|
|
|
By Dr.
Philip Baczewski, Director of Academic Computing and User Services
E-mail was one of the first
applications to be used with the Internet and other wide-area
networks. It is a staple of Internet communication and something we
take for granted. At least we take it for granted until something
changes, such as our e-mail service, our e-mail client or our e-mail
address.
Getting a new e-mail address these days has the kind of impact that
getting a new phone number used to have. In the old days, you'd have
to phone up the people you wanted to keep in touch with or send out
post cards with your new contact information via the U.S. mail. Of
course, you can still do so, but today you'd probably just send out
e-mail.
Most people tend to define their e-mail service by the program they
use to receive and send their messages. In the bigger e-mail picture,
however, that program is only one part of the system. There are a
couple of other key components that accept your messages and send them
to their intended destination as well as accept messages for your
address and route those messages to your e-mail mailbox. Your mailbox
is stored on another part of the system. Still another part of the
e-mail picture lets your client program retrieve and store messages in
your mailbox. It all seemed so simple when e-mail was just your
favorite client program.
Behind the scenes ...
As an attempt to explain things and understand a bit more about how
e-mail works, let's start at the top and work our way down to your
client. Mail received from the Internet comes to a mail transfer agent,
or MTA. There are a number of MTAs that can be run to send and receive
e-mail between Internet sites, and interestingly enough some of the most
popular ones are open source or licensed for public use. Some people
think that there's some Internet mail sorting location, like they have
at the U.S. Post Office, where e-mail is sent and then routed on to its
final destination. In actuality, the MTA connects directly with other
MTAs once it's determined where the e-mail should be sent.
When someone sends a message to an address at unt.edu, their client
drops off that message with their MTA (usually an SMTP server). The
remote MTA consults DNS (Domain Name Service) to find out what server
handles e-mail for unt.edu. In our case, that server is mailhost.unt.edu.
The remote MTA then opens an Internet connection directly to Mailhost
and transmits the message. Mailhost (UNT's MTA) then sends the message
on to the mailbox server associated with the recipient e-mail's address
usually by sending it to another MTA which provides mail service for a
department or application.
So, if there is an e-mail sorting station, your local MTA comes closest
to being one, since it accepts many messages, both incoming and
outgoing, and transfers them to the right destination. The MTA also
handles mail alias translation for incoming and outgoing messages. If
you have an address of "name@unt.edu" any messages coming in for that
address will be routed to the correct mailbox server. Any message going
out from an on-campus mail system will have the from address "rewritten"
using the "@unt.edu" address.
Before you can get an incoming e-mail message, it must be delivered
to a server that maintains your e-mail mailbox, sometimes known as your
"inbox." GroupWise and Exchange are both examples of such a server. They
also allow you to create mail folders on the server to let you store or
organize your e-mail. While GroupWise and Exchange are both commercial
e-mail servers, there are a number of freely available open source
mailbox servers, such as the Cyrus IMAP server used for UNT student
e-mail service, EagleMail.
E-mail clients must connect to mailbox servers using a predefined
communication method known as a protocol. For example the GroupWise
client, which is a separate piece of software from the GroupWise server,
and the Outlook client which is used to access an Exchange server both
use the MAPI protocol for transmitting messages from server to client.
MAPI is a proprietary (i.e. closed) protocol developed by Microsoft.
Some commonly used Internet protocols are POP (Post Office Protocol) and
IMAP (Internet Message Access Protocol) which may be familiar if you
have Internet service at home. Both GroupWise and Exchange include or
make available IMAP and POP servers that allow IMAP or POP clients to
access those systems' mailboxes.
As you can see by now, much has been done behind the scenes before
you read a message in your e-mail client. Your e-mail client is most
familiar to you, but what it really is doing is interacting with your
mailbox server to allow you to read, store, reply to, or delete
messages. A flexible mailbox server will allow you to access e-mail via
different protocols. This means you can use different clients if you
need to, such as a web-based client, or a PDA-based one.
It's easy to take Internet e-mail for granted, but it is somewhat
amazing that it works as well as it does (SPAM notwithstanding.) The
reason that Internet e-mail works is because the standards and protocols
for e-mail communication were established long ago, were defined by the
Internet community, and openly documented for all to use. That's
something to think about the next time you click on the "send" button.
Return to top |
