Hello,

At the moment, the Prosys OPC UA SDK for Java doesn’t provide built-in support for persisting the AddressSpace. If you wish to add Nodes to a Server during runtime and have them stored to memory so that they can be loaded to the AddressSpace when you restart the Server, you will need to implement your own NodeManager that handles storing them to disk. You can use NodeManagerUaNode as a starting point and override methods related to adding Nodes with versions that store the added Node to disk. You can handle loading Nodes from disk in the constructor of your NodeManager.

However, saving Nodes to disk won’t be easy as you would have to solve how to serialize UaNode interface in a way that preserves all the information contained by it and interfaces that extend it. You could use NodeSet2 XML files at OPC Foundation github (https://github.com/OPCFoundation/UA-Nodeset/tree/v1.04) as an example. They are used to model OPC UA information models in XML form and the Prosys OPC UA SDK for Java can use them to generate AddressSpaces when a Server is started. If you serialized your Nodes in a similar form, you should be able to load your stored Nodes when the Server is restarted.

BR,

Matti

Hello,

The functionality for adding Nodes permanently to the AddressSpace by persisting the AddressSpace to disk in NodeSet2 XML format has not been added to the Prosys OPC UA SDK for Java.

But as a clarification, is the use of information models required? If the full capabilities of OPC UA information modelling are not required, then sufficient information on the Nodes could probably be persisted by using some custom XML or JSON format used by the Server application. Though, such format is not provided by the SDK either and would need to be designed by the developer of the Server application.

Do you have a specific use case where persisting Nodes would be required?

Hello,

The definition of “sufficient information” depends on your application, which makes defining that outside the scope of the SDK. At SDK level, it would only make sense to persist all information, but at the moment such feature is not planned. Therefore, you will need to define what information you need to persist and how to persist it by yourself.

Can you be on more specific on what types of Nodes Clients are adding and what those Nodes are used for by the Server and possibly other Client applications?