Microsoft Exchange Server 2007: Undisputable Champion?

Can Open Source replace Microsoft Exchange?
Now I'm busy on research for my company messaging & collaboration system. Working on small company is quiet painful. The bosses ask full feature with less even no budget. It seem i'm heading open source (free source actually). This article is written by Steven J. Vaughan-Nichols posted on LinuxWorld.com as my reference why almost only one name matter in this field.
Once upon a time at a NASA space flight center a long way away, I was an e-mail administrator. At the time, the 1980s, e-mail was still chaotic. The RFC 822 standard was only beginning to bring rhyme and reason to e-mail. One of RFC 822's competitors, the Common Messaging Calls (CMC) X.400 standard, wasn't making much progress, but then Microsoft adopted it in 1992, added the concepts of folders to it, and re-named the result Mail Application Programming Interface (MAPI). And, ever since, the e-mail world can broadly be divided into two camps: the RFC 822 Internet compliant e-mail group and the MAPI-compliant Microsoft Outlook/Exchange pack.

Many of us assume that all e-mail works by using such RFC-822isms as e-mail addresses that look like "name@SomePlaceOrTheOther.com." Not so. MAPI takes a quite different approach. In addition to simply handling e-mail, extended MAPI and Collaboration Data Objects (CDO), which became Microsoft's default protocol set in Exchange 2003, added the power to manage calendars and addresses. So it is that Exchange and Outlook, while primarily used for e-mail, is also a groupware package. And, I might add, a very popular one.

A recent survey by Ferris Research revealed that Exchange has about 65% market share across all organizations. Lotus Notes/Domino is a distant number two with 10% of the market. POP/IMAP, (Post Office Protocol/Internet Message Access Protocol), the usual way incoming RFC-822 mail is handled? All the dozens of RFC-822 mail servers, including Sendmail, Qmail, and Postfix combined, have only 15% of the business/organization e-mail market.



As for the open-source groupware servers that try to directly compete with Exchange, such as Scalix, Open-Xchange, and Zimbra, Richi Jennings, a Ferris Research analyst, dismissed them as being mere 'noise' in the business e-mail market. If you look outside the U.S., it's a somewhat different story. Sarah Radicati, CEO of The Radicati Group, estimates that in EMEA (Europe, Middle-East and Africa) open-source e-mail servers, lead by Germany-based Tobit and Axigen Bucharest, Romania have about 10% of the business e-mail market.

Still, Radicati also estimates that Exchange is the dominant e-mail/groupware server with 37% of the world e-mail business market. In other server areas, open source has made great gains against Microsoft's and the Unix vendors' proprietary programs. You only need mention Apache, Linux and MySQL to see this.

Why is it that in business e-mail that while open-source e-mail servers are very popular with Internet service providers, they've been unable to make any substantial gains against Exchange? One of the reasons is that Microsoft Outlook has been, continues to be, and looks to remain the e-mail client of choice for businesses. While open-source end-user applications in other categories, such as Firefox for the Web browser and OpenOffice for the office suite, have made impressive inroads on Windows desktops, open-source e-mail and groupware clients are still niche products. Even Thunderbird, Mozilla Messaging's answer to Outlook and the most popular open-source e-mail client, has made little progress in denting what's estimated to be Outlook 80%+ business e-mail client marketshare. That's because Mozilla has largely neglected Thunderbird in favor of Firefox. Even now, Thunderbird's security fixes tend to lag behind Firefox's fixes, and while Firefox's developers are looking well beyond the current 3.0 version, Thunderbird has only released its second alpha for version 3.0. Without a viable, open-source e-mail client contender for Windows, Outlook remains businesses' e-mail client of choice—despite Outlook's infamous security problems.

Since that's the case, any rival to Exchange has to be able to communicate with Outlook in the same ways that Exchange does and duplicate Exchange's functionality. That's not easy. Jennings said that open-source developers have tried three different ways to do this. "The simplest way is to run a background task on the PC that does a sync with the mail server, probably using IMAP, and then outputting the result to an Outlook Personal Storage Table (PST) file." Bynari, to the best of his knowledge, is the only program that uses this approach.

Another path to Outlook is to "reverse-engineer Exchange's wire-protocols," says Jennings. In this approach, "you forget about installing software on the desktop and try to make the server look to the client just like Exchange. Outlook really does think it's talking to Exchange. The only server that tries to do that is PostPath. One advantage to this route is that by not requiring any client-side software, installation and maintenance costs are reduced. While PostPath doesn't have all of Exchange's fancier features, it does work well for providing Outlook with the basics of integrated e-mail, calendar, and contact management.

