Professional Documents
Culture Documents
Rsaccons MBA Thesis
Rsaccons MBA Thesis
Rsaccons MBA Thesis
A Thesis
in Partial Fulfillment
by
Roberto Saccon
August 2003
This thesis studies the possibilities and limits of open source software at the enterprise. The thesis
analyzes the history of the open source movement, describes the open source community and
collaboration model, analyzes the open source development process, describes business models
based on open source software, analyzes possible cost savings and presents case studies of popular
open source projects.
Recommendations are presented, how companies and organizations might benefit from open source
software and in which cases it should be avoided, because the hidden costs will not pay off the
license costs savings.
Along the process of writing the thesis, more than 20 IT managers from companies and
organizations all over Switzerland had been interviewed. The evaluation of their experiences
contributed significantly to the conclusions of the thesis.
Table of Contents
1 Introduction.................................................................................................................................... 11
2 What is Open Source Software.......................................................................................................13
2.1 Myths and realities.................................................................................................................. 14
2.2 Free and Open Source Software FOSS................................................................................... 15
2.3 Open Source Definition.......................................................................................................... 15
2.4 OSS Licenses.......................................................................................................................... 18
2.4.1 General Public License (GPL)........................................................................................ 18
2.4.2 Lesser General Public License LGPL............................................................................. 18
2.4.3 Mozilla Public License MPL.......................................................................................... 19
2.4.4 BSD License....................................................................................................................19
2.4.5 Apache Software License................................................................................................19
2.4.6 Public Domain.................................................................................................................19
2.4.7 OSS License Summary....................................................................................................20
2.4.8 Non-Open Source licenses.............................................................................................. 20
2.5 OSS and intellectual property................................................................................................. 21
2.6 The Open Source Community................................................................................................ 23
2.6.1 Development Model........................................................................................................25
2.6.2 The Source Code Repository...........................................................................................26
2.6.3 Software Quality............................................................................................................. 26
2.7 OSS and the Internet............................................................................................................... 27
2.7.1 The Internet as driving force behind OSS....................................................................... 27
2.8 OSS and Digital Media........................................................................................................... 27
3 History of OSS...............................................................................................................................29
3.1 Linux Operating System......................................................................................................... 29
3.2 Apache Webserver.................................................................................................................. 30
3.3 Netscape and the Mozilla Web Browser.................................................................................30
3.4 The Halloween documents......................................................................................................31
3.5 OpenOffice..............................................................................................................................32
3.6 JBoss Enterprise Application Server...................................................................................... 32
3.7 Eclipse Development Platform............................................................................................... 33
3.8 City of Munich decides for Linux...........................................................................................33
3.9 Microsoft's call-to-arms.......................................................................................................... 34
3.10 The SCO / IBM lawsuit........................................................................................................ 34
4 Engaging with the Open Source Community................................................................................. 36
4.1 Commercial OSS and Services............................................................................................... 36
4.2 Deploying OSS directly from online repositories..................................................................36
i
4.3 Active Participation................................................................................................................ 37
4.4 Active Participation with leading function............................................................................. 38
4.5 OSS without community.........................................................................................................39
5 Global trends in OSS adaption....................................................................................................... 40
5.1 Europe..................................................................................................................................... 40
5.1.1 Germany.......................................................................................................................... 40
5.1.2 Spain................................................................................................................................41
5.2 Asia......................................................................................................................................... 41
6 OSS from economic point of view................................................................................................. 43
6.1 Total Cost Of Ownership TCO............................................................................................... 43
6.1.0.1 Possibilities and limits of TCO calculations........................................................... 43
6.1.1 Components of a TCO calculation..................................................................................44
6.1.1.1 Acquisition costs..................................................................................................... 44
6.1.1.2 Quality components.................................................................................................44
6.1.1.3 Administration.........................................................................................................44
6.1.1.4 User involvement ................................................................................................... 44
6.1.2 TCO - Summary.............................................................................................................. 46
6.2 Return on Investment (ROI) and Business Value................................................................... 46
6.2.1 The decision.................................................................................................................... 49
6.3 Open source effect on value of software.................................................................................50
6.4 Devaluation of software..........................................................................................................52
6.4.1 Devaluation as a competitive advantage......................................................................... 52
6.4.1.1 Aggressive patent use.............................................................................................. 53
6.4.1.2 Monolithic Software ...............................................................................................53
6.4.1.3 Competition............................................................................................................. 54
6.4.2 The OSS based business model for software vendors.................................................... 55
6.4.2.1 OSS and the Apple Mac OS X................................................................................ 56
6.4.3 How does Microsoft deal with devaluation of its software.............................................56
7 OSS and usability........................................................................................................................... 59
7.1 Approaches for better OSS usability.......................................................................................60
8 OSS examples and case studies ..................................................................................................... 62
8.1 Server Operating System: Linux ..........................................................................................62
8.1.1 SWOT analysis................................................................................................................62
8.1.1.1 Strength................................................................................................................... 62
8.1.1.2 Weaknesses:............................................................................................................ 63
8.1.1.3 Opportunities........................................................................................................... 63
8.1.1.4 Threats..................................................................................................................... 64
8.1.2 Linux Server summary and conclusion........................................................................... 64
8.2 J2EE Application Server: JBoss ............................................................................................ 65
8.2.1 SWOT analysis................................................................................................................65
8.2.1.1 Strengths.................................................................................................................. 65
8.2.1.2 Weaknesses ............................................................................................................ 66
8.2.1.3 Opportunities........................................................................................................... 67
ii
8.2.1.4 Threats..................................................................................................................... 67
8.2.2 JBoss summary and conclusion...................................................................................... 67
8.3 Office Suites: OpenOffice ......................................................................................................68
8.3.1 SWOT analysis:.............................................................................................................. 68
8.3.1.1 Strengths.................................................................................................................. 68
8.3.1.2 Weaknesses ............................................................................................................ 68
8.3.1.3 Opportunities........................................................................................................... 69
8.3.1.4 Threats..................................................................................................................... 69
8.3.2 OpenOffice summary...................................................................................................... 69
8.4 Web browser: Mozilla ........................................................................................................... 70
8.4.1 SWOT analysis:.............................................................................................................. 70
