This week at Microsoft TechEd 2003, more than 10,000 IT administrators, professionals, and developers are convening in Dallas to learn how the products in the Windows Server System, combined with those in the Microsoft Office System, will not only save them time and money but also, presumably, cure cancer and solve the energy crisis in one fell swoop. OK, maybe that's a bit of an exaggeration. In his keynote address Monday morning, Microsoft Senior Vice President Paul Flessner pointed out several times that his company's products still have plenty of room for improvement--but listening to the bizarre variety of servers, services, and applications that Microsoft will foist on unsuspecting enterprises this calendar year, I was struck by how confusing (and expensive) the entire situation has become.
As I've discussed previously in Windows & .NET Magazine UPDATE, the cost of Microsoft's server products can quickly add up. After a Microsoft Exchange Server 2003 Reviewer's Workshop Sunday, I told Exchange Server Group Product Manager T.A. McCann that most of Microsoft's customers probably share a similar decision-making process before licensing any of the company's products. First, they determine all the benefits Microsoft's various products will supply, tack them into a list, and tally the cost. Then, they start trimming back on the least essential products until the price matches their budget. This somewhat backward approach to licensing is unfortunate but necessitated by Microsoft's high prices. Let's hope the company can rectify this situation over time.
Adding to the burden is the problem of convincing companies that they need new versions of Microsoft's products and services. Certainly, a hefty percentage of Exchange users, for example, are still using Exchange Server 5.5 running on Windows NT 4.0; Microsoft says the figure could be as high as 60 percent. This market, presumably waiting to upgrade, has a confusing challenge ahead.
For Exchange 5.5 users wanting to upgrade to Exchange 2003, I have good news and bad news. The good news is that the upgrade is simpler and less error prone than the upgrade to Exchange 2000 Server. The reason is that the Exchange 2003 setup routine won't let you continue until you meet all prerequisites; Exchange 2000 doesn't check for prerequisites, often with disastrous results. The other good news is that you can mix and match Exchange versions in one environment. For example, you can run Exchange 2003/Windows Server 2003 boxes, Exchange 2000/Windows 2000 boxes, and Exchange 5.5/NT 4.0 boxes, in the same topology, interoperating together.
So what's the bad news? Determining which servers run on which OSs is a confusing and daunting task. You can't run Exchange 2003 on NT 4.0, so you'll have to install Exchange 2003 on a new server OS, probably Windows 2003 (likewise, you can't run Exchange 5.5 on Windows 2003, go figure). The upgrade to Exchange 2003 will require new hardware and a migration-style upgrade rather than an in-place upgrade. Even this situation isn't totally horrible because you can use a new Move Mailbox tool to move Exchange mailboxes to the new version relatively easily. And with Exchange 2003's server consolidation capabilities, you'll likely be able to consolidate numerous older servers into one modern box, which will be easier to manage and less expensive.
Accurately showing this information in a text-based newsletter is difficult, but here's a rough product matrix that attempts to explain which Windows versions can run which Exchange versions:
- Windows 2003: Exchange 2003
- Win2K Server Service Pack 3 (SP3): Exchange 2003
- Win2K Server: Exchange 2000 SP3, Exchange 2000 SP2, and Exchange 5.5 SP3
- NT 4.0 Server: Exchange 5.5 SP3
So why might a company decide to upgrade to Exchange 2003? In this release, Microsoft has significantly updated Outlook Web Access (OWA); it's virtually indistinguishable from Outlook 2003. This change means that enterprises can save money by not rolling out a new Outlook version, and OWA users can benefit from the larger reading pane, spell checking, server-side rules, 80 percent better performance than OWA in Exchange 2000, and numerous other new features. Exchange 2003 includes improved versions of all the features found in Mobile Information Server (MIS) 2002, including a new Outlook Mobile Access (OMA) tool that lets users access their email, calendar, and contacts from virtually any mobile device, including Pocket PCs, Smartphones, or virtually anything else with a Web browser.
In a nod to one of the company's biggest customer requests, Exchange 2003 now supports HTTP Secure (HTTPS) connections to the server from Outlook 2003, negating the need for expensive or complicated VPN solutions. Other good reasons for using Outlook 2003 with Exchange 2003 include a new cached mode, which silently copies a local copy of your Exchange mailboxes to the local machine so that you always have a quicker, more easily accessible copy, especially when offline. Also, network bandwidth requirements between Outlook 2003 and Exchange 2003 are significantly lower than before, resulting in faster performance. And when you combine Exchange 2003 with Windows 2003, you gain the benefits of the better performing platform, as well as enhanced versions of backup and restore functions that use Windows 2003's Volume Shadow Copy Service (VSS) to provide Exchange database snapshots and backups that don't require the server to go offline.
The largest problem with Microsoft's big happy product family is expense. Because of the company's mad dash to integrate features across several products, only those customers who can afford the full meal deal--in this case, Windows 2003, Exchange 2003 and Office/Outlook 2003--will realize all the benefits the new "system" supplies. For those who can afford only part of the product matrix, the capabilities drop off quickly, as does the resulting cost/benefit ratio. Microsoft's marketing plan essentially boils down to the old adage that you have to spend money to save money. And I'm not sure that's a path many enterprises are prepared to travel.