OSEC

Neohapsis is currently accepting applications for employment. For more information, please visit our website www.neohapsis.com or email hr@neohapsis.com
 
From: Kevin J. Menard, Jr. (kmenardWPI.EDU)
Date: Tue Jun 05 2001 - 10:32:43 CDT

  • Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]

    Hey Peter,

    Monday, June 04, 2001, 5:25:47 PM, you wrote:

    PH> this problem occurs:

    PH> after a user is "authenticated", it is possible for him to send mail
    PH> with any account name, so everybody after popping succesfully could send
    PH> mail as e.g. rootmachine.com or bobby could send as bossbigcompany.com
    PH> ... ok, that´s no wonder, because the script, as I understand it, just
    PH> checks for ip-addresses.

    This is not a problem as you put it. It's the way it's supposed to work. It's
    fast and simple. If you can't trust your users to Do the Right Thing, then you
    can always check your logs afterwards and see who had relay access at that time.

    PH> That´s also a bad idea for people living behind
    PH> a proxy... do I really have to enable my whole department to send mail
    PH> via my account just because we´re using the same gateway?

    PH> Wouldn´t it be better to check usernames also?

    That would require a hell of a lot more resources, and inner knowledge of how
    the mailboxes are set up. It's just impractical.

    PH> And how to put this together with postfix to check not only for
    PH> ip-adresss but also that from-address matches the pop-user-id?
    PH> (something with smtpd_sender_restrictions should be the way?)

    PH> I am new to this and I may be wrong!

    PH> I haved hacked one or two little perl scrips, but I do not feel in state
    PH> of solving this myself - in fact I wonder that there is not a readymade
    PH> solution for such everyday-situations like pop-before-smtp - no wonder
    PH> spammers have easy life...

    PH> Please, could anybody give me a hint in how to build a pop-before-smtp
    PH> solution that is as reliable and secure as postfix itself? I mean, why
    PH> do you use a secure mailer if you patch something around it, that makes
    PH> it unsecure?

    I'm no mail guru, but I can't think of a viable way of doing this. SMTP
    provides no means of checking that you are who you say you are. It provides
    relay or no relay. Nor does SMTP really care about domain names. You're asking
    for a massive bridging of data between IMAP/POP and SMTP, and doing user lookups
    on that, if possible, would be agonizingly slow.

    PH> Is there a solution that plays well with postfix, that can handle all
    PH> situations needed with virtual hosting, roaming users?

    Recall that these are your users. you're not just opening up an open relay.
    Part of being a mail administrator is dealing with your users. If they give you
    crap, revoke email privileges if necessary. Also note, this is not just a
    pop-before-smtp issue. You could spoof the From: address on any machine that
    has relay access (try it from your local machine with "mail" if you'd like).

    PH> thank you very much for your attention,
    PH> Peter

    --Kevin

    -
    To unsubscribe, send mail to majordomopostfix.org with content
    (not subject): unsubscribe postfix-users