8.4.1.1 Strengths.................................................................................................................. 70
8.4.1.2 Weaknesses ............................................................................................................ 70
8.4.1.3 Opportunities........................................................................................................... 71
8.4.1.4 Threats..................................................................................................................... 71
8.4.2 Mozilla summary and conclusion................................................................................... 71
8.5 ERP/CRM Software: Compiere .............................................................................................72
8.5.1 SWOT analysis:.............................................................................................................. 72
8.5.1.1 Strengths.................................................................................................................. 72
8.5.1.2 Weaknesses ............................................................................................................ 72
8.5.1.3 Opportunities........................................................................................................... 73
8.5.1.4 Threats..................................................................................................................... 73
8.5.2 Compiere summary and conclusion................................................................................ 73
8.6 Content Management Software (CMS): ez Publish ...............................................................74
8.6.1 SWOT analysis:.............................................................................................................. 74
8.6.1.1 Strengths.................................................................................................................. 74
8.6.1.2 Weaknesses ............................................................................................................ 74
8.6.1.3 Opportunities........................................................................................................... 74
8.6.1.4 Threats..................................................................................................................... 74
8.6.2 Ez Publish summary and conclusion...............................................................................75
9 OSS survey in Switzerland ............................................................................................................ 76
9.1 Expert interviews.................................................................................................................... 76
9.2 General observations...............................................................................................................81
9.2.1 Small, privately owned companies .................................................................................81
9.2.2 Larger enterprises ........................................................................................................... 81
9.2.3 Internet service provider( ISP)........................................................................................ 82
9.2.4 Governmental organizations........................................................................................... 82
9.2.5 Educational institutions...................................................................................................83
10 Conclusions.................................................................................................................................. 84
10.1 Summary of advantages of OSS........................................................................................... 84
10.2 Summary of disadvantages of OSS.......................................................................................84
10.3 Possibilities for OSS............................................................................................................. 84
10.3.1 Governmental organizations......................................................................................... 85
10.3.2 Educational Institutions.................................................................................................85
10.3.3 ISP................................................................................................................................. 85
iii
10.3.4 Software Vendors..........................................................................................................85
10.4 Limits for OSS...................................................................................................................... 85
10.4.1 Linux for the desktop.................................................................................................... 85
10.4.2 OSS for niches.............................................................................................................. 86
10.5 First Steps for migrating to OSS...........................................................................................87
11 Bibliography................................................................................................................................. 89
11.1 Working papers.....................................................................................................................89
11.2 White papers......................................................................................................................... 90
11.3 Internet resources: ................................................................................................................ 90
11.3.1 Expert Interviews:......................................................................................................... 91
iv
List of Figures
v
List of Tables
vi
Abbreviations
FLOSS Free / Libre / Open Source Software
FS Free Software
IE Internet Explorer
vii
Preface
I became interested in open source software in summer 2000. As software engineer, I had to
customize applications to fulfill the requirements of customers. I discovered that Linux and other
open source software offered possibilities for customization far beyond what comparable
proprietary software offered. During the next years I used open source software personally and
professionally wherever it made sense. And I learned that often it did not make sense, neither form
economic nor from usability point of view.
This thesis was written in summer 2003, as word processor I used the open source software
OpenOffice1.
Acknowledgments
I would like to thank my supervisor, Dr. Helmut Steigele, for all the help I received over the
research. Had not been for his guidance, this paper would not have materialized.
I also wish to thank my wife, Rosemary, for inspiring me, motivating me and for pushing me to get
work done in time.
And finally a very special thank to all those IT professionals who shared their experience and
valuable insight with me by accepting to be interviewed.
1 For further information about OpenOffice visit http://www.openoffice.org, accessed: Aug. 28, 2003
viii
1 Introduction
Many enterprises are struggling to find a solution that will help them control costs and remain
competitive. One approach many IT managers are investigating is the inclusion of open source
technologies as a key part of their overall IT strategy to reduce licensing fees and operating
expenses.
During the last decades, open source software was more of academic than economic importance.
But things have changed. These days everyone is talking about Linux and open source software.
Business analyst Chistopher Koch believes2, that CIOs who will not develop an open source strategy
in 2003, will be paying too much for IT in 2004. Linux got adopted by big companies like IBM and
Oracle. Many enterprises and governmental organizations are cutting costs, gaining flexibility and
discovering powerful new sources of business value with open source software. On the other hand,
IT costs often are dominated by service and consulting fees and not by license fees.
The objective of this paper is to identify the possibilities and limits of OSS at the enterprise and to
analyze how companies and organizations can benefit from OSS.
• Beside of license costs savings, what are the other benefits of OSS ?
• In what cases OSS solutions are more expensive than proprietary solutions ?
• What are the advantages and disadvantages of software solutions which consists of proprietary
and open source components ?
• Is the common belief that volunteers are developing OSS for free still valid in the context of
professional OSS at the enterprise level.
• Will there still be a market for OSS, if proprietary software is getting better and cheaper ?
11
In examining these questions, this paper discusses existing relevant research, presents examples and
case studies of popular open source projects. Theories about business models and cost analysis will
be applied to the concept of OSS.
Beside of the the theoretical analysis, IT managers of more than 20 companies and organizations in
Switzerland have been interviewed. The summary and the statistical results of the experiences of
these IT managers are an integral part of this paper.
Finally recommendations are presented, how companies and organizations can benefit from OSS or
when OSS should be avoided. An example of an action plan is presented, with all the initial steps,
which should be followed when deciding for the OSS path.
12
2 What is Open Source Software
Before exploring the characteristics of OSS, an attempt is made to locate software development as
such in the larger picture of the IT industry and its history.
The evolution of the computer industry has been driven by the emergence of standardized platforms
which allow modular substitution of complementary assets such as hardware and software3. In the
early years of the IT industry, software was provided by the manufacturer of the hardware, and it
was only running on the hardware of its manufacturer. A big step towards standardization was the
emerge of hardware independent operating systems – Unix and Windows – which shifted platform
control form hardware manufacturers to operating system vendors. While Windows has always been
under full control of a single company – Microsoft – the Unix operating system had many variants,
some of them developed and maintained by academic institutions, other variants were under control
of commercial companies. The Unix operating system also inspired to a more radical shift of
platform and vendor independence, the open source movement which represents a new software
development methodology and licensing model.
The are many differences between proprietary closed source software and OSS. The most
significant ones are the availability of the source code and the development model of OSS, which is
community based. Today OSS software is associated with the following characteristics, some of
which are usually not found in commercial closed source software:
13
2.1 Myths and realities
There exists a common belief that open source software is developed by alternative, beard
individuals who represent the classical computer freak or “hacker”. That might have been true in a
general sense when the first open source projects emerged at the horizon. Today open source
software became professional and the businesses built around open source became conventional
ones. Many important open source projects are sponsored by commercial companies like IBM4 and
SUN5. There are successful OSS projects which are under the control of commercial companies
grown up with an open source project. An example which is analyzed later in this paper is the open
source project JBoss, controlled by its commercial entity JBoss Group6.
Another myth is that OSS must be necessarily for free. The availability of the source code does not
imply that. There are many commercial software products which ship with source code, but for
commercial use of the software, a license fee must be paid. One example is Resin, a Java based
software for serving dynamic generated Web pages. Caucho7 , the company behind Resin lets
everybody download source code and binaries8 of their software product. The license however
states that one may use the software to develop, evaluate, and demonstrate applications, but it is not
allowed to use it if one is paid to use the software or is paying anyone else to use the software.
There exists a common belief that open source is the same as Public Domain. This happens simply
because the idea of open source is confusing to many people, and they describe these programs as
public-domain because that's the closest concept that they understand. The programs, however, are
clearly copyrighted and covered by a license, just a license that gives people more rights than they
are used to.
