Approaching the Linux Desktop

Novell Linux Desktop Cool Solutions Article

by The Novell Solution Creation and Marketing - Linux Team

November 17 2004

Table of Contents

Approaching the Linux Desktop: Enterprise Ready or Not?


Is Linux* enterprise ready? Does this sound familiar? We've been asking the same question about Linux for almost a decade.

When Linux first started making its way into the enterprise as early as 1997, many organizations and market analysts were skeptical. They watched as early adopters made the move. Google (1998) created its own Linux distribution to run its search engine-the core of the Google enterprise-in a 4,000 PC-node cluster (currently up to 100,000+ servers). Burlington Coat Factory (1999) began with 1,000 Linux PCs and now has 7,000 across 350 stores. Other early adopters included the United States Postal Service,, Morgan Stanley, and Siemens Business Services.

Where Linux was once relegated to supporting the edge of the network (DNS, FTP, file, print, and Web servers), it is increasingly used to support mission-critical applications. Now that Linux is receiving such positive mainstream attention, chances are good-that even if you are a late comer to the Linux movement-your organization is beginning to recognize the value of Linux in the data center. You may even be among those organizations taking advantage of Linux as an enterprise platform. In fact, Linux has become so accepted in the data center, even as the core operating system, that we are now asking the obvious companion question: The Linux Desktop-is it enterprise ready or not?

Currently, most market analysts advise caution in making a wholesale move to Linux on the desktop: caution in making a move just to spite proprietary vendors; caution in too generously estimating ROI and underestimating TCO; caution in assuming that all desktop environments can be, or should be, moved to Linux; and caution in assuming that switching operating systems will automatically resolve desktop security and management problems.

But even against this thoughtful advice, there are early adopters who are using Linux-including the Linux desktop-across the enterprise: the City of Munich, Ernie Ball, Novell-and don't forget that both Google and Burlington Coat Factory included PCs in their original Linux deployments.

And other companies are considering a Linux desktop; for example, according to a Forrester survey, 45% of respondents who use or plan to use open source say that they are either already using Linux on some desktops or they plan to use or evaluate Linux desktops within the year.1

The organizations cited above as well as other organizations that are currently reaping the benefits of open source didn't wait for the market to tell them when the Linux desktop might be ready for their enterprise. They evaluated-for themselves-what benefits the Linux desktop might bring to their organizations and then acted.

The purpose of this paper is to help organizations do just that-evaluate the Linux desktop against their own enterprise needs-market hype aside.

Evaluating Enterprise - Readiness

Just as Linux began its tenure in the data center by filling specific roles, the Linux desktop has begun to move into segments of the desktop environment that are particularly well-suited to its enterprise-class strengths: task-oriented workstations, help desks and call centers, developer environments, and point-of-sale terminals. But some organizations, such as Ernie Ball and Novell, have also found the Linux desktop to be a viable solution in their general corporate environment-on every desktop-with a few exceptions (but only a few).

Because each organization's desktop needs and requirements are unique, it's important to approach a Linux desktop evaluation openly-by considering a wide variety of adoption opportunities. Gartner has encouraged organizations to "fight the temptation to make Linux desktops an 'all or none' decision, and to select the right users to migrate."

A thorough evaluation of the organizational structure of your enterprise will ensure that you have indeed selected "the right users to migrate." Use the guidelines outlined in this paper to help you evaluate the fit between your enterprise environments and the Linux desktop:

Analyze the organizational structure of your enterprise

Begin your evaluation by analyzing the organizational structure of your enterprise from various angles so that you can get a holistic view of the desktop environment. For example, you might create a graphical representation of your organization based on clear dividing lines, such as business units, geographical location, or LAN or WAN topology.

This view will be useful for segmenting users according to functional job roles, but it may be insufficient for understanding how user groups interact with each other. It's probably best to base your analysis on several organizational views of your desktop environments.

It's often useful to take a three-tiered approach to the analysis, starting first with processes, particularly processes that cross organizational (business unit) boundaries; then the applications that support those processes; and finally the application data that needs to be exchanged. Completing such an analysis will help illuminate and isolate dependencies that will impact, limit, or hinder your migration.

Much of the data you need may already exist if your help desk (or other) group is responsible for maintaining default desktop images for your organization.

