% fortune -ae paul murphy

Domino vs. Exchange - the personal picture

Lets pretend that you're responsible for email and scheduling for 60,000 employees in 60 functional groups ranging in size from 15 members to 3000. They're all currently using Microsoft Exchange on Windows/XP desktops with Active Directory as installed in mid 2005.

In that situation both the Windows desktop and active directory will be organizational givens you have no power to affect - in fact, you would typically be in the frustrating position that you get no respect for keeping Exchange running while the CIO simultaneously ignores your contributions to the organization and uses the relative stability you've achieved for email and scheduling to justify Active Directory, other Microsoft tools purchases, and the further imposition of centralized management on users.

You would therefore have two complementary sets of reasons for looking at alternative messaging systems. On the business side, the changes coming in with Microsoft's expected transition to its Server 2008 product set look serious enough to trigger your professional responsibility to investigate reasonable alternatives before committing to the upgrades.

On the personal side, messaging products that co-exist well with Active Directory but are functionally independent of it offer significant potential payoffs; most notably the opportunity to raise your organizational profile relative to that of others at your level. If you saved a few bucks by switching to Domino, for example, the organization's cost of running Active Directory would be completely unaffected, but the guy running that side of the business would suddenly have to justify a much larger share of his own costs - because right now you're carrying him, and getting no recognition for it.

For a small business the licensing cost side of this decision can be both important and non obvious - but for an organization of any significant scale the actual dollar amounts involved are too small to matter except in terms of decision politics and the calculation itself can be utterly trivial. In the case of a large organization seriously considering dropping Exchange for Domino, for example, IBM's competitor license trade-in program gives its representatives almost total freedom to match whatever number the Microsoft clique comes up with - meaning that Domino can always be made to net out a dollar cheaper.

Domino has a much more important edge on the hardware side of the cost discussion. There are three complementary reasons for this, all arising from the freedom to choose that comes from stepping outside the Microsoft tarball.

First there's the obvious: both Exchange and Domino are performance pigs, but Domino marginally less so and especially less so if ADsync is used to allow identity management to run on Domino's built-in directory services. As a result our hypothetical organization could switch to Domino on its existing Windows 2003/XP servers instead of upgrading hardware, the Windows OS, and the Exchange servers in 2008/9. Once users have transitioned, it's possible, furthermore, to extend the life of the hardware by another couple of years by switching the servers to Linux and relying on Domino's multi-point backup and recovery capabilities to compensate for the increased failure risk associated with hardware aging.

Note, in this context, that Domino's Domain Monitoring (DDM) facility on Windows or Linux counter-balances the centrist argument that only Microsoft's management tools on Windows are effective enough to trust with messaging servers - and that success in getting this accepted not only removes a significant point of control through which the core Microsoft clique now impacts messaging, but also positions you to bring in Linux as a means of reducing the number of servers needed.

Second, Domino isn't limited to x86 or other Intel hardware. Both Sun and IBM have new generations of RISC hardware either on the way (IBM) or in the market (Sun) and both offer significant performance and reliability advantages over Wintel.

For example, Sun and HP recently filed competing notesbench messaging performance reports in the same week (10/04/07 and 10/09/07 respectively). HP's report covered a ProLiant BL480c server blade (an eight core Intel Xeon X5355 with Hyper Threading at 2.66GHz;) which achieved an average 0.764 second response time with 27,000 nominal users at a unit cost of $13.27. In comparison the Sun T5240 "coolthreads" machine achieved an average response time of only 0.584 seconds (30% better) with 43,000 nominal users (60% better) at an average cost of $2.89 -almost five times better.

Third, Domino's integration magic isn't limited to Microsoft's products. Thus Domino 7 resurrects something from the mid nineties Informix/Lotus alliance: direct access to DB2 from within any Domino process - including control processes. How this plays out depends on what the target organization has in place, but it's worth a close look at what else users need access to, how that access can best be delivered, and whether development decisions or existing applications could be affected by the Exchange/Domino decision.

The bottom line, overall, is quite simple: messaging executives now developing budget documents featuring a future transition to Microsoft's next generation servers have options - options they have a clearly justifiable professional responsibility to investigate as part of the budget development process.

Paul Murphy wrote and published The Unix Guide to Defenestration. Murphy is a 25-year veteran of the I.T. consulting industry, specializing in Unix and Unix-related management issues.