Say if I sent an email to user@theirdomain.com, the exchange server and the user will receive it, but if I sent an email to fakeuser@theirdomain.com using SMTP client, there will be no fakeuser mailbox, and the local exchange server may bounce it back to the sender, but will the destination server receive the mail at all?
I was trying to send testing emails, but don't want to use the real server name, if it will cause trouble on the server side.
There certainly is a difference in how mail servers behave in that case. Ideally the receiving mail server will reject the RCPT command, in which case the mail content will never be delivered to the target server in the first place.
In a much less ideal scenario (which is unfortunately often seen in practice anyway), the mail server will accept the RCPT command even though the target address does not exist. If the server proceed to accept the incoming mail, it now becomes the responsibility of the receiving server to generate an error message.
This is a problem, because some of the mails will be spam mails with a forged sender address. By accepting and bouncing, the mail server in question is causing spam mails to end up in other people's mail box.
It is easy to test a specific address yourself. You just need to use the telnet command and issue the following SMTP commands HELO (or EHLO), MAIL, RCPT, and QUIT with the proper arguments. By following the RCPT command by QUIT rather than DATA, you will not actually send a mail to the server.
Here is what it could look like when testing the address invalid@example.com
telnet mx.example.com 25
Trying 2001:db8::1...
Connected to mx.example.com.
Escape character is '^]'.
220 mx.example.com
HELO mx.example.net
250 mx.example.com
MAIL From:<>
250 Bounce OK - Please include Message-ID from original message
RCPT To:<invalid@example.com>
550 No such user
QUIT
221 Closing connection
Connection closed by foreign host.
