1: I meant like this:
Without whitelisting, you can't login to SMTP or IMAP, password isn't valid at all.
To enable SMTP and IMAP, you then either surf ro webmail, or the 2FA gateway, and login with: Username + password + 2FA code + captcha.
When all is valid, then your IP is whitelisted for SMTP and IMAP access. This still means you have to use usename/password for SMTP/IMAP.
So how would this be a security hole? Instead of using only username+password for SMTP/IMAP? The whitelisting procedure ADDS to the security. The baseline security with username+password is already there, but now you ALSO need a whitelisted IP to even get a chance to authenticate.
Kind of stupid that there doesn't exist some common standard for 2FA that works in email clients. Some clients do support TLS client certificates, and some clients do support certain "extensions" for 2FA auth. But only common supported in all clients is password auth without 2FA, which is pretty insecure.
Outlook have solved 2FA auth with a webview that uses OAUTH to create a authentication token, for use with SMTP/IMAP using some proprietary extension with gmail and hotmail. But that webview is not something you can trigger from a third party service.
Captcha is there to prevent bruteforcing. If a valid captcha is submitted along with a 2FA code, you could lock out the account for 1 minute for each invalid attempt. If a invalid captcha is submitted, you ignore the request completely. This then prevents a attacker from flooding the server with invalid auth requests for the sole purpose of keeping a user locked out. (Account Lockout DDoS attack)
I had problems with my mail password getting hacked all the time. The instant I added IP whitelist to my system and blocked all non-approved IPs from authenticating at all (so you must have username + password + correct IP to gain access) - then all hacking of my passwords have stopped. IP lockout was the solution to my problems.
2: The idea with the reverse-proxy gateway, is only to prevent auth-bypass or non-authenticated security holes. If you have a web service that has a suspected vulnerability that could be used without authenticating, or could be used to bypass authentication, then you put a reverse proxy in front. The reverse proxy does the authentication, and only forwards requests belongning to authenticated users. Even if the webservice behind, has a auth-bypass hole, it cannot be exploited, as the reverse proxy is behind the service, and non-authenticated users cannot even touch the webservice at all.
-----Ursprungligt meddelande----- Från: dovecot-bounces@dovecot.org dovecot-bounces@dovecot.org För @lbutlr Skickat: den 27 oktober 2020 15:57 Till: dovecot mailing list dovecot@dovecot.org Ämne: Re: Looking for a guide to collect all e-mail from the ISP mail server
The second way, is to not have webmail at all, but instead have a authentication gateway in browser, where you must auth with 2FA and captcha. The only purpose of this gateway, is to authenticate users with 2FA before
On 25 Oct 2020, at 22:47, Sebastian Nielsen sebastian@sebbe.eu wrote: their IP is whitelisted.
I mostly agree with the sentiments in your email, but whitelsiting IP addresses is a HORRIBLE idea and a massive gaping security hole and using a captcha is only slightly less horrible and user-hostile. If you are using 2FA there is absolutely no reason to use a captcha.
A 2FA gateway that reverse proxies the webmail is quite good, but enforcing good passwords and using TLS is good enough for nearly all use cases.
(I recently upped the minimum password length from 12 characters)
-- Ah we're lonely, we're romantic / and the cider's laced with acid / and the Holy Spirit's crying, Where's the beef? / And the moon is swimming naked / and the summer night is fragrant / with a mighty expectation of relief