Many people claim that OSS usually is not user friendly. This seems to be true, with some
exceptions. The explanation can be found in the nature of the OOS developers. They are experts in
providing functionality, but they give little priority to software ergonomics. User experience often is
related with the graphical appearance and quality of documentation of the software. And OSS often
lacks in both of the areas, because OSS projects which do not have a straight commercial
background usually do not have graphical artists and technical writers in the team of volunteers and
budget does not allow to outsource these tasks9. However there are exceptions, for example Mozilla
4 IBM donated the eclipse development platform to the OSS community. See chapter 3.7
5 SUN donated StartOffice to the OSS community, see chapter 3.5
6 Fleury, M., Why I love Professional Open Source, JBoss Group LLC, 2003
7 For further information about Caucho and its license strategy visit http://www.caucho.com
8 Binaries: executable code, as result of compiling the source code during the software development process
9 Nichols, D. M., Twidale, M.B., Usability and Open Source Software, University of Waikato, 2002, P.5
14
Firebird10, a streamlined open source web browser which is comparable to commercial browsers like
Internet Explorer from Microsoft in regard to usability. Most of the successful and widely used OSS
is targeted to the server side, where the aspect of usability is less relevant compared to stability and
performance.
1. Free redistribution
“The license shall not restrict any party from selling or giving away the software as a component
10 For further information about Mozilla Firebird web browser visit http://www.mozilla.org, accessed: Aug. 28,
2003
11 Fink, M., The Business and Economics of Linux and Open Source, P.36
12 The official website of the Open Source Initiative is available at http://www.opensource.org, accessed: Aug. 28,
2003
13 Sandred, J., Managing Open Source Projects, P.47
14 Open Source Definition, Version 1.9. Available at http://www.opensource.org/docs/definition.php, accessed:
Aug. 28, 2003
15
of an aggregate software distribution containing programs from several different sources. The
license shall not require a royalty or other fee for such sale”
It is common for providers of OSS to charge a fee for the media, the manuals and support
offerings. This provision usually results in free or low-cost software.
2. Source Code
“The program must include source code, and must allow distribution in source code as well as
compiled form. Where some form of a product is not distributed with source code, there must be
a well-publicized means of obtaining the source code for no more than a reasonable
reproduction cost–preferably, downloading via the Internet without charge. The source code
must be the preferred form in which a programmer would modify the program. Deliberately
obfuscated source code is not allowed. Intermediate forms such as the output of a preprocessor
or translator are not allowed.”
Since the purpose is to make evolution of OSS easy, modification of the source code must be
made easy. Accessing obfuscated15 source code makes modification of OSS more difficult.
3. Derived Works
“The license must allow modifications and derived works, and must allow them to be distributed
under the same terms as the license of the original software.”
For rapid evolution of OSS, the ability to read source code is not enough. Developers need to be
able to redistribute modifications.
“The license may restrict source-code from being distributed in modified form only if the license
allows the distribution of "patch files" with the source code for the purpose of modifying the
program at build time. The license must explicitly permit distribution of software built from
modified source code. The license may require derived works to carry a different name or
version number from the original software. “
In the normal process of development, within the open source community, the main source code
is usually controlled by a “maintainer”. There may be cases where the maintainers rejects
changes to the official source tree16. With patches plus the base source code, there exists a way to
make available unofficial modifications.
15 The only purpose of obfuscating source code is to make the source more difficult to read.
16 Software projects usually have their different versions of source coded organized in an hierarchical tree.
16
5. No discrimination against person or groups
“The license must not discriminate against any person or group of persons.”
This provision prevents an open source licensor form controlling or limiting the way the software
is used.
“The license must not restrict anyone from making use of the program in a specific field of
endeavor. For example, it may not restrict the program from being used in a business, or from
being used for genetic research.”
The clause is intended to prohibit license traps that prevent open source software form being
used commercially.
7. Distribution of license
“The rights attached to the program must apply to all to whom the program is redistributed
without the need for execution of an additional license by those parties.”
The intention of this clause is to forbid closing up software by indirect means such as requiring a
Nondisclosure Agreement NDA.
“The rights attached to the program must not depend on the program's being part of a
particular software distribution. If the program is extracted from that distribution and used or
distributed within the terms of the program's license, all parties to whom the program is
redistributed should have the same rights as those that are granted in conjunction with the
original software distribution.”
“The license must not place restrictions on other software that is distributed along with the
licensed software. For example, the license must not insist that all other programs distributed on
the same medium must be open-source software.”
This clause encourages companies to distribute OSS and proprietary software an the same
distribution medium.
17
10.The license must be technology neutral
“No provision of the license may be predicated on any individual technology or style of
interface.”
The clause is intended to prohibit OSS to rely on any specific distribution mechanism.
18
distributed. This is particularly interesting for libraries and frameworks. The LGPL license
facilitates common used libraries and frameworks to become de-facto standard. Many open source
projects have chosen the LPGL because it is a good compromise between protecting the copyright
owners and their work, but still allowing to use LGPL-ed software for commercial purpose. An
examples of a popular open source project using the LPGL license is the JBoss21 Java application
server.
19
removing the original author's name.
When using and modifying public domain software, it should be considered to apply ones own
copyright to the program and re-license it. The original version will still be in the public domain,
but the modified version will be under a license that others must heed if they use it or derive from it.
Most commercial companies which release source code, create there own license. Some companies
let approve these license by the OSI.
1. The copyright owner wants to give the unrestricted ability to the user to debug27 an application
which depends on a library. For example, Microsoft ships source code to most of its foundation
classes MFC, which are used to develop Windows applications.
2. The software developer or company does not own all the rights. In such cases, even if it would be
26 Fink, M., The Business and Economics of Linux and Open Source, P.50
27 Debugging an application, in this context, means to observe with the help of a software tools in real time each
step of a running application at source code level.
20
admirable to apply a real open source license, there are restrictions which prevent form doing so
because of contracts with other parties.
3. The software developer or company wants feedback from customers or the development
community without giving up intellectual property rights. If a company wants to let their code
peer review by the developer community, but the business model of the company is dependent on
the the software license revenue, than this type of license may be appropriate. The Microsoft
Shared Source License28 falls into this category. This license is most often used for software
components which are running on non standard devices such as operating systems for handheld
computers. The Shared Source License provides no redistribution rights and Microsoft gains all
the rights for bugfixes or enhancements a developers makes.
Briefly: the software is delivered with the source and the user can do with that software what
ever he wants, including redistribute it and selling it. The only restrictions, which depend on the
open source license, might be the requirement to provide the source code when redistributing or
to give back modifications of the software to the license owner.
