Can someone tell me or direct me to reading that will let me learn about email sending and delivery?

As I understand things, modern email works like this:

• I type my message and SEND
• My message gets written to a local file along with some sort of local file "work order"
• local software gathers all of the work orders for processing
• processing involves connecting to the target domain
• after connection, the message file gets passed to the target
• after receiving a message file, a server turns some cranks
• some of these cranks pass the message along to other servers
• eventually, one crank causes notifications of the target addressee "new mail"

I know this is a vast simplification of the efforts of "Mail Transfer Agents[MTA]" and other parts of the system. My questions involve elapsed times and permitted delays.

I remember that email began when system operators had to dial phones and connect modems for each hop in the process. The modems relied on telephone lines that were rarely "leased lines" because they were so expensive based on accumulated connection time. This meant that each email hop might require an elapsed work-day.

We pat our feet and drum our fingers after pressing SEND and can mostly expect delivery in a couple of minutes or less. I know that I've received messages saying, "... still trying to deliver ..." or similar.

What determines how long my local system can take to pass a new message along to the first server?

Are there standards for email forwarding and delivery? For example, does an RFC state governing time-to-delivery, when to give up, and such?

Merci d'avance,
~~~ 0;-Dan

Your mail client passes the mail directly to YOUR mail server (the servername configured in your client - typically your ISP's).
It may be passed to a specific server that handles the outbound mail or may do the job itself.
This should happen within seconds.

This server then determines the server responsible for the domain of the recipient (by querying DNS for the MX record for the domain) and attempts to deliver it directly (there may be several more "internally routed" mail servers within the domain for your mail to actually pass through, but the listed MX servers are the only ones that recieve that domains mail directly from the world).

If it gets rejected (typically with a 5xx series error code) or fails to find a server for that domain, it will generate a 'bounce' message to let you know.

If it gets deferred (typically a 4xx series error code) or encounters a temporary DNS error, it will requeue the mail to try again on subsequent queue runs.

The retry period, retry notification time, and give up time is configurable and will vary from server to server.
Commonly, retry is on the next pass, then 10 mins, then 30, then hourly or so.
"Retrying" notification isn't commonly done.
4 days is a common time until it gives up and sends you a FAILED notification.

1 members found this post helpful.

I believe there is still a message-file handoff on my local workstation.
Eons ago, I think things ran on a cron cycle looking for outbound messages.
Today, I think that creating a new outbound message pokes and wakes the local daemon.

Is this "server" is at my ISP or employer rather than at my workstation?

Do you know the RFC or similar "authority" that describes how all this is supposed to work?

Merci,
~~~ 0;-Dan

If you have installed and configured a local MTA and set your client to use it, then yes.
Otherwise, if you enter an external servername in your client, no.
Your client may have a queue or "outbox" or such to temporarily store messages until they are successfully passed to whatever server it is configured to use.
Yes.

No. But teh interwebz produced this, which should keep you entertained for a while.

1 members found this post helpful.

Just what I wanted (grin) a full time job reading ... for weeks ... (laugh). I did ask for it, though.

Cheers,
~~~ 0;-Dan