AI systems increasingly handle sensitive data, making protection of data-in-use essential. Confidential Computing enables this leveraging hardware-based TEEs. Confidential Containers (CoCo) extend this protection to cloud-native workloads. Simultaneously, Arm-based architectures are gaining popularity in cloud environments due to performance and power efficiency.
This session presents an end-to-end architecture for running confidential AI workloads on a Kubernetes environment using Arm CCA. We demonstrate an orchestration of architecture integrating CoCo, Trustee, & secure container image delivery to enable trusted AI inference. Additionally, we highlight Fujitsu’s recent contributions to the open-source ecosystem including enhancements to KubeVirt & our approach to enable CoCo Peer Pods on Arm CCA.
This session is ideal for platform engineers and security practitioners looking to operationalize confidential workloads on next-generation Arm-based infrastructure.