OASIS Mailing List ArchivesView the OASIS mailing list archive below
or browse/search using MarkMail.


Help: OASIS Mailing Lists Help | MarkMail Help

pki-tc message

[Date Prev] | [Thread Prev] | [Thread Next] | [Date Next] -- [Date Index] | [Thread Index] | [List Home]

Subject: RE: S/MIME vs DKIM

Hi Anders,

I do not think I said gateway encryption should not be used. It
certainly depends on the scenario how it is used. If there is a
mechanism deployed to secure the data being delivered to the gateway
machine and the gateway machine is properly secured, that is fine.

Maybe I missed your point now. 
At first it seems to me that you cautioned the use of PKI, but now you
seem to encourage the use of SSL for solving/answering the question I
raised by pointing out "web mail uses SSL". But SSL is still PKI based
technology. If that is the case, SMTP with TLS (RFC2487) can be used too
along with client authentication in the SSL/TLS that will solve spamming
problem to certain extent. But again cost vs security, esse of use and
other factors etc.  

My main point here is that we should not discourage the use of PKI
because some new technology like DKIM is coming along. It is certinly a
nice addition, because no technology can solve all levels of needs of
business and security as we all agreed.

Thanks for your comment that inspired me for the discussion. It is a
good one. 

-----Original Message-----
From: Anders Rundgren [mailto:anders.rundgren@telia.com] 
Sent: Thursday, April 13, 2006 1:17 PM
To: Yu, Jiafu; PKI TC
Subject: S/MIME vs DKIM

Isn't actually this product of yours actually a gateway server?

The document does not talk about end-2-end security for the most
obvious reason of all:  It was never a particularly good idea in an
organizational environment.  Until the PKI community get to grips
with simple fact, progress will be limited.

Yu, Jiafu wrote:

>Since DKIM relies on dns, the vulnerabilities in dns will cause
>to the keys there too, which in turn will cause legal challenges in
>court if any.

What the vulnerables may lead to is spoofed mail.  Such messages
cannot be traced back to the impersonated domain's key which IMHO
means that such messages can be repudiated and rejected without going
to court.

That there are vulnerabilities is of course a problem as the point is
really that you should trust without doubts.  I believe that
Phillip Hallam-Baker may change DKIM to include support for
Gateway PKI based on the existing SSL PKIs.  The reason for not
using PKI was actually to cater for hobbyists who run their mail-
servers who cannot afford a $200/y gateway certificate.   I believe
that we could live with two clasess of DKIM mail, "first class" and
"tourist class".  Probably 99% of all users mail trough IPSs or
companies and should therefore be able to use first class mail.

>I remember DKIM mentions long list of the issues or
>conditions still waiting to be resolved in order to make DKIM more
>secure than it is. With those potential holes, I would not think it has
>a strong legal binding than traditional PKI.

Legally binding signatures "on masse" will *never* be carried out
through non-interactive and static e-mail.  For that people will
use web signing.   This is already a fact in the EU.  The EU is
BTW years ahead of the US when it comes to mass market PKI.
For low-volume, community based person-to-person messaging
S/MIME is probably sufficient as is.

>As for gateway level encryption in DKIM, it would be a nice add-on in
>the future, but it does not solve the content security issue between
>email client to its local gateway unless email client can talk securely
>with its local gateway. Then the question would be: how complicated it
>would go? Will it face similar problem as PKI faces now. 

SSL does this job with ease.  Web mail uses this as standard.


-----Original Message-----
From: Anders Rundgren [mailto:anders.rundgren@telia.com] 
Sent: Thursday, April 13, 2006 10:52 AM
To: Yu, Jiafu; PKI TC
Subject: Re: [pki-tc] Candidates for OASIS PKI TC Chair

Hi Jiafy,

You are right, the goals with DKIM are not the same as with S/MIME.

That DKIM does not have legal binding is though something that can
be elaborated a bit.  Legally binding signatures have two distinct

1. As a direct replacement of wet signatures (=automation).  This is
   where DKIM and S/MIME currently differs since DKIM has no
   support in signature directives.

2. As an evidence in court.  The legal systems have shown to be very
   pragmatic, putting folks in prison based on things like unsigned mail
   and IP addresses.  DKIM, particularly in an organization context
   is likely to be as hard to repudiate as an S/MIME signature.

S/MIME encryption is probably the least working PKI application
there is, since it is in conflict with an organization's need for virus
checking on in-bound messages as well as content monitoring
on out-bound dittos.  Not to mention how hard it is to use.
But there is more: That millions of US public sector employees
have certificates does not help much as there is no easy way you
can get hold of their public keys due to privacy concerns.  I always
thought that the primary reason for having a public sector was for
serving the society at large!  Some recent, associated papers:


My hope is that DKIM will get an update so that even encryption
is handled at gateway level.  This should be a no-brainer AFAIK.

Anders Rundgren

----- Original Message -----
From: "Yu, Jiafu" <Jiafu_Yu@stercomm.com>
To: "Anders Rundgren" <anders.rundgren@telia.com>; "PKI TC"
<pki-tc@lists.oasis-open.org>; "Arshad Noor"
Sent: Thursday, April 13, 2006 16:02
Subject: RE: [pki-tc] Candidates for OASIS PKI TC Chair

I would question statement like "That secure e-mail currently is being
redesigned from the ground and up (DKIM)". The goal of DKIM is to
identify who sends email mainly from smtp gateway point of view (for
supressing spamming). It does have the advantage of ease of key
distribution using dns for key lookup and retrieval, but it is only for
message origin verification, not for message content security (encrypted
like SMIME). It does not have a legal binding at this point. Secondly,
depending on how we use PKI, if we mainly use PKI without legal binding
as DKIM, PKI will be more popular than it is now. I still think PKI at
this point is still the most effective solution in certain industry or
region for its original goal(authentication, integrity, non-repudiation

-----Original Message-----
From: Anders Rundgren [mailto:anders.rundgren@telia.com]
Sent: Thursday, April 13, 2006 8:37 AM
To: PKI TC; Arshad Noor
Subject: Re: [pki-tc] Candidates for OASIS PKI TC Chair

>Businesses continue to search for the elusive silver bullet to
>solve their security problems - but it is my belief that until
>they start using PKI in many different aspects of their IT infra-
>structure (along with appropriate changes to applications,
>business processes and employee training), that silver bullet
>will continue to elude their grasp.

I think businesses should be cautious embracing a technology
that not even the people who claim to know PKI, know how to
apply to everyday business processes such a e-purchasing.
Until such knowledge becomes common, agreed upon, and
published[*], businesses betting on PKI are at risk being stuck
in pretty "consultant-intensive" activities.

That "secure e-mail" currently is being redesigned from the ground
and up (DKIM), is another indication that the previous generation
of PKI "theologists" did not actually foresee the Internet revolution.
The problem is that S/MIME effectively delegates security policy
enforcement down to the [nowadays often rather novice] users.
The following is how secure e-mail should have been:

    "If I send a mail via my company, it is my company that secures it"

If any of the TC chair candidates have the guts to address any of
these issues, he or she has my full support.

Anders Rundgren

Go to NIST's PKI pages.  Nothing
Go to PKI-TC AGSC pages.  Nothing
Go to PKI-*.  Nothing

To unsubscribe from this mail list, you must leave the OASIS TC that
generates this mail.  You may a link to this group and all your TCs in

To unsubscribe from this mail list, you must leave the OASIS TC that
generates this mail.  You may a link to this group and all your TCs in

[Date Prev] | [Thread Prev] | [Thread Next] | [Date Next] -- [Date Index] | [Thread Index] | [List Home]