Could you guys with experience of mail servers tell me what you'd see as the best option for an integrated mail/calendaring solution with public folders for a Windows network (even if the server is non Win)?

We've had Exchange (5.5), and Scalix with Outlook. Both worked as far as our users were concerned, but neither was particularly easy for me to manage. One of the reasons we're moving to Exchange is that we now have a few users who want mail through their phones and I've been told this is is easier via Exchange.

I like Thunderbird a lot, but it's sufficiently different from Outlook for me to have to go through it with all my users, likewise (I guess) any other client, although Outlook 2007 is hugely different from 2000 (our current version)...

I haven't looked into this in a while, which means that I'm not totally familiar with current products, but I'm happy to provide some general input.

First off, I would suggest avoiding close integration between the mail and the calendaring. I see this as one of Exchange's larger design flaws, along with its proprietary datastore that makes backup and restore difficult. I'm not saying you can't have email notifications, have the same client for mail and calendaring, or even potentially have the same datastore for both, but I firmly believe that the server side should be kept separate.

From that point, though, the sky's the limit. I would personally install an IMAP server (I have the most experience with Cyrus, but there are many others that are reasonable), an LDAP server for an address book (I feel like there ought to be a read-write LDAP-based addressbook client by now, but I could be wrong), and a separate calendaring server.

The calendaring server is the hard part. There don't seem to be many out there. A quick survey suggests that the Sun calendaring server might be okay, but you should probably do your own research on that.

Thanks Bitt, now this is where I guess proprietary solutions win, 1 widely installed app vs 3 separate ones that I'll never find decent support for (one of the reasons we're moving away from Scalix). I'll look into Cyrus, but the address book and calendar are integral to what we do, and the ability to set permissions on shared calendars and address books relatively straightforwardly is invaluable.

How visible would a setup like this be to the user, i.e. would all the functionality be available within a single client app (and what would that be)?

Well, a big problem with a lack of calendar servers is a lack of calendar clients. As such, most calendar servers will work with Outlook, which you can also set up for IMAP mail.

Call me a snob, but I feel that the reason that there is so much Exchange support available is that it's a piece of junk that needs constant handholding. In my experience, mail servers tend to need very little regular maintenance.

Evolution and Thunderbird/Sunbird both support CalDAV servers (allegedly -- I've never tried it).

Peter

Gotcha, Outlook with an IMAP mail server and a discrete calendar server, public address books?



I agree, BUT on the few occasions I've had need of support on my Scalix server the Scalix forum has been more useful than my support organisation, not exactly ideal, and kind of defeats the object of paying someone for support.

Well, I don't know anyone that's used a similar setup either (although I have looked at it)...

My own view is that mail/calendar servers should effectively be appliances where platform really doesn't come into it, but life's not like that, yet. Is it?

With mail servers, it is like that. It's just that the most visible email server is not. Virtually everything else follows standards, Lotus Notes, I think, being a notable exception.

It certainly isn't with Scalix, haven't sucessfully managed to export to a PST, and forget public folders.

Part of me wants to outsource the whole thing to some offsite service, but that's bound to have it's own issues. One day I'll get a full time IT guy in.

Perhaps you should look at Zimbra, then?

There's also SmarterMail and Open-Xchange. I don't have any experience with any of them, but I do try to keep my "Exchange Replacements" bookmarks folder vaguely up-to-date.

Isn't MS buying Yahoo? Can't see them being too keen on developing an Exchange rival.

Perhaps not, but, seeing as it's open source, anyone could pick up that ball and run with it.

Then again, if I were running Microsoft, I think I'd be looking at ways for Exchange to not suck, and throwing it away and replacing it with something that shares no code sounds like an ideal solution. I seriously doubt that anyone in charge of Microsoft shares my sentiment, though.

I think Google would be happy to help you with that, among others.

I just wish Google sold (or rented) an appliance — anything that could be kept in-house. Many companies don't want their email in someone else's hands.

I understand why Google doesn't want to do that; I'm sure they're not happy with the idea of their proprietary information being placed in someone else's hands, but that's the same reason that people are reticent to put their email in Google's (or anyone else's) hands.

We looked at Google, and MS now do hosted Exchange (forgotten what they call it) but the feeling in house was that with a lot of image files being sent and received a hosted solution would be less responsive than our own server.

Gmail, if you use it over IMAP, would be problematic in this regard. Using it with the web interface, however, you might get real benefits, since forwards and replies that retain the attachments don't actually use any network bandwidth. Everything just stays put on Google's servers. I suppose you could profile your current activity and sort out how often people forward big attachments or otherwise include the same ones in replies.

It's not single big attachments, more multiples of small ones (jpg, pdf), we normally have to view the whole doc to extract the relevant data (e.g. a 3 page purchase order may only have a single data line). I guess the other worry with using a hosted service is what happens when you lose connectivity, hopefully our dual WAN setup should mitigate this but there's always the risk that someone digging up the road cuts through all your telco cables (we've had it happen in the past). I'm giving myself a headache here.

I spoke to Open-Exchange before we went for Scalix, at the time they had a single UK reseller who had a single tech trained on it, but at least they were honest about it. Each of our Scalix resellers (3 of them) has assured us they have numerous trained techs on board, but it's always the same guy that answers all my support calls...

I don't know, Exchange just looks like the lowest risk solution.

That's great as long as you never want to look at them.

(Sorry; I seem to be argumentative today.)

"No one ever got fired for buying IBM."

First off, I think that Exchange is full of fail, from needing a lot of handholding, to costing a huge amount, to failing to implement standards, etc. To me, that's risk. It means an excessive amount of IT expenditure and an assurance that you're always going to be a 100% Windows shop.

Let me give you an example of a non-obvious Exchange-related IT expenditure. One would assume that backing up your users' mail is important. And, one would assume, if that's the case, it's important to be able to restore their mail. Now, you can back up Exchange's mailstore blob and restore it without significant effort. However, if a user realizes that they need an email from a few weeks ago, the only thing you can do is restore the entire mailstore blob. Which means that all of the emails received in the last few weeks are gone. For all your users. Now, you can restore that blob, let the user copy the pertinent email out, and then restore the current blob, but that means probably hours of downtime for your entire mail system. If you want to be able to restore individual messages without downtime, you'll have to invest in a commercial backup solution and their Exchange client. (As far as I know, there is no free/open backup server that can deal with Exchange datastores; I believe that Microsoft licenses that specification to the backup vendors.) But that's the sort of thing you need to consider, and that Microsoft doesn't tell you. In fact, it's entirely possible that you could implement an Exchange solution without realizing this and then be up the creek when your CEO requests that email.

Second, even if you don't buy that argument — that those factors are not risk — you have to ask yourself: is that reduction in risk really worth that increase in cost? Now, you can argue that as long as your users are happy with mediocre (sadly, not a terrible assumption) that Exchange/Outlook is good enough. Personally, I feel it's my responsibility as an IT guy to improve the workflow of my users where possible. (That said, Exchange and Outlook do work together well. They're like two Legos that fit together perfectly, as long as you're willing to overlook that all of the other faces have been chewed up by the dog.)

Third, is it risk to your users, to your company, or to your job? I'm not going to make any arguments about the virtue of those assessments, but make sure you know what risk you're reducing. Don't lie to yourself.