Deployment Guide
- AI-VAULT can be deployed on a container or a VM.
- AI-VAULT is deployed on the user management node.
- AI-GUARD is deployed on the user host.
- Crypto-fs and KMSAgent are deployed on the running node.
- Solution is provided for AI-VAULT (containerized) deployment. For details, see section "Installation and Deployment" > "AI-VAULT Installation" in the MindX DL AI Model Protection User Guide.
- For details about how to configure independent communication between AI-VAULT and APIGW, see "Installation and Deployment" > "AI-VAULT Installation" > "Network Deployment Scheme".
- Solution is provided for AI-GUARD deployment. For details, see section "Installation and Deployment" > "AI-GUARD Installation" in the MindX DL AI Model Protection User Guide.
- Solution is provided for Crypto-fs deployment. For details, see section "Installation and Deployment" > "Crypto-fs Installation" in the MindX DL AI Model Protection User Guide.
- Solution is provided for KMSAgent configuration. For details, see section "Installation and Deployment" > "KMSAgent Configuration" in the MindX DL AI Model Protection User Guide.
- To deploy AI-VAULT on a VM, you can use tools such as XFTP to upload the AI-VAULT component package to the working directory on the VM and then decompress the package.
- Use XFTP to upload Ascend-mindxdl-aivault_{version}_linux-{arch}.zip to the working directory on the VM.
- Decompress Ascend-mindxdl-aivault_{version}_linux-{arch}.zip to obtain the Ascend-mindxdl-aivault_{version}_linux-{arch}.tar.gz file, decompress the package again to obtain the ai-vault binary file and the dynamic library file on which AI-VAULT depends.
unuzip Ascend-mindxdl-aivault_{version}_linux-{arch}.zip tar -xvf Ascend-mindxdl-aivault_{version}_linux-{arch}.tar.gz - The file directory is as follows:
. |-- ai-vault `-- lib |-- libkmc.so |-- libkmc.so.20 |-- libkmc.so.20.1.0 |-- libkmcext.so |-- libkmcext.so.20 |-- libkmcext.so.20.1.0 |-- libprocwk.so |-- libsdp.so |-- libsdp.so.20 |-- libsdp.so.20.1.0 `-- libsecurec.so