This kind of closed source software provides a functionality which can be imitated by other
software. With other words, this kind of software is open to re-engineering. A typical example
are file formats. Some companies have released the specification of proprietary file formats to
facilitate the interoperability with other software. An example is Macromedia, which has released
the specification of its Flash vector file format30. An example of a protocol which has been kept
secret, but third party developers managed to re-engineer, is the Microsoft network filesystem
protocol SMB. The result of this re-engineering effort by OSS developers is Samba31, a popular
21
Linux project. Samba allows sharing files between a Linux and Windows computer over a local
network. Microsoft has not taken any legal actions against Samba so far, however there are parts
in the Microsoft network filesystem protocol which are protected by patents32, but which are
currently not used by the Samba project.
The least open kind of software is entitled to patents. The patent owner can prevent anyone else
from using it and the idea and/or method behind it - unless they get permission, which usually
means paying a royalty fee.
Unfortunately most patents go beyond that - they can put controls on the process that software
allows to happen. A example is Amazon.com and its One-Click patent33. Amazon thought it
would be a good idea to allow its customers to click just one button to buy a book, rather than
have to go through various screens confirming their credit card, address, delivery etc.
It was a very good idea but not a huge leap in imagination. The tough part was writing software
that provided the functionality. Amazon.com patented the one-click idea, for allowing them to
take legal actions against competitors from offering the same thing - even if the code was written
from scratch. Barnes & Noble was prevented from offering a one-click option until the two
settled years later out of court. Amazon still holds the patent.
Patents exist for a very good reason. Let's take pharmaceutical companies for example. It costs
billions to develop and test new medicines until a commercially feasible one is found. Without a
patent on this, competitors would be able to copy the drug and benefit from the other company's
research and development. It is easy to imagine that without this protection, there would be no
new drugs. One of the major arguments for software patents is the protection of investment. This
encourages companies to spend money developing new ideas. This argument is simple and easily
explained and lawyers instinctively understand it.
The counter-arguments are more complex. The strongest one is that software works by building
on top of other software. Patenting aspects of software breaks down the process by which new
software is created.
Another very strong argument is the huge success built on top of common, patent and royalty-free
standards. The Internet is so successful because there weren't constraints or patents on it. Due to
this, it has grown hugely, created new markets and so benefited everyone.
32 Howorth, R., Samba avoids patents costs, IT week, Issue 03-05-2002, available at
http://www.itweek.co.uk/News/1131485, accessed: Aug. 28, 2003
33 For further information about Amazon's one click patent visit
http://swpat.ffii.org/patente/wirkungen/1click/index.en.html, accessed: Aug. 28, 2003
22
2.6 The Open Source Community
OSS is not different than any other kind of Software, not technically at least34. The most important
issue in open source development is organizational35. In the broadest sense, an open source
community includes everyone who uses OSS. However, many users of OSS are passive consumers,
they do not actively contribute to the evolution of an OSS project by posting questions on mailing
lists, participating in online discussions, posting bug reports or developing and contributing source
code. For the purposes of this paper, the author refers to the active open source community when
using the term open source community. Each OSS project has its community. An individual open
source community is characterized as a loose association of people who have a common interest in
developing and using a particular piece of software. These people are typically geographically
dispersed, communicating via email and instant messaging and using various open source
development tools in order to collaborate on the software that they are jointly developing, testing,
using, and improving.
Eric S. Raymond has pointed out that the OSS development approach violates many of the well
established ideas in software engineering36. He introduced the concept of the “cathedral” and the
“bazaar”. He associates the traditional, corporate, closed source methods to the “cathedral” model
and the open source development process to the “bazaar” model.
Questions which can be found at most academic papers, that try to analyze the phenomena of OSS
are:
• Why do programmers write Open Source codes if no one pays them to do it?
• How do hundreds of developers located all over the world and without hierarchical structure
based on the ownership of the developed product manage to coordinate each other ?
Menzolas Tzouris articulates in his research about the motivation of people to contribute to open
source projects37: “When open source programmers volunteers their time into writing code for an
FS/OSS project, they freely give away the code they write for the good of the FS/OSS community; at
the same time, they benefit from of the contributions of others within this community by getting
really good software, and so the whole movements benefits and grows.”
34 Tzouris, M, Software Freedom, Open Software and Participant's Motivations, The London School of Economics
and Political Science, 2002. P. 9
35 Sandred, J., Managing Open Source Projects, P.73
36 Raymond, E.S. The Cathedral and the Bazaar, 1999, O'Reilly, Sebastopol, CA.
37 Tzouris, M, Software Freedom, Open Software and Participant's Motivations, The London School of Economics
and Political Science, 2002, P. 43
23
There is also a strong relationship between the academic world and the OSS movement. Many
successful OSS projects have an academic origin. Developing OSS is often part of scientific
research, and motivation to invent something new or improve something existing, is a driving force
for developing OSS. And the interchange of ideas without restrictions of intellectual property
protection for the benefit of all involved and not involved parties is common to academic research
and development of OSS.
Some people see writing code as an art. And the more independence an artist has, the better are his
results. Under the umbrella of OSS, code artists enjoy independence and they reach a large
audience, what is another motivating factor.
Those who contribute to an open source community typically are not compensated financially
within the community for their work. Members join the community because they have an
intellectual interest in the software or because they have an association with some other
organization that is interested in or actively using the software. However, many open source
community members are paid for their work directly by those enterprises that employ them, and
their contributions to the community are simply made in the course of their normal employment.
Intangible status and rewards within an open source community invariably come as a result of
contributing significantly to the software developed by the community.
Most successful open source projects display a clear hierarchical organization, but the roles within
the hierarchy are are not strictly assigned since the beginning38. The leadership deeply influences the
evolution of the project, but does not have to carry out the bulk of coordination effort, that would be
necessary within a comparable corporate closed source project. Members of the open source
community typically choose for themselves what piece of programming or other development tasks
they will work on, depending on their interests and skills.
Typically one or more senior members of the community will organize a prioritized list of things
that need to be done, but it is up to the individual members to choose their own self-assignments. It
is considered good form to work on tasks from the prioritized list; consequently the recognition is
greater for those who cooperate with the agreed-upon priorities. Most open source communities are
fairly cooperative, but members are free to contribute however they wish.
There is typically a group of core developers who write the most formidable constructs and
algorithms. A larger peripheral group of less committed developers often exists that contributes
interfaces, subroutines, device drivers, and the like. Some members may focus their efforts on
38 Bonnacorsi, A. Why Open Source can succeed, Sant' Anna School of Advanced Studies, Pisa, 2003
24
functional or performance testing, while others might write documentation. Perhaps the largest
group are the involved users who contribute bug reports and feature requests. Everyone involved is
free to make comments, argue for one approach over another, and generally discuss the direction
and technical details of the software under construction. Peer recognition, honest and direct
criticism, and the weight of one's own reputation among other members of the community combine
to make these open source communities self-regulating.
With OSS any teenager with a personal computer and an Internet connection can compete - on a
level playing field - with seasoned systems development professionals in academia and industry.
