This annex is intended to provide some guidelines on how to design the address space of an Analyser Server. It covers the following topics:
- Main questions to answer before starting the address space design
- Configuration process
- Parameter definition
- OPC UA key elements
- General rules
This annex should be used as a check list of points to address during the design and the source of description of the rationale behind some elements of the ADI specification. The annex is written in the FAQ format.
This section describes the basic questions that must be answered before starting the address space design.
Is the number of AnalyserChannels constant for this analyser or does it change frequently? For example, the GC has the concept of software AnalyserChannel, and new AnalyserChannels may be added over the time. Is the number of Streams per AnalyserChannel constant for this analyser or does it change frequently? Does the analyzer have a default configuration that allows the user to start the analyser without loading a configuration? This is true for small dedicated analysers that have a single mode of operation. This can also be true if the last configuration is automatically recalled at analyser power-up. Will the analyser continue to acquire data if the connection with the client is broken? Does the analyser server implement access control? What is the access control scheme: user id based, role based? Does the analyser server expose the same Parameters to all users? Knowing that this specification has been developed partly to support analyser deployment in the pharmaceutical industry, how do you plan to support 21 Part 11 regulations? This is not mandatory in the ADI specification, but a good practice to plan for it knowing that it does not require more development, but rather follow some basic design concepts. How do you plan to do the internationalization of the text elements that can be translated? What is your error handling philosophy? How do you report error to the client: through status, events…? What do you put in the audit log? Do you support more than one model of analyzer with this analyser server? Does this analyser server handle more than one analyze simultaneously?