is a must. Intel platforms provide built-in security features to make it possible.

This tutorial provides steps and guidance to enable built-in Intel security features, using the Metro AI Suite Smart Intersection& Smart Intersection Agent applications as an example. With some adjustments, you will be able to equip your solutions with:

  • Comprehensive Data Privacy: Protect sensitive personal data with strong encryption protocols and blockchain technology, ensuring secure transactions and compliance with regulations.

  • Infrastructure Protection: Safeguard your critical infrastructure with regular firmware updates and advanced threat detection systems, mitigating vulnerabilities in distributed edge devices.

  • Secure AI Processing: Leverage privacy-by-design principles and hardware-level security features to prevent AI model manipulation and ensure secure, compliant AI operations.

This guide covers:

  • IFWI Flashing Guide: Flashing an Integrated FirmWare Image (IFWI) guide.

  • UEFI Secure Boot: Secure Boot technology maintains system integrity through a verified boot chain, preventing devices from launching unauthorized software. On Intel platforms running Ubuntu OS, the secure boot process initiates from a hardware root of trust, guaranteeing that only trusted operating systems can boot.

  • Full Disk Encryption (FDE): FDE technology prevents unauthorized access to disk data, particularly in scenarios involving device theft, loss, or physical compromise. Additionally, FDE keys are secured using hardware-assisted TPM/Intel PTT, ensuring that device data can only be decrypted by trusted software.

  • Total Memory Encryption (TME): TME technology provides runtime protection for sensitive algorithms and detection models by safeguarding against cold boot attacks and physical threats to the memory subsystem. The TME key is a hardware-generated ephemeral key that cannot be accessed by software components.

  • Trusted Compute (TC) deployment for secure, isolated execution of AI workloads: Trusted Compute is a set of software-defined security extensions that leverage hardware security capabilities of Intel platform. These extensions provide enhanced protection for AI workloads through runtime isolated execution environments.

These security features protect AI applications, such as the Smart Intersection system, from unauthorized access and ensure data integrity for traffic monitoring and analysis.

By leveraging Intel’s cutting-edge security solutions, you can harness the potential of Edge AI and Agentic AI while blocking potential threats. This paves the way for smarter and safer solutions, where AI-driven threat detection and digital twins for cybersecurity scenarios continue to evolve. Trust Intel technologies to meet the demands of the modern AI era, such as affordable and secure protection of your transportation infrastructure.

This Page