|Common sense about Windows 2000|
|Written by Maggie BiggsHits : 768|
|Wednesday, 31 May 2000 19:00|
Call me a practical sort of gal, but I believe in ascertaining the business benefits of a technology before considering whether to adopt it. Far too often these days I see colleagues of mine in business and in IT who jump onto a particular technology bandwagon because something is new or highly hyped in the media--or even for emotional reasons.
Such is the case with Windows 2000--Microsoft's new server and desktop operating system--which was finally delivered this spring after much delay. You've undoubtedly heard Microsoft's heavy-duty marketing messages, read articles on the topic, or seen advertisements for the massive operating system.
By contrast, you've probably also heard or seen opposing viewpoints or articles on the Windows 2000 release. Some will not adopt the release because they have concerns over Microsoft's alleged monopoly status. Others knock the new Windows version as being too expensive or too client/serverish with Web extensions. Still others feel that the new version locks the customer into Microsoft-only technologies.
With all of these differing thoughts, messages, and opinions, it is difficult to determine the business benefits and detriments of installing or upgrading to the new version of Windows. But time is on your side.
Why? Well, computer industry research firms, such as International Data Corp. (IDC), and GartnerGroup, have estimated that more than 70 percent of companies will skip this release of Windows 2000.
The estimates are based in part on the fact that Windows 2000 is not just an upgrade to Windows NT. More than 50 percent of the operating-system code found in Windows NT was reworked for Windows 2000. In addition, developers added thousands of lines of new operating-system code.
Industry wisdom says that these huge, sweeping changes and additions to the Windows code base will yield bugs, issues, and instabilities. In addition, current Windows versions will be supported for some time to come (through 2004, according to IDC).
These things being the case, this is a great time to closely assess whether your company needs to adopt Windows 2000.
Pondering the Pros
A good portion of your assignment is to examine how well Windows 2000 maps to your business requirements. Microsoft has said that its design goals for Windows 2000 are a direct result of customer input.
Chief among Microsoft's stated goals for Windows 2000 are increasing manageability, reliability, availability, and scalability. It will be up to you to determine if these new additions to and changes in Windows 2000 are viable and if they represent functionality important enough to warrant a commitment to the new operating system.Administrative ease: Windows 2000 does include a hefty amount of new and enhanced features that are geared toward improving administration functions. Some of these include a common interface (called MMC) for all administrative tools, improved monitoring capabilities, and a new directory service solution called Active Directory.
Active Directory in particular deserves a good bit of the focus during your company's assessment of Windows 2000. Until now, many companies have leveraged directory services as a means of better managing network resources.
Renewed interest in directory services is occurring because many people believe the same directory-service management qualities can be applied to business applications to increase efficiency and decrease costs.
IDC predicts that native LDAP directory services will gain the lion's share of the directory service market in the next five years because LDAP is an open Internet standard. The IDC report also predicts that in the same five-year period, Novell's more mature eDirectory solution will garner the next biggest chunk of the marketplace.
The report also shows that Microsoft's Active Directory is expected to gather about one half the number of licenses as eDirectory in the same period. The story behind the numbers is that Microsoft's delay in delivering Active Directory caused many businesses to bank on LDAP or eDirectory. Furthermore, this first release of Active Directory contains issues and problems that will take some time to iron out.Reliability: Microsoft has listened well to its customers on the subject of reliability. Many people have complained about sudden crashes and the famous "blue screen of death" in prior Windows versions. Windows 2000 attempts to address these concerns.
The new Windows version includes improved error-handling mechanisms and protected subsystem technology that helps prevent a crash when a single process is having trouble. When a crash does occur, automated restart features will help preserve conditions prior to the crash and restart the system in an unattended mode.
Windows 2000 also supports disk mirroring, disk duplexing, and disk striping. Support for distributed file systems and protection for essential files should improve reliability (and an end to "DLL Hell," according to Microsoft).
