Re: Microsoft takes over British Telecom

Paul Smith <[email protected]>
Newsgroups gmane.ietf.asrg
Message-ID <[email protected]>
On 31/10/2011 18:28, Douglas Otis wrote:
> Kerberos was suggested as a way to avoid much of the overhead related 
> with processing certificates at each exchange.   It also provides a 
> way to offer layered protections, such as selectively enabling 
> firewall subsequent to completion of a Kerberos exchange.  Kerberos 
> exchanges would be at much much lower rates than that demanded by SMTP. 
I'm struggling with the Kerberos idea

If you want to send a message to me, that means you need to authenticate 
with 'my' kerberos server? What authentication details do you use? Do I 
have to contact you to give you authentication details to my server? 
Surely that can't be the case, but how else would the authentication 
work?    (or would the sender authenticate *with itself* via the 
receiver? I can see that working in theory, but would be complicated and 
I can't see how it would work with a standard kerberos implementation)

Given that most of our customers are business with < 25 users, many of 
whom have their MX records pointing to their own MTA, how would this 
work? Most want to use their own MTA because their ISPs are useless, and 
switching to their own MTA gives them enhanced reliability and control.

I can't see them rushing to sign up to a third party kerberos server 
(probably with extra fees and unknown reliability). Does this mean we'll 
be needing to create our own kerberos server to run alongside our mail 
server software? Most of our customers use Windows desktop OSes (eg 
Windows XP, Windows 7 etc), which, AFAIAA don't give you kerberos services.

Using kerberos only makes sense if everyone who has a legitimate mail 
server has kerberos, which obviously isn't the case.

> Kerberos was suggested as a way to avoid much of the overhead related 
> with processing certificates at each exchange.
Does this mean that STARTTLS is undesirable as well?

Is the calculation of a signature that problematic? It's relatively CPU 
intensive, but low bandwidth. I'd have thought most mail systems would 
be I/O bound rather than CPU bound (unless they're doing 
antispam/antivirus, in which case a calculating/checking a signature is 
a relatively miniscule extra load). DKIM already generates/checks 
signatures, and with much more data & complexity.
lmpx.com only provides a reader for public news (NNTP) servers. It is not affiliated with the servers or forums shown here and is not responsible for the content of articles, which is written by their respective authors.