Another characteristic of open source development is vigorous criticism. It's well known that email
communications often lack the tact of personal interactions. Any open source developer who
submits a badly formatted and gratuitously complicated, spaghetti-code module or patch to the
maintainer of an open source project, gets publicly criticized within the community. Corporate
programming managers cannot get away with such enthusiastic correction of subordinates in
modern office environments with strictly enforced human resource policies. But open source
contributors are all volunteers, so they either work up to the standards of their peers, or drop out,
and someone more able takes their place. Open source communities are very capable of enforcing
25
high standards among their contributors, either quickly training or losing any poor performers.
Another piece of Open Source philosophy is characterized as "many eyes make all bugs shallow."41
The continual review process used by open source communities produces a "many eyes" effect of
massively parallel peer review that has been demonstrated to produce very high quality oversight of
the software development process and products. Constant, repetitive peer review, coupled with a
release schedule tied to objective software quality rather than marketing deadlines, consistently
results in open source software quality orders of magnitude higher than that of commercial releases
of similar software.
Open Source software is constructed primarily by those who want to use it, so they don't release it
until they're satisfied that it's usable by the majority of the user community involved, that is, when
it's done.
39 Open source code repository SourgeForge is available at http://www.sourgeforge.net, accessed: Aug. 28, 2003
40 Fink, M., The Business and Economics of Linux and Open Source, P.56
41 Raymond, E.S., The Cathedral and the Bazaar, O'Reilly, Sebastopol, CA, 1999
26
2.7 OSS and the Internet
The Internet as it has evolved over the years and as we know it toady, is tight together with OSS.
The following list shows that there would have been no Internet without the following examples ,
which opened the path and most of them still play a dominant role:
• Domain Name Server - The most visible element of the Internet are domain names. All this
domain names are mapped to IP addresses. This mapping called Domain Name Service (DNS) is
provided by nameservers. The most popular nameserver software is BIND42, an open source
software for Unix servers developed during the early years of the Internet at Berkeley University.
Any other nameservers software has just a marginal market share.
• E-mail Server - The most popular software for E-mail servers is Sendmail43. It has come to age,
it was first released more than 20 years ago. Eric Allman, one of the main developers has built a
for profit organization around Sendmail. Initially Sendmail had many bug and security issues, but
it has matured and today Sendmail still processes about three quarters of all Internet mail.
• Web Server - The most popular web server with a market share of 64% is Apache44.
27
legal download.
Beside of the online availability, which OSS and digital entertainment media have in common, there
are fundamental differences between these two types of digital information. First there are licensing
differences. OSS comes with a license, which exactly states what one is allowed to do with the
downloaded product. OSS is usually much more complex than digital content, and has an extended
life cycle and requires often large amount of interaction with the end user. A major difference
between OSS and other digital information available on the Internet is the life cycle of OSS, which
usually is tied to the Internet from its very first idea appearing on mailing-lists, to the actual
development of the software in an collaborative approach, to the distribution. OSS development
process might represent the process of global collaboration between individuals, something which
can be applied to other activities than software development.
28
3 History of OSS
In the early days of computing, software was only available together with hardware. Otherwise the
hardware would simply not operate without it. In the 1980s commercial Unix Systems47 dominated
the computer industry. It was at that time when people started to write free available and
distributable UNIX imitations48. One of them was Minix, developed by Andrew Tannenbaum.
Another popular Unix imitation was BSD developed at the University of Berkeley in California. I
later formed the basis of Solaris, Sun Microsystems Unix operating system.
In the 1980s Richard Stallman founded the Free Software Foundation49 which served as vehicle for
making real his vision of an open source operating system. At that time, Stallman developed an
editor called EMACS (which is still the favorite editor for many software developers today).
Stallman gave away the editor including the source code for free, but charged $150 for actually
making the copies and for providing support. He asked to sent back to him all modifications and
improvements. In this way he developed the model of the GPL license, which is described in detail
in the further on. More and open source tools which run on top of any Unix implementation and
which are paving the road toward an open source Unix like operating system were developed by
Stallman and the community. These tools are known as the GNU tools and every Linux Distribution
today contains a significant amount of this GNU tool. The term GNU itself comes from “GNU is
Not Unix”
29
since april, and is starting to get ready. I'd like any feedback on
things people like/dislike in minix, as my OS resembles it somewhat
(same physical layout of the file-system (due to practical reasons)
among other things).
I've currently ported bash(1.08) and gcc(1.40), and things seem to work.
This implies that I'll get something practical within a few months, and
I'd like to know what features most people would want. Any suggestions
are welcome, but I won't promise I'll implement them :-)
Linus (torvalds@kruuna.helsinki.fi)
PS. Yes - it's free of any minix code, and it has a multi-threaded fs.
It is NOT protable (uses 386 task switching etc), and it probably never
will support anything other than AT-harddisks, as that's all I have :-(.
Torvalds used and integrated the GNU Tools into Linux. While Linus Torvalds was the Linux
project administrator, many hundreds people form all over the world provided feedback, help, and
support. The Internet made for the first time in history such a joint effort of an open source project
possible. 1994 the version 1.0 of the Linux kernel was released. Over 120000 programmers have
been contributing to Linux about 2 billion dollars worth of labor50.
50 Kenwood, C. A., A Business Case Study of Open Source Software, Mitre, 2001
51 The Apache web server has a worlwide market share of 64% according to netcraft statistics:
http://news.netcraft.com/archives/web_server_survey.html, accessed Aug. 28, 2003
52 Rosenberg, Donald K., Open Source: The Unauthorized White Papers, 2000, P.32
53 For further information about Apache open source projects visit http://www.apache.org, accessed Aug. 28, 2003
54 Rosenberg, Donald K., Open Source: The Unauthorized White Papers, 2000, P.28
30
The codename of the project was Mozilla (a monster to destroy Mosaic). In the meantime Microsoft
also had done its homework. They licensed NCSA Mosaic Browser and studied its functionality and
eventually produced their own browser which they delivered together with their operating system.
The browser war between Netscape and Microsoft started and Microsoft won the battle. Netscape
market share dropped to 20%. After having won the browser war, Microsoft stopped improving
their browser technology. After being bought by AOL, Netscape also stopped to innovate their
browser technology. However they released the source code of their browser, and the project was
called Mozilla again (this time Microsoft was the enemy). Today the Mozilla browser offers
advanced features like tabbed browsing which are available for the Internet Explorer only as plugins
from third party vendors55. Only the future will show whether there will be a round two of the
browser war.
During the next years, Raymond used every official or unofficial Microsoft document, which was
attacking Linux or defending Microsoft vision of enhanced security with proprietary software, to
released to the press, after having added his annotations and a “Halloween” number to the
documents57.
The reason why Raymond released only annotated version of the Microsoft documents, was to
defense himself against possible a copyright-violation attacks from Microsoft. Making un-annotated
copies available would have placed Raymond at significant legal risk.
Up to August 2003, totally nine such “Halloween documents” have been released by Eric S.
Raymond. The documents reveal what Microsoft has been thinking in the past about Linux, how
they have been ignoring it and the FUD58 they spread about OSS in general.
