Domain data / reference data implementation?

Question

Fairly simple question, but I have not been able to find a way to do it via the docs.

I am interested in a core set of reference data that all events can be compared against. In the simple example, there are segmentation contexts that talk about the nodeID - but the implication is that different values for the nodeID as they come in fork partitions that can then be used later (eg aggregates).

What I was wondering was if there was any way to pre-populate or pre-register the whole partitioning process via a feed - and avoid a long winded startup process that feeds an event for each different ID in my domain

Answer 1

My guess is that you are looking for data. That is usually one of the big problems to put your hand on. Otherwise, I am not sure what is that you need, but the Proton system can process data off line, like a simulation rather than work just with real-time events. In that case, once you have your data in a file - that can be done. I think the docs should clarify how.

Hope this helps some. - Uri

Answer 2

The paradigm of CEP is to process real time events as they come in. There is no pre-population of the state of agents and context precisely for that reason - it is an event driven architecture, so any piece of information intended for the system should be represented as an event and injected in real time.

You don't have the need to feed manually different NodeIds. All events are segmented based on the value of this attribute...(assuming you have such segmentation context). Therefore if you feed in event intended for aggregation, it would also need and have such a NodeId attribute. A context instance for the particular NodeID value would be opened on such event arrival and aggregation of following events with the same value will be performed.