Are there any guidelines as to what a 'normal' throughput for replication is using AlwaysOn?
I have a 100 megabit connection that is able to replicate over 1000 databases using log shipping, however the same connection frequently gets a backlog for a system using AlwaysOn.
The connection has a ping time of approximately 200ms, however can always achieve 100 megabits when pushed using UDP.
Are there any settings to increase the number of TCP connections, or can communications between replicas use UDP instead?