Protection of the host PC
Customers are responsible for protecting their own host PCs and preventing unauthorized access to their host PCs.
Siemens suggests taking the following measures to protect the host PC:
Deploy the host PC only in isolated plant network, but not office network.
Enable the screen saver and lock the screen when leaving.
Install suitable anti-virus software.
Install updates and patches for the operating system and software on the host PC in time.
Notes on protecting administrator accounts
A user with administrator rights has extensive access and manipulation options available to the system.
Therefore, ensure there are adequate safeguards for protecting the administrator accounts to prevent unauthorized changes. To do this, use secure passwords and a standard user account for normal operation. Other measures, such as the use of security policies, should be applied as needed.
Notes on metric names
Metric names should consist only of characters accepted by Prometheus (ASCII letters, digits, underscores, and colons), and may contain spaces between words. In addition, they can contain spaces between words. For further information, refer to the related note in the section about metric names.
Notes on secure communication
AI Model Monitor and AI Model Monitor Agent communicate over HTTPS. The applications are shipped with predefined templates that ensure secure communication with almost minimal configuration. The AI Model Monitor and AI Model Monitor Agent applications have been scanned for viruses by an industry standard virus scanner in the course of generating the executables.
Notes on safety critical applications
AI Model Monitor and AI Model Monitor Agent are NOT intended to be included in the control flow of industrial automation. The intended use is limited to supervise the operation of ML pipelines and some aspects of the underlying software and hardware. For example, you MUST NOT feed the output of AI Model Monitor back into a PLC to influence machine actions on the shopfloor.
Data collection in AI Model Monitor is not real time, therefore no alerting in Safety Critical applications should be based on data collected by AI Model Monitor.