We did this last year. We were on Exchange 5.5 with Outlook. Just got tired of monthly reboots, zombie Exchange processes, viruses, and those damn Outlook profiles.
We put together a system we called Steamroller. Courier-IMAP, OpenLDAP, qmail-LDAP, Sympa, Squirrelmail, and SunOne Calendar Server. Works like a champ. Adios Exchange. Adios Outlook. Adios crappy problems with no obvious cause and no obvious answer.
Courier-IMAP = maildir IMAP OpenLDAP = store user passwords and info qmail-LDAP = grok LDAP for info, virus scanning, SMTP Sympa = distribution lists, err mailing lists Squirrelmail = webmail from any computer SunOne Calendar = uh, calendar
We didn't try to clone Outlook or try to claim that our system was Outlook/Exchange. We distilled what was essential in Exchange and added it to our system and added some refinements.
The big selling point for us was that an email system that goes down is like having no email system at all. We were able to trade some, uh features of Exchange/Outlook in return for rock solid stability.
We were also able to whip up some nice Perl cgi scripts to do administration of the system and a plugin for Squirrelmail that allows the users to have out of office functionality.
We have a HOWTO document if you are interested. You may want to check out QVCS as well
http://www.dulug.duke.edu/~icon/qvcs-guide/
There was a handful of users who resisted the change, however, we worked hard to sell the idea to IT staff, managment and the users. A big plus was having the new system running in parallel for several weeks for users to try it out before we migrated. The migration itself was smooth, a few calls from people who did not attend training, a few 'how do i...' questions and that was it.
Is very handy for our users to go to conferences, training, etc and get access to their same mailbox or calendar as if they were sitting at their desk.
End result was that it removed a PITA administration area and greatly reduced the overhead on our help desk. Nothing like setting up 25 or 30 Outlook profiles on a single computer at a fire station.
The first document is what inspired us over a year ago to begin this project. The second project is very similar to what we ended up with. We will be producing a HOWTO next month on how we did this.
Replacing Exchange is not that difficult if you understand how email works and how Exchange is cobbled together. We chose to separate the Exchange functions and put them in a web browser driven context.
The big task you will have is to fight the user conception, built through marketing and fud, that somehow Exchange/Outlook is synonymous with email in the same way that some people see AOL as synonymous with the Internet.
You will have to fight like hell for an Exchange replacement. A replacement has to be feature-rich, a replacement better be secure, and most importantly a replacement needs to be more reliable than Exchange.
If you can do your homework on these issues you should be able to get PHB and upper management to buy in.
The magical thing that Microsoft, and to a lesser extent Lotus and Novell, managed to do is transform the function of email into the monstrousity that Exchange/Outlook is and convince people to lay down gobs of cash money for something which fundamentally is no different than any other email system - its job is to deliver email.
We got tired of Microsoft sticking it to us for licensing. We got tired of virus after virus. We got tired of Exchange problems with no apparent reason and (worse) no apparent cure. We got tired of having our data held hostage by Exchange.
The big question, for you bofhs out there, is whether you can/will do something about Exchange. You can sit idly by while Exchange craps on you again and again or you can do something about it.
I'm the BOFH at a municipal government and we have been working on this for some time.
I would recommend starting with the backend server and network systems. Swapping these systems (DNS, DHCP, email, backups, file, print, etc) out for Linux/BSD systems can be done seamlessly so your users never know (other than everything runs better and faster) and is a smaller task in scope than everything to Linux.
If you can convert your backend to Linux/BSD (Solaris even) you can break the MS stranglehold on your wallets - break down the servers and you can break down the server licensing (sql = nt license, nt cal, sql license, and sql cals) and the CALs which should be a major cash savings.
You should also look at the cost savings, security benefits, and flexibility gained from staying away from proprietary software (PHB love that stuff).
The next step would be to move as many applications to an open source web based model (if you have developers around - may take some work looking at various projects) as you can.
The hard part is going to be the users. Most of them think os = windows. If you can get them used to the idea of computer diversity (Macs may be an avenue to introduce this idea) and get them some exposure to StarOffice/OpenOffice, Mozilla etc you can get them used to the idea of running non-MS applications and non-MS OS.
In our experience once people actually use StarOffice or Mozilla and they see that they are OK and they tend to be pretty open to running OS X, Linux and so forth.
Your ace in the hole may be to find businesses, gubmints, etc in your area that are working on the same thing. It is much easier to get acceptance of using open source when business X or Foobar U is doing the same thing.
We are working with county govt, school districts, universities, etc in the area - they are tired of paying out of their ass for MS and they want something else. Many of their organizations are much larger than we are but we have many of the same problems. The benefit is a partnership where everyone works together and collectively things move much more quickly that in a small group.
Worse case scenario would be Linux/BSD backend which saves you money and time. Best case would be Linux everywhere. Most realistic case may be to have small controlled pockets of MS in your network (those pesky vendors who walk in and ask for a dedicated NT/2000 box for their apps) and the rest Linux.
Slashdot Mirror
We did this last year. We were on Exchange 5.5 with Outlook. Just got tired of monthly reboots, zombie Exchange processes, viruses, and those damn Outlook profiles.
