sync with backend and sync with client

[Former Member]

I am new to MI7.1 and still trying to understand the sync concept .

I know the data sync between MI client and MI.

How does the MI sync with the backends?

Please help. Thanks a lot!

[former_member304703 · Accepted solution]

Hi Jeff,

MI has two separate processes: synchronisation and replication (called "data load" in MI7.1).

1. Synchronisation - is a process of downloading data to the device. Data can come either directly from backend (SyncBos type S01, or transactional data, not used in MI7.1 anymore) or from the replica DB on the moddleware (SyncBos type T01 and T51). Synchronisation in MI is always triggered by the device and is asynchronous, which means that device always sends two messages to the middlware. First message is request for data, middleware processes this request and puts results into qeues. Next time device synchronises results from qeues are downloaded to the device.

2. Replication or Data Load - is a process of maintaining replica DB on the middleware up-to-date to the backend data state. Replication is always triggered by the middleware and normally is triggered as a scheduled job. It is crutial for this process to understand the differnce between T01 and T51s. With T01 middlware always calls getlist and getdetail BAPIs for the whole list of data. Which means that all the data is compeletely replaced, which is not very efficient when there is a lot of objects in the list. Hence, comes T51s, so called backend (or server) driven SyncBos. When data is changed on the backend, backend is making an RFC call to the middleware and puts a key of changed object inot a table. When middlware triggers replication only getdetail BAPIs for keys in that table are executed - so only data objects that has changed since the last replication are replicated in the middlware replica DB.

Regards,

Larissa Limarova

[Former Member · Accepted solution]

Thanks!