The XXXEncoder and XXXDecoders within package com.prosysopc.ua.stack.encoding.XXX purpose is to encode/decode classes that have (and are used in) OPC UA Part 6 defined encodings. They are not something that would work on UaNodes (nothing is). They would work on the Value Attribute’s value of the node though, since they are something that have Part 6 encodings defined to them.

Using UaNodes requires in practice that all of them can be kept at memory all times in the current SDK design (i.e. the references are directly in the nodes referencing other nodes), thus using the MyBigNodeManager-style is basically the only way at the moment for you (since it wont use UaNodes and instead you must provide Attribute and Reference data “somehow” when requested by the SDK). Preferably at some point in the future we can have a better way to combine the 2 strategies.

The current primary purpose for the JsonEncoder is in conjunction with our PubSub MQTT+JSON support. Though in general it should still work the same way as BinaryEncoder for the Part 6 encoding purposes.

Also note that we do not (and wont have in a while) have a counterpart JsonDecoder, because building one is hard for our current architecture for the encoder/decoders. Basically before the JSON encoding was defined in OPC UA 1.04, the Binary and XML have clearly defined sequential ordering for the Structure fields (the order they are defined). However, JSON by the very nature of JSON means non-ordered set of key-value parts, thus it is not suitable the way we do it currently. Since any order is allowed, we can encode them in the sequential order, just that we would have no guarantee that we would receive such JSON back.