55 For more information about the Mozilla web browser visit http://mozilla.org, accessed Aug. 28, 2003
56 Rosenberg, Donald K., Open Source: The Unauthorized White Papers, 2000, P.79
57 All Halloween documents are available at http://opensource.org/halloween/halloween1.php, accessed Aug. 28,
2003
58 FUD: Fear, Uncertainty and Doubt
31
3.5 OpenOffice
In the summer of 2000, SUN released the source code of their office suite StarOffice. Two years
later the project counts many hundred volunteers which contributed to the developments,
OpenOffice 1.0 was released and ten millions of users downloaded the software.
For long time there did not exist an open source alternative MS Office. This changed when Open
Office 1.0 was released59. Historically, Open Office is based on StarOffice, which has been written
by StarDivision, a company founded in Germany in the mid-1980s. It was acquired by Sun
Microsystems during the summer of 1999. One year later Sun donated the source code of StarOffice
to the community and this was the birth OpenOffice. 2002 the project had already hundreds of
volunteers. At that time, OpenOffice 1.0 was released and ten millions of users downloaded the
software. SUN continued to develop and distribute StarOffice, which is based on OpenOffice, but
contains some additional features.
The JBoss history starts in the year 2000, when Marc Fleury, at that time a SUN consultant, started
to work on his own implementation of an application server64. The project attracted some of the best
Java developers which contributed their ideas and code. The JBoss Group was founded, a company
which provides training, consulting and support services for the JBoss Application server. Today
59 For further information about the OpenOffice software visit http://www.openoffice.org, accessed: Aug. 28, 2003
60 Middleware is software which is located between end user applications and the operating system or the backend
system.
61 J2EE stands for Java 2 Enterprise Edition, fro further information visit http://java.sun/j2ee, accessed: Aug. 28,
2003
62 For further information about JBoss visit http://www.jboss.org, accessed: Aug. 28, 2003
63 LaMonica, M., Java servers feel the open.source heat, Feb. 2003, Internet:
http://news.com.com/2100-1001-984476.html, accessed: Aug. 28, 2003
64 Fleury, M., Why I love Professional Open Source, JBoss Group LLC, 2003
32
JBoss is technically equal if not better than its proprietary competitors. The software has been
downloaded millions of time.
The story started with Microsoft's initial $36.6 million proposal for upgrading the software. Suse
and IBM, which formed a powerful Linux partnership, offered a $35.7 million proposal. And
because Microsoft did not want to loose the deal to Linux, in spring 2003 Steve Ballmer visited
Mayor Christian Ude to assure him Microsoft would do what it takes to keep the city's business.
And Microsoft did a lot of efforts for not loosing an important key customer:
• Microsoft lowered its price down to $23.7 million — an overall 35% price cut.
• Microsoft agreed to let Munich go as long as six years, instead of the more normal three or four,
without another expensive upgrade, a concession that runs against its bread-and-butter software
upgrade strategy.
• Microsoft offered to let the city buy only Microsoft Word for some PCs and strip off other
applications. Such unbundling cuts against Microsoft's practice of selling PCs loaded with
software.
• Microsoft offered millions of dollars worth of training and support services free.
65 For further information about the OpenOffice software visit http://www.eclipse.org, accessed: Aug. 28, 2003
66 Acohido, B., Linux took on Microsoft, and won big in Munich, Feb. 2003, Internet:
http://www.usatoday.com/tech/news/2003-07-14-linux-tech_x.htm, accessed: Aug. 28, 2003
33
But Microsoft did not convince the city of Munich. On May 28, the city council approved the Suse /
IBM proposal. Among the reasons, why the city of Munich decided for Linux, are the unpredictable
long-run cost of Microsoft upgrades. With the switch to Linux, Munich will become the largest tech
user to deploy Linux for everyday use on desktop PCs.
What happened in Munich shows how profoundly tech buyers' mindsets have changed. Five years
ago, as Linux was just starting to appear on the tech landscape, companies routinely snapped up
expensive technology without evaluating properly the different options.
In the memo, Ballmer said customers facing lean budgets are examining software such as Linux and
OpenOffice but Microsoft offers products that cost less for companies in the long term. "So-called
'free software' is the latest new thing," he wrote. "We will rise to this challenge, and we will
compete in a fair and responsible manner that puts our customers first. We will show that our
approach offers better value, better security and better opportunity."
The memo was well received among Advocates of OSS such as the OpenOffice suite were
heartened that Ballmer was at last taking their products seriously and mentioning them by name.
34
controversy in the Linux world, when it sued IBM, saying the company had incorporated SCO's
Unix code into Linux and seeking $1 billion in damages68. The company alleged, among other
things, trade secret theft and breach of contract. SCO then updated its demands in June, saying IBM
owed it $3 billion. In the meantime, it sent letters to about 1,500 Linux customers, warning them
that their use of Linux could infringe on SCO's intellectual property.
In addition to the controversial lawsuit over Linux code, SCO announced in June 2003, that it plans
to offer licenses that will support run-time, binary use of Linux to all companies that use Linux
kernel versions 2.4 and later. Eric S. Raymond and Robert Landley from the Open Source Initiative
OSI responded with the OSI Position Paper on the SCO-vs.-IBM Complaint 69, where they collected
relevant information about the history of Unix and Linux and tried to show, that SCO is wrong with
its claims.
68 Associated Press, The whole SCO-IBM-Linux-Novell-MS mess explained, August 2003, Internet:
http://www.usatoday.com/tech/techinvestor/2003-08-07-open-source-wars_x.htm, accessed: Aug. 28, 2003
69 Raymond, E. S.,Landley, R., OSI Position Paper on the SCO-vs.-IBM Complaint, August 2003, Internet:
http://www.opensource.org/sco-vs-ibm.html, accessed: Aug. 28, 2003
35
4 Engaging with the Open Source Community
A company which decides to use OSS has basically two options: Buying packaged OSS and related
services from a commercial OSS distributor ans service provider like Red Hat, Suse, IBM, JBoss
Group or they can get directly involved and interact with the open source community. The second
option includes a lot of different variants. This chapter explores all the different variants and its
advantages and disadvantages.
70 For further information about the Debian Linux visit http://www.debian.org, accessed: Aug. 28, 2003
36
applications make software available at completeness and quality levels ranging from immature pre-
alpha releases to stable, reliable and mature production releases. When choosing the maturity level
of a Linux distribution or software application, the enterprise is trading confidence of stability for
having the latest functionality. Obviously the more mature releases will be more reliable. However a
given enterprise may be willing to put up with a certain level of instability in order to gain access to
a highly attractive but relatively new functionality.
Almost all open source communities provide some form of technical support. In most cases this
support will be more formalized in those more mature projects that are associated with larger,
established communities. In some cases these communities even provide commercial technical
support, but at least email support from the developers in the community is usually available either
directly or via mail-lists, newsgroups or online forums.
