Hi,

You are on your own. As seen in the README:
“This repository is provided by OPC Foundation as legacy support for an Java version for OPC UA. It will not receive further features and updates.”

Though in this case note that how long the write takes is basically entirely dependant on the server, so in practice no change in the client side would help you. A server implementation might e.g. handle each write individually even if the request was a bulk one.

But anyway, as of our SDK version 4.x we no longer depend on that lib, see https://downloads.prosysopc.com/opcua/Prosys_OPC_UA_SDK_for_Java_4_Release_Notes.html#version-4-0-0 “Changes to OPC UA Stack” section. Basically this happend due to the fact the OPC Foundation decided to discontinue all Stacks (keeping only the .NET, as it was more advanced; since with just a Stack you really cannot do much). Basically previously the Stack for us was a very low-level API on top of which our SDK built all the other features you will need in a proper OPC UA Application. Now with 4.x there is no longer a “Stack” and “SDK” separately as far as our SDK is concerned (see the release notes). That is to say, in order to make some parts more useable, some “stack level” APIs now do refer “SDK level” APIs (e.g. Structure with it’s StructureSpecification etc.). Though the fact that there is anymore even a “com.prosysopc.ua.stack” package was to ease the migration process of SDK 3.x -> 4.x.

So in general I would recommend you to buy/try our Prosys OPC UA SDK for Java: https://www.prosysopc.com/products/opc-ua-java-sdk/, as it is true SDK level library compared to the Stack levels and also in this case a well maintained lib :). See this video from the OPC Foundation which explains the differences: .

Though, I would estimate that it would not be any faster even with our SDK due to the server side being the limiting factor here. You should be able to use https://www.prosysopc.com/blog/opc-ua-wireshark/ to verify this (i.e. look how long the server takes to respond with the WriteResponse).

No idea, I’m not an expert in PLC-related things. Everything depends on everything. Depending on how much you are actually writing at once and what kind of data anything can be a bottleneck. Also unless you are running your client application on a realtime OS using a realtime JVM, in general with Java timings should not be that much of importance anyway (though of course performance is important). I would need to see some numbers to have any better guess. Though it should be noted that in general we do not give support for the “stack layer” even with the SDK, since you are supposed to use the SDK APIs (but in some cases we do not have “higher-level” one).