DROOLS规则引擎：在Drools中进行网络调用

Question

One of the transactional systems i work upon deals with data-gathering from multiple external systems, and making some business decision based upon the (transactional entity + gathered-data).

Currently , we follow the following steps:

1. Gathering data from multiple sub-systems.

2. Using (Gathered data + transactional entity) as an input to drools, and deriving a business decision out of it.

One major cons of the above approach is that i have to gather all the data beforehand (expensive network calls) without even bothering about the usefulness of the data.

What i am trying to do is to delay the service calls by moving it to the rule execution layer. The intention is to leverage the drools decision tree to avoid making a service call if i can take a decision on some already available data (within my transactional entity itself).

Just want to get the thought process validated (ie, making a service call as part of the rules execution would be a good practice or not.).

Can anybody please share the pros/cons around the same. Any leads would be appreciated

Answer 1

You cannot reason with facts you don't have (yet). If the facts you do have let you decide what else you need you can write rules according to the following trivial example.

rule "x and y, not z"
when
    Dimension( $id: id, coord == "x" )
    Dimension( id == $id, coord == "y" )
    not Dimension( id == $id, coord == "z" )
then
    fetch and insert missing Dimension
end

rule "x, y and z" ... end

Note that this process of fetching and inserting must not be executed synchronously. You can insert the request as an event queue - use anything from Java.