At this level of engagement the enterprise will gain many of the advantages of using commercial
releases but will have substantially more choices available. By being conservative, an organization
can choose only those releases that are at a mature release level and enjoy a level of quality similar
to commercial software releases. The organization will also enjoy a significantly expanded range of
software choices. Using the advantage of being able to choose from the large number of applications
available, it is possible for an organization to build much of its IT application stack entirely with
Open Source software.
Typically an enterprise actively participates by having one or more of its employees engaged with
the open source communities of the applications that the enterprise considers critical to its business.
These employees, typically software developers, would work directly with the source code that
provides the key functionality as well as the more general capabilities of the application. This active
participation has the additional advantage of enabling increased understanding of the key
applications by the enterprise's employees. Often IT employees have a merely superficial
37
understanding of important software provided by a commercial software vendor. Working directly
with the source code of an important application can give employees a much better understanding of
the software, which in turn enables them to provide much better support for that application within
the enterprise..
This kind of participation we can observe when member of an open source community found a
commercial support organization for OSS. Other examples :
• IBM has this kind of participation with their leading function with the Eclipse Java development
platform71.
• Sun has has recently intensified its involvement with open source software. The company has
opened the open source collaboration platform java.net72 which is mostly were everyone is
invited to participate.
• When SAP switched the default database for their enterprise standard software from the
proprietary database SAP DB73 to Oracle, they released the source code of SAP DB. However,
they continued to lead further development of SAP DB in an open community process.
38
4.5 OSS without community
There are some typical scenarios for OSS projects without a community. In such a case, there is
little or no development, and a OSS project without community can be considered as dead. The
reasons for not having a community can be:
• the project has been initiated and managed by a single person. When this person leaves the
projects, there is nobody else to continue.
• the project is the result of releasing the source of some unpopular closed source software.
Companies and individuals are not interested in bad or unpopular software, even if it is for free.
39
5 Global trends in OSS adaption
5.1 Europe
With the decisions of the city of Munich to switch to Linux, Germany has recently earned a lot of
press coverage about its pioneer role for the adoption of Linux on the enterprise desktop. How does
this look in a European context ?
In the recently published paper "Linking up Europe: The importance of interoperability for e-
government services"74, the European Commission is placing open standards and open-source
software at the center of its efforts to promote inter operable e-government services. The
Commission, the executive arm of the European Union, argues in the report that open standards and
open source are crucial in making new e-government services work with each other and with
enterprise systems. The paper, which will be used as a reference point for policy and decision-
makers around Europe, is the latest document pushing forward the far-reaching European initiative
on e-government services. However, the document does not give exact directions to which extent
governments should encourage the use of open source and open standards.
5.1.1 Germany
An analysis75 performed by SOREON Research in 2003 revealed the following trends76 and
predictions for the German IT market:
2. The market for OSS will raise from 131 million Euro in 2003 to 307 Million Euro in 2007, with
a yearly average growth rate of 24%.
3. The largest growth rate in OSS deployments is expected for governmental organizations.
4. Main reason for using of OSS is the reduction of IT costs, followed by increased stability of OSS
solutions. Main reason for not using OSS are interoperability problems with proprietary
solutions.
5. Large companies do better benefit form OSS, because they often have already inhouse OSS
74 Commission of the European Communities, Linking up Europe, the importance of interoperability for e-
government services, July 2003
75 Soreon Research, The Market for Open Source Software in Germany, July 2003, Main results available at
http://www.soreon.de, accessed: Aug. 28, 2003
76 Blau, J., Update: High growth rates for open source in Germany, July 2003, Internet:
http://www.infoworld.com/article/03/07/02/HNgermanopen_1.html, accessed: Aug. 28, 2003
40
knowhow. The largest potential for cost savings are OSS office applications, databases and
Content management systems (CMS).
6. Many providers of proprietary software will change their business models and release the source
code of their software products.
5.1.2 Spain
In Spain, the government of Extremadura, a western region of Spain with 1.1 million inhabitants,
launched in April 2002 a campaign to convert all the area's computer systems, in government
offices, businesses and homes, from the Windows operating system to Linux77. More than 10,000
desktop machines were switched immediately, with 100,000 more currently being converted. The
regional government paid a local company $180,000 to assemble a set of freely available software,
including operating system, word processor, spreadsheet and other applications. The government
also invested in creating customized software for accounting, tracking hospital patients and crop-
yield management that the agency will distribute free to citizens. The European Economic
Commission is promoting it as a model for the rest of the world, and officials from governments as
far away as New Zealand and Peru have inquired about duplicating the region's efforts.
5.2 Asia
There exists initiatives in Asia78 which try to advocate the use of OSS among the populations and
bring together academic research and industrial and governmental use of open source software.
When analyzing Asian sources about OSS, an interesting fact is that they talk about free software
and not about open source. On the one hand sometimes it might be a translation-into-English issue,
on the other hand, because of economic reasons, free availability in the sense of no license costs is
more important then source code availability.
• FSIJ (Free Software Initiative Japan) FSIJ was created in 2002. The Purpose of FSIJ is to
advocate the principle of free software, collaborate with free software groups from all over the
world and organize international symposiums, and facilitate the release of high quality free
software from Japan to the rest of the world. concrete step towards the international collaboration
Outside of Europe and North America, Linux is beginning to emerge as a serious desktop
alternative. Countries in Asia are adopting Open Source at high speed.
41
In much of Asia, the only real competition to Linux-on-the-desktop comes from software piracy
with bootlegged Microsoft products.
Among the barriers which have prevented in past widespread adaption of Linux at the enterprise is
Microsoft's disinformation campaign about Open Source's reliability and consistency, which has
been particularly effective in discouraging early adoption by many Asian companies and
government agencies.
42
6 OSS from economic point of view
This chapters introduces economic definitions and presents models which can be used to analyze IT
costs. The influence of open source on devaluation of software is analyzed and business models
which allow independent software vendors to benefit from OSS are explored.
One of the major manager's tasks today is the reduction of the total cost of IT projects and solutions.
A good starting point for approaching that task is the cost analysis based on TCO.
There does not exist a general definition of the total cost of information technology. TCO basically
consists of79:
For the calculation of the TCO value there does not exist a clear rule, the range of values is broad.
The TCO analysis does not include effects of service level agreements, availability requirements
and evaluation of the business case80.
TCO is dominated by costs which are given, and can't be changed. Therefore TCO can only be
considered as starting point for a cost analysis. The concept only describes a point in time and
reveals the current cost situation.
43
6.1.2 Components of a TCO calculation
The following sections describe all costs which need to be considered in TCO calculations81:
The cost calculation first considers all investment costs for hardware and software and maintenance
contracts. Additional service costs are calculated for defined functions based on full time employee
(FTE or Full Time Equivalent). Cost are also calculated for outsourced functions.
The calculation also contains quality components like type of installed hardware, operating systems,
standard applications, quality of service (availability, unplanned down time) and the extent and
volume of performed activities (for example at IT support call centers: success rate of immediately
resolved problems).
