wasm-demo/demo/ermis-f/imap-protocol/cur/1600095133.23064.mbox:2,S

46 lines
1.9 KiB
Plaintext

MBOX-Line: From imap at tlinx.org Sun Aug 7 18:40:13 2005
To: imap-protocol@u.washington.edu
From: Linda Walsh <imap@tlinx.org>
Date: Fri Jun 8 12:34:36 2018
Subject: [Imap-protocol] Protocol Q: IMAP(s) svr (want tcp)-> IMAP client
Message-ID: <42F6B7FD.2000408@tlinx.org>
Under what circumstances, or why would an IMAP(s) server try to
open a connection back to a client?
I currently have firewall rules setup to allow outgoing connections
from my clients to an external imaps (secure) server. This seems
to work just fine, but I'm wondering if I am creating some inefficiency
in downloading IMAPS email.
I notice (have for a long time) entries in my firewall' "denied"
traffic log that my ISP's imap server is attempting to initiate
an inbound TCP connection from server port 993 to a high numbered
(>50000) client port.
I'm wondering under what circumstances an IMAP server might try
to open a new secure connection back to a client server. I don't
know what server they are using, but I am using fetchmail as a
daemon running in poll mode using idle. Why doesn't send it
back on the already open connection?
I've been hesitant to allow incoming requests do to the increased
possibility of 'spoofing'. I'm not even sure what process on my
end should even be "listening" for such a request. I am unsure about
what credentials it would try to use, though, presumably (?) the
credentials I used to connect to it? But I don't know that 'fetchmail'
is even configurable to listen to 'random' ports coming from semi-random
mail servers (mail1, mail2...mail25, mail26...).
I've never seen the u.washington IMAP server try to connect back
to any of the clients I use with it -- and I would an immediate
popup notification if my imap server tried to open an IMAP connection
to the client (as my client system is setup to notify me of
non-whitelisted traffic).
Thanks for any insight...
Linda