Re: [Discuss] Delivering mail to folders

[Edward Ned Harvey (blu)]

> From: Discuss [mailto:discuss-bounces+blu=nedharvey@blu.org] On
> Behalf Of Tom Metro
> 
> > Ever-so-slightly better than no encryption.
> 
> Huh? We're talking about using a self-signed cert for IMAP access, right?
> 
> Self-signed certs have all the same cryptographic benefits as a CA
> signed cert, including having your client validate the cert, if you
> install your own root cert on your clients.
> 
> The only down-side to self-signed certs is the inconvenience of having
> to install the root certs on your clients. This is why they aren't used
> for public web sites.

Creating a self-signed cert isn't the same thing as creating your own CA and 
installing the CA root as a trusted root on your clients. If you create your 
own CA and distribute your own CA root to all your clients - as you said - 
you'll get pretty good security (unless you screw something up). A self-signed 
cert is one which certifies itself. The client cannot follow any chain to a 
trusted root, so the client needs to either reject the cert, or prompt for user 
interaction (in which case, users almost invariably click "accept," and thus 
are easy to attack via MITM). If the user accepts the cert, some clients (such 
as firefox) have the option to do certificate pinning, so it won't prompt again 
when it sees the same self-signed cert, similar to the way ssh behaves when 
connecting to a new unrecognized server.

But if you have a client that prompts you to accept a self-signed cert, and you 
accept it, and the client pins it, and at a later time the cert changes (MITM 
attack)... Does the client prompt you again? Openssh refuses to talk to a 
server with a pubkey different from the pinned key, as it should. But every SSL 
client I've ever seen (firefox, chrome, ie, etc) will prompt you again to 
accept the unrecognized cert, so even highly technical and reasonably alert 
people are still vulnerable to the MITM attack on a self-signed cert. ... As 
David in particularly would be, because he mentioned a checkbox for "ssl accept 
any certificate," and asked "is that a good option?"
___
Discuss mailing list
Discuss@blu.org
http://lists.blu.org/mailman/listinfo/discuss

Re: [Discuss] Delivering mail to folders

[John Abreau]

Apparently I've been doing it "wrong" all these years. I've always created
my own CA and signed my certificates with it, and I thought that's what the
term "self-signed" meant.

On Mon, Feb 1, 2016 at 5:50 PM, Edward Ned Harvey (blu) 
wrote:

> > From: Discuss [mailto:discuss-bounces+blu=nedharvey@blu.org] On
> > Behalf Of Tom Metro
> >
> > > Ever-so-slightly better than no encryption.
> >
> > Huh? We're talking about using a self-signed cert for IMAP access, right?
> >
> > Self-signed certs have all the same cryptographic benefits as a CA
> > signed cert, including having your client validate the cert, if you
> > install your own root cert on your clients.
> >
> > The only down-side to self-signed certs is the inconvenience of having
> > to install the root certs on your clients. This is why they aren't used
> > for public web sites.
>
> Creating a self-signed cert isn't the same thing as creating your own CA
> and installing the CA root as a trusted root on your clients. If you create
> your own CA and distribute your own CA root to all your clients - as you
> said - you'll get pretty good security (unless you screw something up). A
> self-signed cert is one which certifies itself. The client cannot follow
> any chain to a trusted root, so the client needs to either reject the cert,
> or prompt for user interaction (in which case, users almost invariably
> click "accept," and thus are easy to attack via MITM). If the user accepts
> the cert, some clients (such as firefox) have the option to do certificate
> pinning, so it won't prompt again when it sees the same self-signed cert,
> similar to the way ssh behaves when connecting to a new unrecognized server.
>
> But if you have a client that prompts you to accept a self-signed cert,
> and you accept it, and the client pins it, and at a later time the cert
> changes (MITM attack)... Does the client prompt you again? Openssh refuses
> to talk to a server with a pubkey different from the pinned key, as it
> should. But every SSL client I've ever seen (firefox, chrome, ie, etc) will
> prompt you again to accept the unrecognized cert, so even highly technical
> and reasonably alert people are still vulnerable to the MITM attack on a
> self-signed cert. ... As David in particularly would be, because he
> mentioned a checkbox for "ssl accept any certificate," and asked "is that a
> good option?"
> ___
> Discuss mailing list
> Discuss@blu.org
> http://lists.blu.org/mailman/listinfo/discuss
>



-- 
John Abreau / Executive Director, Boston Linux & Unix
Email j...@blu.org / WWW http://www.abreau.net / PGP-Key-ID 0x920063C6
PGP-Key-Fingerprint A5AD 6BE1 FEFE 8E4F 5C23  C2D0 E885 E17C 9200 63C6
___
Discuss mailing list
Discuss@blu.org
http://lists.blu.org/mailman/listinfo/discuss

Re: [Discuss] Delivering mail to folders

[Rich Pieri]

Dovecot can be a little quirky about the location of INBOX. You can
specify it with the mail_location directive in Dovecot's local.conf file.

-- 
Rich P.
___
Discuss mailing list
Discuss@blu.org
http://lists.blu.org/mailman/listinfo/discuss

Re: [Discuss] Delivering mail to folders

[Edward Ned Harvey (blu)]

> From: Discuss [mailto:discuss-bounces+blu=nedharvey@blu.org] On
> Behalf Of David Kramer
> 
> would
> it be reasonable and possible to use a self-signed cert for starters

Ever-so-slightly better than no encryption. The only difference is whether or 
not it's possible for someone to accidentally see your traffic, or if they have 
to make a point of intentionally looking into it.
___
Discuss mailing list
Discuss@blu.org
http://lists.blu.org/mailman/listinfo/discuss

Re: [Discuss] Delivering mail to folders

[Tom Metro]

Edward Ned Harvey (blu) wrote:
>David Kramer wrote:
>> ...would it be reasonable and possible to use a self-signed cert for 
>> starters...
> 
> Ever-so-slightly better than no encryption.

Huh? We're talking about using a self-signed cert for IMAP access, right?

Self-signed certs have all the same cryptographic benefits as a CA
signed cert, including having your client validate the cert, if you
install your own root cert on your clients.

The only down-side to self-signed certs is the inconvenience of having
to install the root certs on your clients. This is why they aren't used
for public web sites.

Even without installing a root cert, many clients will warn you about
the invalid cert, and if you agree to connect anyway, they give an
option to let you store the exception. If implemented correctly, the
client will warn again if the cert fingerprint changes, raising the bar
(but not preventing) a MITM attack.

 -Tom

-- 
Tom Metro
The Perl Shop, Newton, MA, USA
"Predictable On-demand Perl Consulting."
http://www.theperlshop.com/
___
Discuss mailing list
Discuss@blu.org
http://lists.blu.org/mailman/listinfo/discuss