Other views that might be useful in your evaluation include the following:

There may be other ways to organize and analyze the various groups within your enterprise that will become apparent as you work with the suggestions above. Having this holistic view of your desktop environment helps you make sure your decisions are based on real-life business processes and the day-to-day activities of your user communities. It also provides the foundation for the rest of your evaluation process.

Evaluate and prioritize selected environments

Begin evaluating your user and desktop environments by focusing first on environments that other organizations have found to be well-suited for Linux–namely, task-oriented workstations, help desks and call centers, developer environments, and point-of-sale terminals. One of the next logical groups to consider is workgroups that are self-contained (don't have significant interaction with other groups who may be using different processes and applications). But just because adopting Linux desktops in these environments proved beneficial for other organizations does not guarantee they will be beneficial for yours. Pay careful attention to the needs and priorities of your own enterprise and how well Linux meets your requirements.

The same recommendation applies when considering Linux as an alternative for knowledge workers. While only a few organizations have embraced Linux desktops across their entire user environment, this does not automatically mean that your enterprise will come to the same conclusion. Even if you decide that now is not a good time for a wholesale migration for your entire user population, you'll have a clearer understanding of your future requirements.

Priorities will, of course, differ from enterprise to enterprise. Consider the suggestions below in identifying your priorities for ranking user and desktop environments:

Document both the positive and negative aspects of your evaluation. Remember that Linux desktop adoption does not have to be all-or-nothing. The positive results can be translated into short-term opportunities for segments of your enterprise that are ready to move to a Linux desktop. The negative results can be translated into specific steps or requirements that need to be met before considering Linux desktop adoption for other user segments.

Evaluate application alternatives

An important part of any bid to move from Windows to Linux desktops is evaluating the application changes such a move requires. Begin by looking at the applications in the desktop environments that you have already identified as higher priorities for Linux adoption and move on from there. This is also an excellent exercise in simply understanding your enterprise's application portfolio and what may be possible in the near to long term.

We recommend using the following three questions as a baseline for identifying and evaluating application alternatives:

Is there a Linux version of the existing application?

As the Linux desktop continues to gain attention, more vendors are making their applications run on both Linux and Windows. For example, SAS*, SAP*, and Novell GroupWise(TM), all have clients for both Windows and Linux desktops. Check with your current application vendors to find out if they have a Linux version or are planning one anytime soon. (If they don't, and they're not currently planning on it, this is a great opportunity to encourage them to adopt a Linux strategy by letting them know that you'll be looking for an alternative Linux-friendly application.)

If not, is there a comparable Linux alternative?

As we've already indicated, it's important to view each of your current applications as a set of functionalities that allow users to accomplish their daily tasks. In many cases, fully comparable Linux applications are available; for example, replacing Internet Explorer* with Mozilla* Firefox* or Microsoft* Access* with MySQL* should provide one-to-one and for some activities better. functionality. And while is not exactly the same as Microsoft Office, for most users it has more than enough of the same functionality to make little difference.

If the current application will not run on Linux, first determine what functionality users really need and then look for a Linux-friendly application that provides it. And don't be surprised when you discover that some Linux-friendly applications provide functionality that is more specific to the needs of your users than their Windows-dependent counterparts.

If neither, is there another viable option?

When you can't identify a Linux version or a viable alternative, you'll want to investigate the feasibility of other options such as porting the application to Linux, serving it from a Web or terminal server, or using a utility such as CrossOver Office* by CodeWeavers*. Porting is particularly effective for internally developed applications and should be considered as a long-term solution. Web or terminal services can also be long-term solutions, particularly if your enterprise adopts a Linux thin-client strategy or finds it more cost effective to manage certain applications via the Web or terminal services.

In some cases, you'll discover that alternatives are more expensive than your current solution; watch licensing fees and development costs carefully. In this case, you will need to determine whether it's better to wait, to retain your current system, to support two systems, or make to an across-the-board move to matter the cost and inconvenience.

Evaluate Linux desktop distributions

Evaluating Linux desktop distributions is part of deciding whether Linux can meet the needs of your users and whether a Linux desktop is a worthwhile pursuit for your enterprise. Your analysis should include an assessment of which distribution to use and what criteria to use in your evaluation. Fortunately, there are many good options, with new Linux desktops being announced almost daily. Since this paper is being provided by Novell, we could end the discussion by recommending one of our own desktop solutions, particularly the Novell Linux Desktop released this fall (2004). Instead, let's look at some of the criteria for evaluating distributions so you can make your own decision.

