I am currently testing the effects of worst-case latency and throughput on the ePO to SADR Full Replication task.
My topology includes: ePO+DB---vSwitch---NetworkImpairmentTool---vSwitch---SADR
I have manually set the throughput between ePO and SADR to 1 Mbps and the latency to those listed below:
|Tested Latency (one-way)||Ping RTT||Duration||Success?|
|0 ms||~2 ms||3h 27m||Yes|
|150 ms||~302 ms||5h 54m||Yes|
|300 ms||~602 ms||9h 05m||Yes|
|325 ms||~652 ms||2h 17m||Fails @ 30%|
|350 ms||~702 ms||3h 31m||Fails @ 30%|
|410 ms||~822 ms||4h 03m||Fails @ 30%|
|450 ms||~902 ms||4h 25m||Fails @ 30%|
So here are my questions:
1.) What is the reason, from an application standpoint, that this failure would occur at latencies over 300 ms?
2.) What can I do to make the ePO more tolerant of high latency and low throughput?