We put together a system we called Steamroller. Courier-IMAP, OpenLDAP, qmail-LDAP, Sympa, Squirrelmail, and SunOne Calendar Server. Works like a champ. Adios Exchange. Adios Outlook. Adios crappy problems with no obvious cause and no obvious answer.
Courier-IMAP = maildir IMAP
OpenLDAP = store user passwords and info
qmail-LDAP = grok LDAP for info, virus scanning, SMTP
Sympa = distribution lists, err mailing lists
Squirrelmail = webmail from any computer
SunOne Calendar = uh, calendar
We didn't try to clone Outlook or try to claim that our system was Outlook/Exchange. We distilled what was essential in Exchange and added it to our system and added some refinements.
The big selling point for us was that an email system that goes down is like having no email system at all. We were able to trade some, uh features of Exchange/Outlook in return for rock solid stability.
We were also able to whip up some nice Perl cgi scripts to do administration of the system and a plugin for Squirrelmail that allows the users to have out of office functionality.
We have a HOWTO document if you are interested. You may want to check out QVCS as well
http://www.dulug.duke.edu/~icon/qvcs-guide/
There was a handful of users who resisted the change, however, we worked hard to sell the idea to IT staff, managment and the users. A big plus was having the new system running in parallel for several weeks for users to try it out before we migrated. The migration itself was smooth, a few calls from people who did not attend training, a few 'how do i...' questions and that was it.
Is very handy for our users to go to conferences, training, etc and get access to their same mailbox or calendar as if they were sitting at their desk.
End result was that it removed a PITA administration area and greatly reduced the overhead on our help desk. Nothing like setting up 25 or 30 Outlook profiles on a single computer at a fire station.
I work for a government agency in Arizona and we are ditching Exchange next weekend for a system built from open source components.
S quirrelMail
We expect other government agencies in Arizona and beyond to do the same in the near future.
FreeBSD
Courier-IMAP
OpenLDAP
qmail
Sympa
SunOne Directory/Calendar server
This can be done. The calendar was the hardest thing to find a replacement for.
Take a look at:
Replacing Exchange HOWTO and QVCS.
The first document is what inspired us over a year ago to begin this project. The second project is very similar to what we ended up with. We will be producing a HOWTO next month on how we did this.
Replacing Exchange is not that difficult if you understand how email works and how Exchange is cobbled together. We chose to separate the Exchange functions and put them in a web browser driven context.
The big task you will have is to fight the user conception, built through marketing and fud, that somehow Exchange/Outlook is synonymous with email in the same way that some people see AOL as synonymous with the Internet.
You will have to fight like hell for an Exchange replacement. A replacement has to be feature-rich, a replacement better be secure, and most importantly a replacement needs to be more reliable than Exchange.
If you can do your homework on these issues you should be able to get PHB and upper management to buy in.
The magical thing that Microsoft, and to a lesser extent Lotus and Novell, managed to do is transform the function of email into the monstrousity that Exchange/Outlook is and convince people to lay down gobs of cash money for something which fundamentally is no different than any other email system - its job is to deliver email.
We got tired of Microsoft sticking it to us for licensing. We got tired of virus after virus. We got tired of Exchange problems with no apparent reason and (worse) no apparent cure. We got tired of having our data held hostage by Exchange.
The big question, for you bofhs out there, is whether you can/will do something about Exchange. You can sit idly by while Exchange craps on you again and again or you can do something about it.
I'm the BOFH at a municipal government and we have been working on this for some time.
I would recommend starting with the backend server and network systems. Swapping these systems (DNS, DHCP, email, backups, file, print, etc) out for Linux/BSD systems can be done seamlessly so your users never know (other than everything runs better and faster) and is a smaller task in scope than everything to Linux.
If you can convert your backend to Linux/BSD (Solaris even) you can break the MS stranglehold on your wallets - break down the servers and you can break down the server licensing (sql = nt license, nt cal, sql license, and sql cals) and the CALs which should be a major cash savings.
You should also look at the cost savings, security benefits, and flexibility gained from staying away from proprietary software (PHB love that stuff).
The next step would be to move as many applications to an open source web based model (if you have developers around - may take some work looking at various projects) as you can.
The hard part is going to be the users. Most of them think os = windows. If you can get them used to the idea of computer diversity (Macs may be an avenue to introduce this idea) and get them some exposure to StarOffice/OpenOffice, Mozilla etc you can get them used to the idea of running non-MS applications and non-MS OS.
In our experience once people actually use StarOffice or Mozilla and they see that they are OK and they tend to be pretty open to running OS X, Linux and so forth.
Your ace in the hole may be to find businesses, gubmints, etc in your area that are working on the same thing. It is much easier to get acceptance of using open source when business X or Foobar U is doing the same thing.
We are working with county govt, school districts, universities, etc in the area - they are tired of paying out of their ass for MS and they want something else. Many of their organizations are much larger than we are but we have many of the same problems. The benefit is a partnership where everyone works together and collectively things move much more quickly that in a small group.
Worse case scenario would be Linux/BSD backend which saves you money and time. Best case would be Linux everywhere. Most realistic case may be to have small controlled pockets of MS in your network (those pesky vendors who walk in and ask for a dedicated NT/2000 box for their apps) and the rest Linux.