>How much traffic does a master/slave configuration produce beyondEssentially, it seems to be the queries that changed something>the inserted/updated data?
sent from the master to the slave. The size of the binary
logs on the master is a good indication of this. If, for example,
you increment a field in *EVERY* row, it sends the query (which is
likely much smaller), not a copy of every row.
There's the usual TCP overhead. And the slave does have to log in>As to say: How much "meta"-traffic is to be expected?
on the master.
It sounds like you are dealing with traffic that could easily be>It's a very small database with around 20 tables, each around
>5 - 10 fields, evenly spread varchars and numerics.
>There are about 10 inserts and 50 updates a day.
handled over a 300 bps (that is, really ancient and very slow)
modem. At about 100 characters per query, and double it for overhead,
that's about 1.1 bits per second (not bytes, not kilobits, not
megabits, just *bits*). It sounds like this could easily be handled
with carrier pigeons or mailed floppy disks.
Gordon L. Burditt