I’ve had no problems extracting 20-60 (and more) million records from ECC 6.0 on HANA using background-enabled RFC extracts.
SAP always seems to recommend using ABAP data flows with RFC Streaming, which certainly does work as well but we’ve had no issues extracting large volumes using the simple RFC Streaming interface as well, nor did we notice any real performance differences. Which really isn’t a huge surprise as both methods use the same output mechanism, that being streaming records out in buckets of 5000 records.
Are you extracting from an on premise based SAP ERP environment into a cloud based SQL Server instance? Or is your SAP ERP environment also in the cloud?
If your ERP is on premise, I would staging the data locally first, preferably in a database otherwise in text files if you have to - and then load into the cloud based SQL instance. At least to investigate where you are running into problems.
Also ensure that your RFC connections are background enabled as they will time out as foreground jobs which such volumes of data.
ErikR (BOB member since 2007-01-10)