pa and mpos

[Former Member]

dear all

we are looking at a pilot project for a transformer mfg company and they have 3 lines of business

these 3 businesses will have diff ppl for handling the planning

we hv gone thru and came to understand that they all do the plg at prod / loc / customer level.

the no of cvc's will not be more than 30000. its incremental increase will be around 500 / year for all the 3 lines , at the max.

the KF's for all the 3 lines - hardly any change.

can anyone share , whether i can go for -

1 PA and 1 MPOS

3 PA and 3 MPOS

what criteria should we use to take the above decision ?

regards

KK

[rajkj · Accepted solution]

Hi KK,

With given minimal context, I would go for 1 MPOS, 1 aggregate MPOS, and 1 PA to support the forecast generation at product/location/customer level. 

Your design should consider the business requirements, extent of master data (i.e. CVC) shared to support different planning scenarios, planning aggregation level, support to features like characteristics based forecasting or DP-BOM, reporting needs, performance, and load on the server. More the MPOS and PA in system more will be the load on server and hence performance degrade.

More over, for any pilot or demo, start with one simple setup and extend only if it does not meet the requirements 100%.

Thanks,

Rajesh

[Former Member · Accepted solution]

Dear Rajesh and Ankit

Thanks for your inputs . Helped a lot.

@ Rajesh - For one Planning Area I can have 3 infocubes or one Infocube , no doubt , with 3 data sources. Any advise on this ?

We have done with one Infocube only - though we need to create more Variants in the Process chains.

Thanks in advance.

Regards

KK

[rajkj · Accepted solution]

Hi KK,

Typically your MPOS (both main and aggregates) and PA influence the data source generation and you are allowed to customize the field settings to decide which are to be selectable and hidden. However, your infocube design is entirely depends on the reporting needs. It can pull data from multiple data sources. With proper transformation rules, we can manage most of the requirements (DP related) with one InfoCube.

Depending on the data pull, temporary storage, transformation, and indexing activities, we definitely need more variants or steps in our process chain.

Thanks,
Rajesh

[Former Member]

Hi,

As per your poblem, it will be good to have 1 MPOS, 3 PAs and 3 Infocubes.

Since you have a big amount of CVCs and they will increase every yeat so it will gud performance wise also to have atleat 3 infocubes for each PA.

Thanks,

Ankit Singhal