With the release of Windows Server 2008 approaching next month, many IT managers are faced with determining the costs and benefits of a migration.
But IT managers must also determine when to do this migration. Remember that in Microsoft's Lifecycle timeline for Windows Server, the clock is ticking on Windows Server 2003, and that makes migrating to Windows Server 2008 inevitable unless you switch to Linux.
When conducting a return-on-investment analysis for Windows Server 2008, there are some important points to remember to ensure that your analysis is realistic as well as accurate.
A number of years ago, I attended a conference where the speaker was claiming that a move to Windows 2000 from NT could result in a savings of about $2 million per year based on improved power management alone. His reasoning was that this new feature would allow a faster boot process, saving two minutes a day. Multiply two minutes a day times the number of days, times the number of employees, times average salary and voilà -- an annual savings of $2 million. Although this is an extreme example, it demonstrates faulty reasoning that can sabotage an ROI analysis.
There are several points here that show how to avoid using phony numbers:
- Use realistic data. There are a lot of better ways to calculate savings than power management. This is funny money. Use something that demonstrate "hard savings" -- something you can tie directly to labor savings, such as reduction in downtime or reduced support costs.
- Use data that you can verify. In the example, even if it were true that power management would save time in bootup and logon time, how could you verify that every person in the company gets two minutes a day more work and that translates into more productivity that results into cash? You couldn't.
- Find features that will actually benefit your environment. If you aren't using Terminal Services, then new Terminal Services features won't help an ROI calculation.
- Stick with measurable savings. Labor costs, reduction in help desk costs, reduced operating system crashes and other events that reduce downtime can be directly tied to productivity increases.
The bottom line is that IT managers have to use data that they can be held accountable for later. For instance, Windows 2000 Server eliminated hundreds of potential causes of bluescreens and about 75 issues, such as maintenance and configuration operations that required a server reboot. Server crashes and reboots disrupt user productivity and can cause the loss of data – that's a reasonable calculation and it will hold up.
Another aspect of ROI calculation is new income -- not just cost savings. Consider how new features could increase income.
I once presented a proposal for a network upgrade to a company president. I was confident of my ROI calculations and that this would benefit the company. After looking at the numbers, he asked, "If I give you this money for the upgrade, how many people can I lay off?" I told him that wouldn't be necessary because it would increase our manufacturing capability and would generate more income. I knew there was a contract we were working on that would directly benefit from this upgrade. He agreed and gave us the money. The payback period was about five days.
In another case, a company had an AD site in every state. Trying to save money, it had only one server in each site that was a global catalog server, Exchange server and file and print server. Frequently, a DC failure would cause the server to be re-promoted causing disruption of Microsoft Exchange. Or Exchange failures took the DC out of service. All of these instances rendered file services to fail.
Several times the company actually shipped the server to the home office so the software problems could be fixed. Calculating a return on investment to justify adding even one server to each site would be easy because of the loss of productivity and disruption of email, not to mention support and shipping costs.
These types of savings exist -- you just have to find them. Look for projects, applications and problems that will benefit from Windows 2008 features:
- New Terminal Services features can make data access easier and more secure.
- By putting read-only domain controllers in remote sites, remote users can get better access to data and more reliable authentication.
- New Hyper-V virtualization technology can be used to consolidate servers.
- Network Access Protection can prevent costly viruses from invading the network.
- Powerful PowerShell scripting for administration benefits can have a direct impact on cost savings.
Of course, it isn't always a slam-dunk. New licensing fees, hardware upgrades and -- in many cases -- a requirement to upgrade to Windows Vista to realize the benefits can dampen perceived savings. But, with a little ingenuity, many other features in Windows Server 2008 can be exploited to enhance partner access and income generation.
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. Olsen is a Microsoft MVP for Windows Server-File Systems.
