Question
Is there an implementation tutorial for implementing Stream data set in Financial domain?
We are trying to implement real time decisioning for a bank,
One of the scenarios is to count the number of transactions for a customer and impose a fee if the number exceeds a certain count.
We are combating on two different approaches:
1. Call Pega through a service whenever a transaction occurs and determine the count through a strategy.
2. Create a stream dataset and use event strategies to determine the count.
The count can change frequently depending on the business needs, initially the volume of transactions are low.
What is the best approach and the dependencies?
If stream data is the best one, what are the dependencies on the client end?
Any kind of advice is appreciated.
Thanks,
Chetan
***Updated by moderator: Lochan to update Categories***
Hello Anand,
Both options could work but have very different characteristics. When using a standard decision strategy you will have to pass in all the data and pull in the data before the strategy can make a decision. You would be doing the counting using some expressions. If the volumes are low enough, this could certainly work.
Event strategies are designed to handle these types of use cases. You would receive the events via a stream data set, as you pointed out, then do the counting and detection straight inside the event strategy. Only the event strategy management system deals with volume, and you pass on relevant events with lower frequency to the next subsystem - which could be a normal strategy in which other factors are taken into account to determine the next best action.
If the threshold changes, you can have flexibility either way.
But it is hard for us to give concrete advice for particular customer use cases here. We lack the context that is required to make a well informed decision. It might be best that you contact the Pega Consulting team. Please work with your Account Executive who will be able to get this service for you.
Best regards
Otto