Does XDCR use persistent connections
when replicating documents from source to destination?
Basically, are you reusing tcp/ip connections when replicating data from source to destination? Or does each HTTP POST of document(s) (CS batches the posts of course) open a new TCP Connection (which is then closed upon completion of the POST?
Can you explain what you mean by "persistent connections"?
Today XDCR is using HTTP request (REST API), to replicates and send keys to remote for conflict resolution. These REST end point are revs_diffs and bulks_doc.
Do you have specific requirements around XDCR?