LIS25-239 Confidential AI Processing Powered by FUJITSU-MONAKA and Arm CCA

• Thursday, 15 May 15:30 - 15:55
• Room: Session room 3 | Opala III

With AI models handling sensitive data, security must extend beyond data protection at rest and in transit to also protect data during execution. Arm Confidential Compute Architecture (CCA) introduces Realm Virtual Machines (Realm VMs) that provides a trusted execution environment (TEE) with Hardware-backed isolation for AI workloads, memory encryption and attestation to verify the integrity of the realm vm. With Arm CCA, AI inference runs inside a TEE, ensuring that confidential data is processed securely, and AI model remains protected from external tampering. Fujitsu-MONAKA is a next-generation, high-performance, and energy-efficient Arm-based CPU designed for data centres. With a focus on AI acceleration, security, and open-source integration, MONAKA aims to provide a robust and secure computing environment for modern workloads through Arm CCA. In this session, we will showcase how FUJITSU-MONAKA will be leveraging essential technologies to support Confidential AI ecosystem through Linaro-supported QEMU Emulator. The related development work has also incorporated some of the quintessential features related to the implementation of Boot Loader, Secure Monitor, RMM and hypervisor enhancements to support Arm CCA based development. This ensures secure AI model execution with remote attestation for trusted AI processing and provides encrypted input/output data to safeguard privacy.

Add to My Agenda