Replication remains the most efficient way to keep two servers in sync. Apart
from replication, SQL talks to one server at a time.
To run an SQL comparison of versions of a table from two different servers, you'd need to copy a table from server B to server A (eg mysqldump it from B, then import it to A), then on server A run a table-comparison query like that described under "Compare data in two tables" at
https://www.artfulsoftware.com/queries.php.
Or, though it'd be slower, you could in a language with a MySQL API, eg PHP, connect to server A, make a second connection to server B, load the table from B into memory, and write a module in that language to compare the in-memory version with the table on server A.