>>> which region will get first updates from CDC and how do I conclude this? Can any one help on this!!
No. Absolutely not. Nobody can help with this.
Is this a question originated by a (perceived?) business requirement, a 'nice to have', or just a curiosity?
If it is deemed to be a business requirement to control this, then you need to figure out why you think this is even relevant and then design an other solution such that it is no longer relevant. Perhaps by cascading update from source to A from A to B; Perhaps adding a replicate server local timestamp onto critical rows. Perhaps by using the commit or change time.
Even assuming that the A and B task are exactly the same, the main problem is that there will be two independent source-capture ""retryInterval" loops each at their own uncontrollable start time, with no guarantee/(Qlik)commitment to stick to that. Other activity on the respective servers, the network latencies (including distance as per @Steve_Nguyen ) and even the performance of data storage and the individual target databases can make the actual retry drift from the expected 'loop_start' plus <integernumber> times <retryinterval>.
I could even imagine that sometimes the perceived to be slower task, ends up committing a particular transaction on target faster because it was slow enough to catch that transactions end where the fast task did not yet see the end and is waiting a second or two before seeing if there is more data - including that end.
Cheers,
Hein.
