What's the best / most popular open-source IMAP server these days?

On Thu, 28 Aug 2008, Joel Reicher posted:
>> Put another way, the market doesn't want an IMAP server. It wants a
>> turnkey complete mail server solution (of which IMAP service is a
>> small piece).
> I would have thought security and spam filtering concerns make
> (boundary) mail exchange a distinctly separate problem to that of
> mailstore access.

Technically, yes. However, as delightful as architecting mail servers may
be to people like us, it isn't what customers want.

The tendency is to make things be Someone Else's Problem with one-stop
shopping. It's Wal-Mart vs. the little stores.

When you have a mail server based upon multiple products then getting
those products to interoperate with each other becomes your problem. Put
another way, when (not if) something goes wrong, more often than not the
multiple vendors will start finger-pointing.

Although there is certainly still a market for add-ons to an existing mail
server (the boundary exchange problem being an excellent example of a
market with a HUGE amount of activity), I think that we're seeing a market
heading for single products that address all the phases of a mail server.

Mind you: I don't think that this is a necessarily good thing. It is damn
difficult to build a product that does many things well, as opposed to
building a product that is excellent in one piece of the task. I also
believe that it is better to think in terms of layered products rather
than monoliths.

-- Mark --

http://panda.com/mrc
Democracy is two wolves and a sheep deciding what to eat for lunch.
Liberty is a well-armed sheep contesting the vote.

Mark Crispin writes:

> When you have a mail server based upon multiple products then getting
> those products to interoperate with each other becomes your problem. Put
> another way, when (not if) something goes wrong, more often than not the
> multiple vendors will start finger-pointing.

Ummm, yeah.

> Although there is certainly still a market for add-ons to an existing mail
> server (the boundary exchange problem being an excellent example of a
> market with a HUGE amount of activity), I think that we're seeing a market
> heading for single products that address all the phases of a mail server.

Hmmm, yeah. I reached the same conclusion a little while ago.

# revision 1.1
# date: 1999/12/29 01:49:13; author: mrsam; state: Exp;
# Initial CVS checkin.

> Mind you: I don't think that this is a necessarily good thing. It is damn
> difficult to build a product that does many things well, as opposed to
> building a product that is excellent in one piece of the task. I also
> believe that it is better to think in terms of layered products rather
> than monoliths.

Well, unless someone has objections, I'll just continue to try to make
Courier do everything that it does, SMTP, IMAP, POP3, Webmail (with basic
webmail-based calendaring), email-to-fax, and mailing list management, well.
I'm pretty sure that I'm at least halfway there.

But, yes, I saw which way the wind was blowing many, many years ago.


-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.9 (GNU/Linux)

iEYEABECAAYFAki3KdAACgkQx9p3GYHlUOJ4cgCeMDz53wdQz+ AeJZceI6LuhZfA
RngAn1S1fhEEhXOJ5OpuSphjERVgtZTZ
=k2gx
-----END PGP SIGNATURE-----

Mark Crispin <mrc@panda.com> writes:

> On Thu, 28 Aug 2008, Joel Reicher posted:
> >> Put another way, the market doesn't want an IMAP server. It wants a
> >> turnkey complete mail server solution (of which IMAP service is a
> >> small piece).
> > I would have thought security and spam filtering concerns make
> > (boundary) mail exchange a distinctly separate problem to that of
> > mailstore access.
>
> Technically, yes. However, as delightful as architecting mail servers
> may be to people like us, it isn't what customers want.

That's the inference which surprises me. Customers want a secure and
reliable email service, and I just didn't think that could be provided
with a monolithic architecture. Either I'm wrong about the
vulnerabilities of a monolithic architecture, or customers have some
very low standards for security and reliability.

> When you have a mail server based upon multiple products then getting
> those products to interoperate with each other becomes your problem.
> Put another way, when (not if) something goes wrong, more often than
> not the multiple vendors will start finger-pointing.

I think what you're saying is that the burden of providing security
and reliability is shifted from the sysadmin to the software vendor if
a turnkey solution is adopted, but I didn't think that could provide
the same kind of security and reliability that a site run by
sufficiently skilled sysadmins could provide. Have customers
(i.e. email users) not been exposed to enough well-run email services
to appreciate how good these things can be, or are turnkey solutions
really good enough?

Cheers,

- Joel

On Fri, 29 Aug 2008, Joel Reicher posted:
> That's the inference which surprises me. Customers want a secure and
> reliable email service, and I just didn't think that could be provided
> with a monolithic architecture. Either I'm wrong about the
> vulnerabilities of a monolithic architecture, or customers have some
> very low standards for security and reliability.

It is the latter (very low standards for security and reliability)
combined with a strong desire to make the whole thing Someone Else's
Problem and (in the short term) save money.

The monolith vendors and outsource providers say, "give us all your data,
and let us do all your IT, and you don't need to have an IT department".
The IT department managers understand it somewhat differently: "farm all
the work out to the outsource providers, and we don't need to have IT
engineers and sysadmins."

> I think what you're saying is that the burden of providing security
> and reliability is shifted from the sysadmin to the software vendor if
> a turnkey solution is adopted, but I didn't think that could provide
> the same kind of security and reliability that a site run by
> sufficiently skilled sysadmins could provide.

The flaw in your reasoning is the assumption that such considerations are
at all important to the people who make the decisions.

> Have customers
> (i.e. email users) not been exposed to enough well-run email services
> to appreciate how good these things can be, or are turnkey solutions
> really good enough?

This assumes that email users make the decisions. To a decision maker
whose sole contact with email is via Blackberry, all the world is
Exchange. And given gmail and other cloud based email, why spend the
money on a mail server for underlings who don't rate a Blackberry?

The way to counter this is to have a turnkey solution that does Exchange
better (and cheaper!) than Exchange, does cloud better (and cheaper!) than
the cloud, and is secure and reliable. You're not going to get this
today, but lots of people are racing feverishly to create it.

-- Mark --

http://panda.com/mrc
Democracy is two wolves and a sheep deciding what to eat for lunch.
Liberty is a well-armed sheep contesting the vote.

On Aug 29, 10:04 am, Mark Crispin <m...@panda.com> wrote:

> The way to counter this is to have a turnkey solution that does Exchange
> better (and cheaper!) than Exchange, does cloud better (and cheaper!) than
> the cloud, and is secure and reliable. You're not going to get this
> today, but lots of people are racing feverishly to create it.

I wonder it that will happen before cheap fusion power ...

?