Hi -

So assuming 1 MB is intercepted by the Splitter and sent to the source RPA, the RPA receives the write and applies any link policies (Write folding, compression, dedup), then the RPA transfers this write (along w/ other writes part of the snapshot - which also has a configurable granularity, default is Dynamic) to the remote RPA. Once the remote RPA receives the write(s), it writes it to a stream in the copy Journal called the DO stream. It then reads it from the DO stream to memory and also reads the data from the actual replica and writes it to the undo stream, finally the write is written to the actual replica. This is a normal 5-phase distribution.

Hope that helps.

-Zahid

RP Corp Sys Eng.

Thanks Sahid, however, I'm still unclear on where the writes occur.  Can you please be more definitive in your description.  Writing it out step by step in the same format I did above would help.

This should help DrWhy - check out the RecoverPoint Admin Guide...it will detail all of the write flows in different scenarios. This illustrates the normal distribution phase that Zahid talks about above.