The Simple Mail Transfer Protocol (SMTP) enables applications to
deliver email messages to one or more recipients. The control provides
an interface for addressing and delivering messages, and extended
features such as user authentication and delivery status notification.
This control is typically used in conjunction with the Mail Message
control to create the messages, and the Domain Name Service control to
determine what servers are responsible for accepting mail for a
When a message is delivered to a user, the application must
determine what mail server is responsible for accepting messages for
that user. This can be accomplished using the Domain Name Services
(DNS) protocol, a protocol that is most commonly used to resolve host
names such as www.microsoft.com into Internet addresses. This is
typically accomplished by sending a request to a nameserver, a computer
system that provides domain name services. In addition to resolving
host names, nameservers can also provide information about those
servers which are responsible for accepting mail for a given domain.
There can be multiple servers which process mail for a domain with each
server assigned a priority as part of their mail exchange (MX) record.
If there is no mail exchange record for a domain, then the domain name
itself is used.
To deliver a message directly to the recipient, you must examine the
recipient address and request the list of mail exchanges for that
user's domain. Using the DNS control, this is done by reading the
MailExchange property array. If the recipient address is
firstname.lastname@example.org, you would want to enumerate the mail exchanges for the
example.com domain. This will give you the name of the servers that
will accept mail for users in that domain. For example, the property
may return the host name mail.example.com as the name of the server
which will accept mail for users in the example.com domain. Note that
it is possible that one or more of the mail exchanges for a domain may
not be in the recipient domain itself. In other words, it is possible
that smtp.othercorp.net could be returned as a mail exchange for
example.com. This is frequently the case when another organization is
forwarding mail for that domain.
Therefore, there are three general steps that you must take when
delivering mail directly to the recipient:
- Parse the address of each recipient in the message. If you are
using the MIME control, the Recipient and Recipients properties can
be helpful in extracting all of the recipient addresses. Everything
after the atsign (@) in the address is the domain portion of that
- Perform an MX record lookup using the DNS control by setting the
HostName property to the domain name and reading the values returned
in the MailExchange property array. This property will return the
name of the servers responsible for accepting mail for that user. If
there are more than one server, they will be returned in order of
their relative priority, with the highest priority server having a
lower index value. This means that you should attempt to connect to
those servers in the order that they are returned by the property,
starting with an index value of zero.
- Attempt to connect to the first server returned by the
MailExchange property array. The connection should be on the default
port, and you should not attempt to use any authentication. If the
server accepts the connection, then use the SendMessage method to
deliver the message. If the connection is rejected or the message is
not accepted, attempt to connect to the next mail exchange server
until all servers have been tried.
- If no mail exchange servers were returned by the MIME control's
MailExchange property, or you could not connect to any of them,
attempt to connect to the domain specified in the address using the
default port. If the connection succeeds, then deliver the message.
If you cannot connect or the message is not accepted, then report to
the user that the message could not be delivered.
One last important consideration is that many Internet Service
Providers now block outbound connections on port 25 to any mail servers
other than their own. If you are unable to establish any connections,
either with the error that the connection was refused or it
consistently times out, contact your ISP to determine if port 25 is
being blocked as an anti-spam measure. If this is the case, it will be
required that you relay all messages through their mail servers.
In some situations it may not be possible to send mail directly to
the server that accepts mail for a given domain. The two most common
situations are corporate networks which have centralized servers that
are responsible for delivering and forwarding messages, or an Internet
Service Provider (ISP) which specifically blocks access to all mail
servers other than their own. This is usually done as either a security
measure or as a means to inhibit users from sending unsolicited
commercial email messages. If the standard SMTP port is being blocked,
then any connection attempts will either fail immediately with an error
that the server is unreachable, or the connections will simply
time-out. In either case, a relay server must be specified in order to
send email messages.
A relay server is a system which will accept messages addressed to
users who may be in a different domain, and will relay those messages
to the appropriate server that does accept mail for the domain. Using a
relay server is generally easier than sending messages directly to the
recipient. In order to send a message through a relay, you need to
perform the following steps:
- Connect to the relay server as you would normally.
- Authenticate the client to the server. This may or may not be
required, depending on how the server is configured. Some servers may
be configured to only require authentication if you are connecting
from an IP address that is not recognized as part of that system's
network, for example, if you are connecting using a different
Internet Service Provider. Others may always require authentication.
Check with the server administrator if necessary to determine if and
when authentication is required.
- Use the SendMessage method to deliver the message to the
recipients through the relay server. If there are multiple
recipients, you can use the MIME control to enumerate the recipient
addresses and then pass them to the SendMessage method.
It is important to note that using a mail server as a relay without
the permission of the organization or individual who owns that server
may violate Acceptable Use Policies and/or Terms of Service agreements
with your service provider. Systems which relay messages from anyone,
regardless of whether the message is coming from a recognized domain,
are called open relays. Because open relays are often used to send
unsolicited email, many administrators block mail that comes from one.
It is recommended that users check with their network administrators or
Internet service providers to determine if access to external mail
servers is restricted and what is the acceptable use policy for
relaying messages through their mail servers.
The following methods are available for use by your application:
Initialize the control and load the Windows Sockets control for the
current process. This method is normally not used if the control is
placed on a form in languages such as Visual Basic. However, if the
control is being created dynamically using a method similar to
CreateObject, then the application must call this method to
initialize the component before setting any properties or calling any
other methods in the control.
Establish a connection to the SMTP server. Once the connection has
been established, the other methods in the control may be used to
deliver messages to the server.
Authenticate yourself to the server using a username and password.
This method should be called immediately after the connection has
been established to the server. This is typically required if you are
attempting to use the mail server as a relay, asking it to forward
the message on to the server that actually accepts email for the
recipient. Many Internet Service Providers (ISPs) require that users
authenticate prior to sending mail through their servers. You may
need to contact the server administrator to determine if
authentication is required.
Disconnect from the server and release any resources that have been
allocated for the client session. After this method is called, the
client session is no longer valid.
Unload the Windows Sockets library and release any resources that
have been allocated for the current process. This is the last method
call that the application should make prior to terminating. This is
only necessary if the application has previously called the
There are two general methods that can be used to deliver messages
through the mail server. In most cases, it can be done with a single
method call. However, there are some circumstances where it would be
more appropriate to perform the transaction in stages. The SMTP control
supports both methods.
This is the simplest method for sending an email message through the
server. You provide the sender and recipient addresses, along with
the message contents and the method will submit the message to the
server for delivery.
This method begins a transaction in which a message is dynamically
composed, addressed and delivered in stages. You provide the sender
address and message size to this method, and after it returns you
begin the next stage, which is addressing the message.
This method adds a recipient address to the recipient list for the
message. This should be called once for each recipient, as well as
for any recipients who are to receive "blind copies" of the
message. A blind copy is when the message is sent to a recipient, but
that recipient's address is not listed in any of the headers of the
message; the other recipients will be unaware that the message was
delivered to him. Most servers have a limit of approximately 100
recipients per message. It is possible that this method will return
an error for a specific recipient address; the address may be
malformed or it may not be acceptable for some other reason. This
does not mean that the message will be rejected in its entirety, only
that the specified recipient is not acceptable.
This method should be called after all of the recipients have been
added. It is used to send the contents of the message to the server.
It is also possible to use the lower level Write method to send data
directly to the server, however AppendMessage is generally easier to
use and can write data from memory, the system clipboard or from a
file on disk.
This method is called after the entire message has been sent to the
server. This terminates the transaction and the message is submitted
for delivery. Note that it is possible for the server to accept the
message up to this point and then reject it at this final step due to
some restriction, such as the message being too large.