I was looking at the MLM engine described in the link: https://neo4j.com/blog/multi-level-marketing-world-of-warcraft-neo4j/
I am trying to build an engine like this. The tree is mostly binary and I am new to graph technologies. I am trying to change from the old system because of scale issues. The original one was MySQL DB, and we had traditional hierarchical queries running through our system all the time. AS userbase exploded, since this being one giant binary tree with almost near to two million, we started running out of computing. We can’t distribute the system into a micro-service too because the tree is where all the magic happens. So I was looking for a graph implementation of the same. However, how are the time complexities working out in this MLM implementation? I am concerned about that, and the compute load – whether it us linear or logarithmic. And the degrees of freedom I would have in changing the dynamics or conditions of sales compensation – and stuff like that. Would you say, this graph implementation of an MLM engine – is it production-ready?