Increased uptime is a positive outcome for those considering Windows 2000. Microsoft has gone beyond its two-node cluster limitation in previous Windows server versions. Included administrative tools ease cluster set-up and management.Clustering: Microsoft's Internet Information Server, Transaction Server, and Message Queue Server work well with the clustering technology provided. You might use the clustering capabilities to increase performance or to insure the uptime of critical business applications that leverage Microsoft technologies. Scalability: Windows 2000 does show marked growth in the area of scalability. Previous Microsoft server versions have not fared well against other enterprise-grade network operating systems. The company has responded to these customer concerns in several ways.
First, Windows 2000 (Advanced Server) supports up to four-way symmetrical multiprocessing (SMP) systems in retail form and up to 32-way SMP systems in a third-party arrangement. Tuning parameter support for CPU, I/O, and memory will help those who need to improve performance.
Windows 2000 Advanced Server and DataCenter versions also support the addressability of greater amounts of physical memory. Job objects, load balancing, I2O (I/O subsystem improvements), and other new features aim to take Windows 2000 further up the food chain in enterprise settings.
Considering the Cons
Thousands of Microsoft developers spent countless hours creating Windows 2000 with all its stated improvements. Are there things you need to consider before deciding on whether to adopt the release? You bet!
I mentioned earlier that Windows 2000 contains a huge amount of new and reworked operating-system code. That alone should make you spend significant time evaluating the release in a test environment before making a decision. Several service packs will be needed to flush out serious problems with the initial release. Also, compatibility and training issues could result in sizable hidden upgrade costs.
Hidden costs: Put simply, you will need to invest in new hardware or significantly upgrade existing equipment before adopting the release. Windows 2000 requires significantly more hardware to run efficiently, whether on the desktop or a server.
You'll also need to invest heavily in server and client licensing, as well as support costs. GartnerGroup reports that upgrading a single Windows NT server to Windows 2000 will run you nearly $56,000, while migrating 50 users to Windows 2000 Professional (the new Windows business desktop) will cost more than $100,000.
Other GartnerGroup cost estimates show that moving from Windows 9.x on the desktop to Windows 2000 Professional will cost between $2,015 to $3,191 per desktop. Those who choose to move from Novell 4.x on the server side to Windows 2000 should expect fees of approximately $430 per user.
If you have to stick to budgets the way I do, these numbers are not small change, and could cut into your pocketbook significantly. Careful assessment of these steep costs versus the final outcome achieved is definitely warranted.
Some of you may be tempted to upgrade to Windows 2000 just because you've already invested in Microsoft technologies in the past. Given the costs and the wholesale changes introduced in Windows 2000, this is an ideal time to examine other options on equal par before deciding whether to continue your Microsoft investment.
While considering your budget with Windows 2000 in mind, know that extensive reworking of your network infrastructure will be needed (because of Active Directory). You will need to replan your existing Windows NT network, and industry estimates indicate a six-to-nine-month planning period before beginning any implementation of Windows 2000.
To support this re-engineering, your network administration staff will require a hefty amount of re-education to comprehend the massive changes and how to successfully manage a Windows 2000based network. You should also plan on training costs for your end-users since they will need to understand how Windows 2000 changes their day-to-day computing activities.
Plus, the number of Windows 2000 natively supported applications is small at this time. Most major vendors are planning to support Windows 2000, however. Their product release schedules may differ, so be sure to confirm the availability of all of the business applications you need before moving ahead.
Another issue to think about is whether you want to run your Web strategy atop a client/server architecture as opposed to a pure Web play. Furthermore, assess carefully how much you really need to support thick clients. Your business applications could be serviced in a flexible, thin-client, Web-based metaphor that might save money and time, and prove more efficient in today's rapidly changing business computing model.
If you want to take a practical approach to considering Windows 2000, this is the year to evaluate the release, determine if it meets a specific business need, calculate the expected costs, and decide whether your company should install the new operating system version from Microsoft.
Maggie Biggs is director of the InfoWorld Test Center, where she evaluates emerging technologies and writes the Enterprise Toolbox column. She has more than 15 years of business and IT experience in the financial sector.
Should I install or upgrade to the first version of Windows 2000? If you determine there is a business case for Windows 2000 at your company, you'll need to factor in both the benefits and detriments of Microsoft's new operating system before making a decision.
