Bidirectional data replication with SDI

[cwolter90]

Hello SAP Community,

we are currently searching a way to implement a bidirectional data replication between two SAP systems with HANA as their database. We want to replicate a limited set of tables of the ABAP system, so the source and the target table are DDIC table, that are managed by the SAP systems.

SAP provides the Learning Journey Provisioning Data to SAP HANA for this topic. We implemented a replication with SDI as a proof of concept. Sadly we discovered, that the replication only work from source to the target table. The backwards direction only works for the virtual table, but not for the target table of the remote subscription (which is a DDIC table).

The learning journey says "SAP HANA data provisioning technologies do NOT restrict the data flow to a single direction (source to target), that is, reading from source system. Depending on the type of source, some scenarios might include writing to the source system." and there is also a picture, that shows arrows for data flow in both directions:

(Source: https://learning.sap.com/learning-journeys/provisioning-data-to-sap-hana/defining-access-to-remote-s...)

The SDI documentation also doesn't clarify, how the "Write-back Functionality" works.

Does anybody know more details about the "Write-back Functionality" of SDI or bidirectional replication?

Best regards
Christian

[xiaoming_wu]

 Hi Christian,

That depends on your replication, are you going to replicate the same table ?
For example, one with A as source and B as target.
Then the other with B as source and A as target? 

Best regards
Helen

[cwolter90]

Hi Helen,

we want to replicate the same table between both systems.

[xiaoming_wu]

Why you want to replicate the same table? If both system will update the system table and you would like to sync the information, then you might see data inconsistence.Usually SDI replication is used to sync the data between different systems.

[xiaoming_wu]

Hi Christian,

SDI does support replication to different target.
For example:
1. Setup DP agent connect to HANA A
2. In HANA A, create two remote sources to connect to HANA B, C
3. Create VT table connect to HANA B, C.(VT_B, VT_C)
4. Setup replication task/flowgraph to do replication from VT_B to VT_C
So you can see in the example above, use VT table as target in your replication, then you can move data to target other than the current hana system.

But keep in mind, performance to other target is not good. But we do provide such function.
In your case. If both the source and target are HANA database, my suggestion is you setup two setup SDI.
Two dp agent connection to two HAHA, within each other setup remote source to connect to the other to replicate data.

Best regads
Helen

[cwolter90]

Hello Helen,

so your advice is to setup SDI in both systems and create a replication with opposite direction? Will this cause an infinite loop?

Best regards
Christian