The most popular approach, though, which is used by both Scalix and Lotus Domino/Notes, is to use a "MAPI plug-in on the client," says Jennings. "This hooks into the MAPI subsystem and then to Outlook. In theory, you're not fooling Outlook, you're using the designed way in to it to talk to an alternative server. The MAPI sub-system, however, leaves something to be desired, because the MAPI documentation was, until the European Commission forced Microsoft to open up its network protocols, a mess." While many people are aware that the European Commission's decision opened up Microsoft's Common Internet File System (CIFS) and Active Directory protocols to Samba, few have picked up that MAPI was also opened up. There is an open-source project, OpenChange, working in conjunction with Samba to build open-source implementation of Microsoft Exchange Server and Exchange protocols from this information.

But, curiously, no commercial company is following up on this work, according to Radicati. It can only be a matter of time, though, before an existing Exchange-replacement company, or perhaps a new one, starts working with OpenChange. As it is, several Linux distributions, including Fedora and Debian incorporate OpenChange in future editions. OpenChange, while already usable, won't be ready for prime-time though until Samba 4, which will include Active Directory and and Heimdal Kerberos network user authentication is released. Unfortunately, Samba 4 is still in alpha. Even with the opening of Microsoft's protocols, Samba 4 isn't expected to see the light of day in 2008. So, for now, any would-be Exchange replacement has to rely on reverse-engineering and an imperfect understanding of Exchange's protocols. Another problem, Jennings points out, is that Exchange is a moving target. "Exchange keeps getting better," Jennings points out.

That said, Exchange is also growing ever more costly and not everyone wants Exchange 2007's unified communications, which now includes everything from e-mail to voice-mail. Cost is often cited as the number one driver for companies that switch from Exchange to an open-source alternative. However, that hasn't proven sufficient to move many customers from Exchange. Jennings said, "Various people have been pushing this idea for years, including yours truly, but it hasn't ended up with a lot of traction. Open-source e-mail servers have remained a niche thing." Why is that? One reason is that none of the open-source programs are really ready to serve as drop-in Exchange replacements. There's also some additional work that needs to be done, and it's not work that Windows administrators are used to doing.

Even a veteran Linux administrator, though, might find setting up a full-powered Exchange replacement for a good-sized company a challenge. For example, Scalix 11.4 requires Apache, PostgreSQL, Tomcat, and either Sendmail or Postfix to be installed before it can work. That's not hard, but when you factor in the need for managing disk performance it becomes more of a problem. E-mail server applications, have trouble scaling, because of disk performance bottlenecks. To run a groupware server for more than a small business really requires shared disk arrays. Put it all together and you have a serious Linux system administrator's job, and it's not one that a former Exchange administrator is likely to be able to handle.

So, while in the long run, the total cost of ownership is likely to favor a Linux/open-source solution, in the short run, getting an IT department to switch over from an Exchange-centric e-mail system isn't easy. The answer? Well certainly advances in OpenChange and adoption of its work by other projects will help smooth out any remaining Outlook compatibility problems. Perhaps, an even more vital change to help open-source Exchange replacements' adoption though will be to make installing and maintaining them easier than it is today. Are the open-source Exchange replacement programs good enough? Yes, I have no question that they are, but it's not just being good enough, or even cheap enough. It's about making it easy to give Outlook users a 99.9999% Exchange-like experience while making it easy for experienced Exchange administrators to deliver that experience to them. Only when these can be delivered will programs like Open-Xchange, Scalix, and Zimbra become important to business IT.

In the meantime, both Jennings and Radicati believe that Google's Gmail and the hosted Zimbra Collaboration Suite are more likely to replace Exchange. That's because they're not actually replacing Exchange, instead they're replacing Outlook with a Web-based e-mail interface. It may not be open-source, but software as a service (Saas) that actually ends up replacing Outlook/Exchange. Of course, Microsoft is aware of this and they have their own SaaS Exchange plans. With that being the case, open-source groupware servers may yet have their chance to make the kind of impact in business e-mail that open-source programs already have in operating systems, Web servers and databases. Trite as it may be, this really is a case where only time will tell.

Next, i'll post my step by step journey to find the ideal & afforable solution for my small beloved company.