Figuring out when to migrate your computing environment to 64-bit technology can be difficult. Knowing how to do it can also be tricky.
Besides understanding the technical aspects of moving to 64-bit, it's important to make a business case for the migration. To do that, first you'll need to evaluate the technical aspects of the technology to determine the business value.
Let's take a look at the decision points and the business value.
In determining the value of upgrading client workstations, consider 64-bit Windows Vista on your refresh program. The x64 systems may be just as cheap as the 32-bit systems.
The apps may not be there yet, but they will be. And when they are, you'll be ready. In addition, if you buy x86 hardware now, by the time your next refresh comes around, it will be antiquated.
Warning: The big gotcha here is device drivers, which cannot run in 32-bit Windows On Windows emulation mode. Before you upgrade clients, make sure that any special requirements you have for device drivers have 64-bit versions. You can't make a 32-bit driver work.
Probably the greatest business value you will have for investing in 64-bit technology will be your servers. They are the ones that are largely responsible for worker productivity. The reliability and performance of 64-bit servers are critical.
The large memory model and virtually unlimited resources in 64-bit servers can go a long way in preventing downtime because of poor server performance. However, consider that most hardware vendors have benchmark numbers for 64-bit systems, but benchmarks are made in an artificial environment. Get an evaluation system and run your own benchmarks – using your apps in your environment. Make sure you know that spending that extra money will indeed give you the benefits you expect.
One critical point to note in evaluating 64-bit servers is that there are two options in the Windows world: Itanium (IA64) and x64.
Itanium-based servers are typically going to run large database applications, computer animation and probably virtualization. These are high-end, expensive machines that will require applications to be able to take advantage of that power to justify their purchase.
The x64 is a lower-cost alternative to IA64. Although the x64 technology is still 64-bit and addresses large-scale memory models like Itanium, it does have limitations in terms of performance and capacity. These servers will be more likely to run everyday business applications and provide basic file service and printing applications. Client workstations will be available only with x64 processors.
The decision, then, is not only whether you go to 64-bit technology but also whether you go to x64 or Itanium. Again, the applications on Itanium are going to be very high end, and there will be few -- if any -- Itanium applications that most organizations will be able to justify. Nevertheless, Itanium does have its place, so it should be evaluated.
In terms of your business case, it really depends on your needs. But rule of thumb would be to start with x64 servers and prove that you need more. Doing a proper benchmark and evaluation is of paramount importance in justifying Itanium.
One place that 64-bit Windows servers will definitely help is in Active Directory. With appropriate physical memory, 64bit domain controllers can literally store the entire active directory database in memory. I know one large, global corporation whose AD database is just under 8GB. So with a reasonably priced x64 server, you can load AD into physical memory and make a huge impact on Active Directory performance.
This is another case where x64 servers are better than Itanium servers because AD can't justify the Itanium resources. Although this will be somewhat hard to quantify for a return on investment calculation, it is a good investment in your infrastructure, which everyone depends on.
It's clear to see that 64-bit is the wave of the future. So begin now to get educated on the technical aspects so you can make informed decisions on how to take advantage of this new and powerful technology. As always, make sure the migration decisions make good business sense -- not only in terms of initial cost but also in terms of performance, reliability, and the ability to host new apps that will gradually abandon 32-bit platforms.
Gary Olsen is a systems software engineer for Hewlett-Packard in Global Solutions Engineering. He authored Windows 2000: Active Directory Design and Deployment and co-authored Windows Server 2003 on HP ProLiant Servers. Gary is a Microsoft MVP for Directory Services and formerly for Windows File Systems.