Tim Pierce <twp @
> Because VERP delivery requires delivering a separate message body for
> each recipient, the total number of addresses should be just about the
> number of times we transmit a message body over the network. Batched
> delivery permits us to deliver a message body just about once for each
> unique domain on the list, so I counted the number of unique domains as
> approximately the number of times we deliver a message body using
> traditional batch methods.
That's ideal batch delivery. I'd be curious how close to ideal batch
delivery you actually arrive in practice. For one thing, you often won't
be able to deliver more than 100 messages for a particular domain at a
time using most standard MTAs since they limit the RCPT count at that
level, and many sites set an even lower limit for spam-control reasons.
If your estimation counted 1,000 AOL addresses as a single delivery, for
example, that may actually be 10 deliveries or more.
Do you have any way of measuring the effective batching rather than the
theoretical maximum batching?
> So I calculated this percentage for each of the thousand lists, then
> added them all together and took the average percentage. The result was
> an average volume increase of 40%.
This number is going to vary somewhat based on your average message size,
since there's a degree of fixed outgoing bandwidth to do the SMTP
negotiation. VERP will increase that somewhat too (longer return path),
but not by 40%. If you send a lot of short messages, this may be
significant (although it's definitely a lesser factor than the
above-mentioned deviations from ideal batch delivery).
Russ Allbery (rra @