SAP Datasphere vs Others (GCP, Azure, etc) - Business Case ??

[pierre-yves_guillou]

Dear all,

I hope you have a very nice day! Sorry for the long question below. I am trying to set up the context and get some other inputs than what I see at the moment.

I am currently working with a client trying to make his way in the great world of Data. Basically, the situation is that the core processes of a Business Area will run with SAP (target is 1 instance and not 3 as today, ECC for now, but S/4 in the near future), and other specific applications. The implementation of SAP involved the implementation of a BW on HANA system (Yes, I know...not the latest stuff). But instead of feeding the BW through regular datasource extractions, they leveraged an SLT system to be near real time for most of their reports sourcing data from their current various SAP landscape. So, they have "rebuilt" lots of calculation views to support their business users. They also have many front ends (BOBJ, PowerBI, Spotfire, and other specialized applications). The maturity of BI is still, not low, but more oriented on checking consistency, very operational and less strategic than it could be. 

So we have : ECCs -> SLT -> HANA DB -> CalcViews -> Composite Providers (for managing authorizations)

They are at a stage of the implementation where the question has been raised to move to SAP Datasphere or to another data manager provider (something like Azure, GCP, Snowflake). The question is raised because they do not have only SAP data, and the overall group (other business areas are in the group) direction is to go to a non SAP landscape for data analytics.

The current exercise is to analyze the options, and choose the most long term viable one: staying with SAP analytics solutions, or moving away. In my current analysis, I am struggling because I don't find any "automatic" ways to convert/migrate the calculation views from BW on HANA directly to SAP Datasphere, provided by SAP. 

I have found the HANACV2SQL software, but I would have had expected SAP to support that kind of scenario. 

So, my question is : What could be the arguments to keep SAP Datasphere when:

• what ever happens, you will need to convert using the same tool at the same costs ?
• connectivity to front ends will be subject to reinstallation
• row level security needs to be reimplemented 
• SAP Datasphere restricts "data replication" to other systems if the data is used for other applications.

I have been working in the SAP ecosystem for a long time, and the Analytics strategy has always been hard to implement with SAP. The 0 copy concepts and the sharing capabilities promoted by SAP Datasphere (even the partnerships with Snowflake and GCP) are not clear form my point of view. Even if the client is SAP Centric, you need dedicated resources on the technology, which could lead to higher costs...

Any arguments that you have seen promoting one side more than another ?

Thanks for reading and sharing 😉

Cheers