Συντάκτης: Bruce Richardson Ημερομηνία: Προς: exim-users Αντικείμενο: Re: [Exim] OT: Need some feed back on Exchange as an SMTP server
On Fri, Apr 02, 2004 at 10:08:14AM -0600, Edgar Lovecraft wrote: > On the flip side of things, I am not a MySql admin, so I do not suggest to
> others how they should implement, correct problems, or use MySql with Exim,
> you will not see me post 'answers' to questions about such things as I do
> not know, so please only give arguments of why an Exchange server is bad or
> faulted when you can give real/verifiable information on those faults, as I
> will ask you to back up those claims. As for any claim that I make about
> Exchange ask me to back it up.
When we decided to ditch our old mail system we did evaluate Exchange.
Despite being largely a Microsoft shop (with Linux playing some key
network and mail roles), we rejected it pretty quickly. The killer
argument was that all the features that might make Exchange attractive
(the groupware features) are proprietary. Once reliant on those
features, we would find it very difficult to move away, while our
options for mail clients and additional mail components would be
severely limited. Having had bad experiences with proprietary mail
systems, we were determined to build the new mail system based on open
standards, so that we would not be restricted in our choice of
components and could replace or add mail components with relative ease.
In that context, Exchange offered us little but the danger of becoming
enforced MS addicts. We have the skills to implement our own calendar
and other groupware components and that's what we have done. Since we
were not going to use Exchange's groupware features, it had nothing
special to offer us. Its implementation of Internet protocols is
inferior to those of several of the other systems we looked at (and
quite deliberately omits some features where MS would prefer you to use
its proprietary protocols).
Our current system is a mixture of open source, commercial and in-house
systems, united by our databases (running on MS SQL Server, so we can't
be accused of being simply anti-MS). It works the way we want to,
precisely how we want it to. We could replace any of the individual
components (mailstore, client, datastore) with a minimum of disruption.
I recognise that many organisations don't have the resources to do this
themselves but on the other hand we are not a large IT team (7 in
total), so you don't have to be Google to do this kind of thing.
Branching off on even more of a tangent, I'd like to take this
opportunity to rant about the paltry number of mail clients, commercial
or open source, that can use a database to store preferences and address
books. Any organisation making any serious use of IT has its own
datastore and it's a natural place to store staff and contact
information (much more natural than an Exchange or GroupWise
pseudo-database). The makers of Windows clients have even less excuse
than the others: it would be trivial to add ODBC capabilities. While MS
obviously have no interest in doing this (why let people build their own
systems when you want them to buy Exchange?), I'm puzzled at the
attitude of the rest.
