Skip to Main Content
Jedox Ideas

Let us know how we can make Jedox even better!

Status Future consideration
Workspace Jedox Platform
Categories In-Memory DB
Created by Guest
Created on Feb 28, 2023

Shared Dimensions / rule performance between different databases

The concept of Jedox models is really great to split a big EPM application
into different manageable modules (modularization).


Example:

You could have 3 different models for planning

  • sales of products,

  • production of products

  • purchasing of material for the products

Each model could have its own database with dedicated cubes and logics.


However there is one big drawback in the concept:

For those example models you need to have some dimensions in all 3 model databases.
-> e.g. you need the same product dimension in sales, in production, in purchasing.

So you would need to care on your own about synchronizing this dim across all db .
(e.g. if you create new plan products)


And a much bigger disadvantage:

If you try to setup rules between cubes in those 3 different model databases to connect different information parts, the performance will go dramatically down.

The processing is unfortunatelly much slower than having all cubes in one big database (with only one product dimension).

--> e.g. for a rule which combines production costs with a sales price for a product
( rule: production cube <- -> sales cube .. by product)


But there could be a solution which I would call "Shared Dimension" (related to similar concepts in othere BI systems).


In this concept one dimension could declared as "Shared Dimension" and therefore to be used across different databases. (with same internal IDs and so on).

I guess this would increase performance and consistency and support directly development best practices in terms of modularization, scoping, encapsulation and so on.

What do you think?

Thanks a lot for checking this idea.

  • ADMIN RESPONSE
    Mar 3, 2023

    Thank you for your great idea. While we cannot commit to a specific target version, we will consider adding it to our future internal backlog. If you can add further information about the context of this feature, please add it here so we can evaluate it more fully.