CI-DS vs CI - MasterData Import

[#1]

Hello Experts,

Hope you all are doing great 🙂

In mid of so many SAP rebranding and innovations I felt to play around with the performance when it comes to On-prem to Cloud connectivity.

"The agenda is to find the best technology among CI and CI-DS respecting their full potential and current possibilities for IBP relevant scenarios."

To be precise as part of my use case I had been trying to compare the Integration performance for CI vs CI-DS reading from Oracle 12c on-premise DB within enterprise network fetching ~76000 rows/records

• In CI I use select sql query via JDBC receiver and then push it to IBP via OData v2 api

  

  

• In CI-DS using native simple dataflow using oracle and IBP datastore

  

However as I notice there is considerable (almost double) advantage CI-DS (58-90secs including IBP postprocessing) have over CI (2min 15sec in bulk and 1min 30-45secs in split excluding IBP postprocessing).

I agree the possible reasons can be

• CI-DS Agent's positioning on-premise
• Oracle installed in the same system as DS Agent

Are there any other explanations or suggestions for performance improvement with JDBC polling/ receiver streaming/OData?

However I am currently trying to test it differently to gain better performance results from CI. Feel free to share your thoughts or suggestions.

[SandeshK]

In my experience, the best option for replicating data into IBP or extracting data from IBP is CI-DS. The tool's main use-case is also for IBP. Although using CI you can replicate data into IBP using the OData APIs, it is not recommended by SAP to use OData APIs of IBP for mass uploads.

Also generally CI-DS is hosted at the same datacenter of IBP, so you will also see an advantage with low latency. Probably one of the reasons why you see low processing times when compared to CI.

[Jitendra_Jeswan]

Hello Folks,

Without going too much into detail..

I must say that CI-DS should be used for "Data Integration" scenarios with use cases like Data Migration, Replication etc. where considerable data volume is considered. Moreover it supports many ETL relevant functions.

CI is meant for "Process Integration" scenarios mainly for transaction processing & cover use cases like A2A & B2B integration. It does not support ETL relevant functions but can be tweaked to address some of them.

Having worked with both the services, I can undoubtedly recommend to use CI-DS for your use case scenario, no matter whether it is IBP scenarios relevant or not..

[Irmi_Kuntze]

Honestly I don't have a clue about CI, just know a tiny bit of CI-DS:

In CI-DS the oData V2 is not enabled for parallelization and hence not recommended for mass load. Maybe you acn check whether this applies as well for CI ? Have you tried V4?

[Aman_V]

Hello irmhild.kuntze,

Thanks for your comment.
There is only OData v2 Api available to connect to SAP IBP and is being used in CI to connect to the IBP system and further master data staging table.

For CI-DS there is a direct connectivity to the IBP system and further master data staging table via Datastore connection.

The agenda is to find the best technology among CI and CI-DS respecting their full potential and possibilities for IBP relevant scenarios.