You could decide to pick one of the more popular distributions-the ones with the greatest market share. You could choose one of the "free" (or nearly free) distributions. Or you could choose the distribution that offers the most comprehensive training and support systems. Whichever distribution you investigate, you'll have to wade through intangibles like perceived value, flashy marketing programs, and other subjective measures. It's easy to get caught up in a wave of hype that clouds the real value. Instead, base your decision on the features and functionality that are necessary to support your business objectives.

Look for specific measurable factors such as Interoperability; does the distribution support the applications you must run? How many applications run on this distribution? A limited number? Are the applications included with the distribution those that you really need?

Look for specific measurable factors such as Interoperability; does the distribution support the applications you must run? How many applications run on this distribution? A limited number? Are the applications included with the distribution those that you really need? Does the organization behind the distribution have strong industry backing? Are hardware and software vendors recommending the distribution for use with their products? Does the distribution offer a variety of windowing environments such as KDE* or GNOME?* What about integration with your workload environment and identity infrastructure?

Evaluate Support for the product. Many distributions are available that generally perform very well, but when you do experience a problem, is there anyone who is responsible for resolving it? Many distributions depend on the Linux/open source community for problem resolution and even though the community's response rate is typically very good, are you willing to bet your business on this method? Will the support organization be able to provide the right level of affordable support.

And if you find a bug, how quickly can you get a patch? Productivity can be significantly impacted when the patch process is difficult and highly manual. Look for a systematic method for patch distribution, one that provides you with discreet control over when, how, and where the patches are applied. This mechanism should not be limited to supporting only the distribution but desktop applications as well.

Consider the Ease of Installation of the distribution, specifically with regard to the default configuration. Do GUI and automated tools make installation easy and automatic or will you need power-user skills to get the job done via the command line? One of the most notable installation concerns is the strength (or lack) of security. Different distributions ship with security settings in various states, some in very insecure states. This often requires a series of modifications that must be handled manually as soon as the basic installation is complete. Anytime manual intervention is required, the stage is set for mistakes to occur and they usually do.

Every operating system requires some level of software Management. it's important to understand what that level is for applications, services, and the end users running them and to select a distribution that provides the level you need. it's equally important to look at how Linux can be managed in context with your existing environment.

Evaluate whether the distribution helps you meet the challenge of managing your Linux resources, especially update and patch management. To be effective, a Linux resource management solution should support the software update process typically employed in the Linux environment.

This process consists of four major phases:

An additional consideration should be the vendor's activity in the Open source community. Are they a significant contributor? How active are they in forums and events? Do they continually help raise the bar in terms of open source innovation?

If the activity is high, but the nature of the contribution seems proprietary, the contributions may do more to "fork" Linux rather than embrace open source. Innovation, on the other hand, can sometimes be confused with proprietary code but ultimately proves beneficial to the community. Weigh this aspect carefully.

And finally, make sure you include application Certification on a specific Linux distribution as one of your evaluation criteria. In many cases, you'll discover that Linux distributions have been certified by both the application vendor and the distribution vendor for use with specific versions of the distribution. In other cases, distribution vendors simply indicate that their own testing indicates that an application is .ready-to-use. with a particular version. This can be a significant distinction; the best solution, obviously, is testing the application in your own Linux test environment.

Create a road map for Linux desktop adoption

Technically, the time to create a road map is after you have decided to move forward with a project; but in evaluations like this, a road map can also be a valuable tool for weighing what it might take to accomplish such an initiative.

Your road map may be as short- or as long-term as you deem appropriate; however, a road map covering three or more years may have stronger strategic significance. Creating a road map will allow you to consider your enterprise's business and technical strategies and examine opportunities for the Linux desktop to benefit both or even help align the two. If, in your evaluation, you find that some environments don't fit with the Linux desktop right now, a long term road map provides an opportunity to examine the potential for adoption in the future, despite current limitations. Just as Linux desktop adoption doesn't have to be all-or-nothing, it also doesn't have to be now-or-never.