PROS Increased administration capabilities over previous Windows desktop and server releases Automated/incremental deployment tools Personalized desktops regardless of location Improvements in scalability and availability when compared to previous Windows versions Added functionality to support application serving Reliability enhancements Support for new types of hardware
CONS Majority of Windows 2000 code-base is new or changed (more than 50%) from previous Windows NT versions. Strong likelihood of major bugs and instability Requires the purchase of new hardware or major upgrade expenditures for existing hardware Licensing and supports costs per Windows 2000 server are very high when compared to other options Licensing and support costs per desktop are very high when compared to other options Requires a commitment to Microsoft technologies Operating-system methodology based on client/server technologies versus current and future Web-based approaches (although Web integration is supported) Fat-client approach may be unnecessary given the move from client/server to server-based, thin-client Web-centric business computing Requires extensive re-engineering of the company network infrastructure Requires extensive re-training of network administrative personnel Added costs to re-train end-users must be factored into your cost/benefit analysis Current Windows server and desktop versions are expected to be maintained through 2004 Small number of natively supported applications available during 2000 (more are expected)
Suppose you begin evaluating Windows 2000 this year, only to discover cost parameters that are well beyond your budget. Maybe you don't want to invest in totally re-engineering your network. Or perhaps you're ready to dump client/server technology in favor of thin-client, server-centric, Web-based computing. What other server and client options might you consider?
For starters, if you're running Windows NT 4.0 on the server side, you needn't adopt Windows 2000 just because you're afraid your network operating system will become obsolete. Official Windows NT support is expected to continue well into 2004.
Also, the operating system won't quit working on some magic date. If you want to continue with a Microsoft-based approach to business computing, perhaps budgeting for a Windows 2000 adoption further out is the best approach, given the costs.
If you do want to consider other non-Microsoft approaches, there are several ways to go. One of my favorites is to leverage IBM's AS/400 platform. There are a bevy of reasons to consider the AS/400. For one, the AS/400 can concurrently run its own operating system (OS/400) as well as Microsoft Windows NT. For a lower entry point, you gain a mature server platform that can meet the demands of legacy, client/server, and Web-based computing. The AS/400 includes not only the operating system, but relational database technology, Web application serving capabilities, and much more.
A wide variety of client platforms can talk to the AS/400, including wireless devices. What's more, the AS/400 interoperates well with other operating systems, and can scale well beyond Microsoft Windows.
You might also consider running Novell's NetWare. Already considered to have mature directory services (and file and print serving), NetWare's newest release shows dramatic improvements in application serving and scalability.
Unix operating systems, such as Sun's Solaris, already run the majority of services on the Internet. If you intend to implement a Web-based approach to business computing, an evaluation of one or more Unix variants is advisable. The same manageability, reliability, availability, and scalability is inherent in these network operating systems.
And then there are the major fans of open-source network operating systems such as Linux and FreeBSD. You'll still need to invest in support costs, but these network operating systems are also dialed in to Web technologies, and are fully capable of handling day-to-day business computing at a fraction of the cost of other approaches.
Networking and application support thrive on these platforms because of continuous improvements by the open-source community. Open-source technologies are still pushing into the extreme high end of business computing. Today, they are ideal on the front end and middle tier of corporate computing environments. And in a short time, they will take on other network operating systems on the back end.
Perhaps you are concerned with abandoning Windows desktops in your business computing environment. A few years ago, that used to be a legitimate concern, but it no longer needs to be. There are other types of client platforms and devices that can provide the same level of support.
Certainly, you may need to change business applications to adopt new platforms, but the lower cost and added flexibility is well worth investigating. For example, several intelligent wireless devices and handhelds now support Web access, e-mail correspondence, complete office suites, transaction-based business applications, and much more.
You might consider implementing lower-cost network appliances that support the same application base, but in a server-centric mode. Or maybe Apple's Macintosh on desktops and notebooks may suit your purposes just fine.
The open-source community is also making great strides toward end-user operating system versions that are easy to use, affordable, and able to support all major business applications. Several companies are helping this process, but Corel's Linux for the desktop is a good indication of where Linux is headed with regard to the end-user. It is well worth evaluating, and it might even remind you of Windows.
The worst mistake a company can make is to not examine all available options before deciding on when and how to change an existing computing infrastructure. Will you look into all of the possibilities?