I noticed that there is an email address validation method, that is to telnet to a email MX server and send rcpt command to the server. But for this method, the MX server may abort connection after several tests.
However, on Internet, there are many websites which claim that they can verify large amount of email addresses in short time, and they provide such services with charges.
I'm wondering how do these email verification companies/websites verify large amount of email addresses?
They don't, it takes less work for them to simply send emails out to invalid addresses than it does to weed out the invalid ones.
In my experience, most MTA's won't actually verify E-Mail addresses if you connect to them for the exact reason that it allows people to enumerate valid addresses. Usually they'll either reject the VRFY command and/or accept mail for any user specified in the RCPT line and then silently drop mail for invalid ones.
As to services that say they'll validate addresses, I'd take that claim with a pinch of salt. Whilst it's possible to do things like send mails to every one on the list and count bounces try to use other methods (e.g. web bugs) to verify live addresses, none of these approaches is foolproof.
One other approach they may take is to validate supplied addresses against lists that they've got from elsewhere, but again not a foolproof method.
Most spam campaigns rely on botnets to do the real work, so having individual nodes blacklisted as spammers due to sending lots of emails to non-existent addresses isn't a big deal for them. Often they'll just harvest as many addresses as possible and go nuts, without any real verification.
One of the smarter approaches I've seen is to send "legitimate" newsletters, which aren't likely to get flagged as spam, and include a few links to product pages and an unsubscribe link. Users that click any of the links are immediately flagged as active. This reduces the chance of having the messages blacklisted by spam filters, and makes the spam efforts more efficient.
There are many methods. You can send mail with a review un subscribe link, that will in fact register you as valid and active. You can create a real service that ask for email to users (newsletter, forum inscription) with a confirmation method, then use them for other purposes. My favourite one would be the creation of a website that can send an image for all request you make to him. Send HTML messages that include an image and uses a personalised link. When the mail client download the image for display, register the mail associated to the link. This method can be used with large images as well as 1x1px images for stealth. Some antivirus may detect some of these attacks.
In some cases they simply don't rely on SMTP methods. Instead they place a personalised link in a spam mail, like http://example.com/news5/display.php?M=537686&C=eee9c997b0e958fa373533c5f4. If someone clicks on that link a spammer can conclude that the email was received and therefore the address is working.
If a spammer provides the kind of verification service you described, he might simply have a database of known working addresses and checks the supplied addresses against the database.
Nowadays most mail servers have countermeasures against checking if email addresses exist. So I assume this is not working on a larger scale.