We recommend including these aspects in your road map:

Make sure your road map is as realistic as possible, even if you don't yet have solid plans for adopting the Linux desktop. The more realistic it is, the more useful it will be in determining whether the identified opportunities for adoption are viable. Road maps are also useful in helping appropriate decision makers understand all the implications so they can make an informed decision.

Putting a plan together

Developing a plan for moving forward is essential, whether you complete your own assessment and migration or use third-party professional services to help you facilitate all or part of your project. Either way, you'll want to plan carefully so that such a transition addresses not only your immediate goals but your long-term strategy as well.

If the recommendations in this paper and your initial investigation suggest that you need experienced help in planning and completing your migration, many consulting organizations, including some hardware vendors, offer experienced and professional help. Compare the services available and choose the one that's right for you.

Novell Professional Services can also provide the consulting help you need. Our engagements span the migration spectrum: from Strategy and Discovery to Requirements Assessment, Planning and Design to Implementation. Our offerings help you assess both current and future strategies and help you discover your readiness for moving to a Linux desktop. We provide help in deciding how to best approach a migration and assist you in planning and designing your entire project. And, finally, our consultants and partners stand ready to help you implement your migration plans. For a high-level summary of Novell assessment criteria and for additional information about Novell desktop migration services, visit us at


A few months ago, an article in Computerworld discussed the advisability of replacing ubiquitous Windows. desktop applications with Linux/Open Source. The writer concluded that "Microsoft owns the desktop, and the open-source folks might as well give up the silly game." We beg to differ (and so do several other Computerworld readers who wrote to object to this assumption). We at Novell have each personally either made the switch or are in the process. Though not without some pain, we've proved that the Linux desktop is enterprise-ready, but there's another, more important, question: "Is your Enterprise ready for Linux"? The discussion items in this paper should help you make this decision.

Open Source Apps Losing Desktop,. by Mark Hall, Computerworld, June 2004.

Because you will be changing the way your entire organization does business, being realistic and conservative is the order of the day. But in our opinion, it's an exciting–and productive–adventure. You don't get an opportunity to change the way the world works very often. Join us–and the growing number of major corporations and governments across continents–in proving that Linux on the desktop is much, much more than a "silly game".

Additional Reading

The following articles will give you some additional perspective on the issues you'll need to evaluate in determining whether a move to Linux on the desktop is right for your organization. URLs are included with these brief abstracts so that you can read the entire article.

Contact Us

For additional information about Novell, please contact your local Novell Authorized Reseller or system house or call, fax, or e-mail Novell at
1-888-321-4272 (U.S. and Canada)
1-801-861-7000 (worldwide)
Fax: 1-801-861-8473
Copyright Novell, Inc., 2004. All rights reserved 14

Disclaimer Novell, Inc. makes no representations or warranties with respect to the contents or use of this document, and specifically disclaims any express or implied warranties of merchantability or fitness for any particular purpose.
Trademarks Novell, the Novell logo, and GroupWise are registered trademarks of Novell, Inc. in the United States and other countries. SUSE is a registered trademark of SUSE LINUX AG, a Novell business. Linux is a registered trademark of Linus Torvalds. Microsoft, Windows, Microsoft Access, Microsoft Internet Explorer, and Microsoft Office are registered trademarks of Microsoft Corporation in the United States, other countries, or both. KDE is a registered trademark of KDE e.V. GNOME is a trademark of the GNOME Foundation. Mozilla Firefox is a trademark of Netscape Communications Corporation. SAP is registered trademark of SAP AG in Germany and in several other countries. SAS is a registered trademark of SAS Institute, Inc. CrossOver Office and CodeWeavers are trademarks of CodeWeavers. MySQL is a trademark of MySQL AB. All third-party trademarks are the property of their respective owners.
Prepared By Novell Solution Creation and Marketing - Linux Team
Title Approaching the Linux Desktop - A Novell Linux Migration Study
Date November 2004
John Beuchert Global Solutions Director
Erica Royer Solution Development Specialist
Norm Rupp Project Manager
Paul Rice Global Solutions Manager
Kurt Brust Global Solutions Manager
Doug Clower Global Solutions Manager
Robert Whitehead Novell Engineering
Joyce Whiting Solution Development Specialist
Philip K. Clark Global Solutions Specialist

Copyright 2004