Another integral part of the TCO view is the quality of service, which will be provided to the user. It
is a big difference, whether in a support contract a response time of two hours or four day has been
negotiated. Additionally all core applications (like ERP-Systems) and data storage (Storage Area
Networks, Tape Robots) and network infrastructure need to be considered, to reflect all those costs
in a per user base.
6.1.2.3 Administration
Another important point is staff. On the one hand there is IT staff responsible for the production
environment and providing support for the users, on the other hand, there are employees planning
the infrastructure and taking design decisions. This is an area where proper planning and usage of
system management tools can help to reduce the demand for human resources.
Well organized IT organizations with a standard compliant software environment, user support
embedded into the organizational work flow, and centralized operations with control functions and
software distribution allow significant costs reductions per user.
The satisfaction of the user is a major factor for the success of business processes. If users are
perceiving IT staff as service provider, then acceptance and satisfaction with IT support is
improving. On the other hand, if users locally try to solve problems by themselves and the help of
others, instead of contacting the IT support, the working hours of the users are consumed to resolve
81 Steigele, H., IT-Sourcing Step by Step (Draft), Cascade IT, 2003
44
tasks which are supposed to be resolved by the IT support staff. The effect of this “hidden costs” or
productivity loss is the user involvement and needs to be considered and determined with the help
of interviews or questionnaires.
Peripheral Supplies
integration
Infrastructure Rental
integration
Integration of Energy
central Components
Decentralized Communication
Hardware
Decentralized Service Level Agreements, spare
Software parts
Modification of IT support operational cost
building / rooms
Organizational and Human resources costs
trainings costs as • Overhead
consequence of • Accounting
platform change • others
Training Training costs
• IT Support • IT Support
• User • Users
Table 2 TCO cost components82
45
6.1.3 TCO - Summary
The TCO value shows a short sighted views of IT costs. It can only be used as an entry point for a
more profound analysis, which also covers longterm aspects.
However the TCO indicates, whether IT services should be externalized in the future. Such an
outsourcing provides cost reduction, because IT operations and application support are provided by
a professional organization and backed up by a Service Level Agreement (SLA).
The ROI is calculated by dividing net profits after taxes by invested capital. When applying ROI
calculations on Information Technology, the following points need to be considered83:
• The invested capital: includes costs of acquisition, installation, system integration and change
management.
• Profit: the profit is the sum of economized process costs and expected additional profits because
of the new implemented systems minus the operational costs and those costs which have been
caused by efficiency loss in the service sector.
The ROI calculation for one year is often too short sighted. Therefore people often use dynamic
methods of investment calculation. With this approach, capital flow of the observed resource over
its whole operational duration can be analyzed and results are getting more accurate.
The results often are represented in a yearly annuity, a so called internal rate of interest, or a real
monetary value calculated for a specific point in time. Important for these values are the following
points84:
3. The annuity or the duration of amortization must be in sync with the operational duration of the
observed resource.
46
Compared to the TCO analysis the ROI analysis is based on a longer duration, which has a strong
impact on process costs. Therefore the two models do not deliver comparable results.
But real life has shown, that the problem is within the process costs. This kind of costs calculation is
not common yet at most companies and process costs can only be estimated.
Costs estimations for a longer term, are integral parts of the “business value”. The term belongs to
the same category like TCO and ROI, but has included aspects of TCO and ROI.
Value
Busines
s Value
ROI
TCO
Time
Figure 1 IT sourcing value
Figure 185 shows the total IT sourcing value as function of time. Figure 286 illustrates how the value
can be divided into its different calculation components
47
political value
image value
risk cost reduction
sourcing value
process cost reduction
service value
product value
total sourcing value
purchse cost
new coordination cost
process tuning cost
sourcing cost
risk cost
image cost
political cost
By taking a closer look at this calculation tree, one can see the cost components but also the value
components of future investment. It is also possible to work with time separated payment flows.
Problems might appear at the elements process cost reduction, risk costs and image value.
Process cost calculation requires to know the individual process costs. Unfortunately it is not
common yet at most companies to determine process costs87. The same can be said for risk costs.
First, awareness of risk costs requires a well established risk management at the company. Second,
it is impossible to calculate precisely long-term risk. Third, risk calculations rely on probabilities,
which need to be defined in the context of the calculation.
It is a calculation with estimated values, and therefore the result will also represent an estimated
value.
The reputation element also contains a strong estimation aspect. If we take the shareholder value,
investors honor platform migration or outsourcing deals with discount or raise of shares. These non
verifiable elements are important in reality, but can only be estimated.
With sourcing not everything can be calculated exactly. But just the consideration of the presented
schema gives some basic guidance.
48
6.2.1 The decision
After having prepared a cost analysis of the different options of an IT strategy, one of the
alternatives need to be selected ? Independent of how much has been invested into preparations, the
decision is what matters.
Beside of financial aspects there are usability and political aspect which have significant influence
on the decision.
A very simple question might help to simplify the whole problem88: “what is the purpose of the
acquisition ?”
IT decisions often follow the pattern to improve its transactions and the related cost situation. In
such a case it is justified, to focus on cost situation and ROI, beside of providing the required
functionality of course.
• customer satisfaction
• user acceptance
A purely financial judgment of the investment based on TCO and ROI will not give precise results,
because the correct parameters required for such a calculation are extremely difficult to obtain.
1. the final decision should be compatible with the overall company strategy.
2. The higher the importance of IT for the transaction, the more should be chosen the alternative
which requires the least expense in regard of implementation, maintenance and risk.
3. The higher the importance of IT the more should be considered aspects like business value,
flexibility, competitive advantage, synergies and value for the user.
Once the decision has been taken, a strategy needs to be developed for having control of
49
success/failure of the investment. This at least has a positive impact on the learning curve for
possible follow up investments.
50
General Availability
Product of open source Product
availability alternative obsolescence
Revenue curve
without open
source competitor
Revenue
Time
Research
Development
Testing
Production, Marketing,
Costs
Martin Fink analyzes in91 the value of software as function of time by comparing the model with
the economic return model for prescription drugs. From this comparison he derives the model for
the open source effect on value of software as function of time. Fink's model is presented in figure
1, which shows the effect of open source on the value of software and the power of commercial
vendors to command a premium price for that value. Almost any commercial software has value.
Over the time, it is normal that the value of the software decreases, from the viewpoint of the
customer. The devaluation comes from competitive forces and the natural commodity effect. For
example, if we take a look at computer hardware, 10 years ago, a computer mouse was usually an
add-on, today, one would be surprised if a mouse would would not be included when buying a new
PC. Software vendors typically respond to the natural devaluation process by adding new features
and new capabilities, which are refreshing the value of the software on an ongoing basis. When it
91 Fink, M., The Business and Economics of Linux and Open Source, P.159-167
51
comes to enterprise software, vendors can maintain value by being the only entity, to support the
software they developed. Another factor which protects enterprise software vendors is the barrier to
entry. Once a customer has committed to a particular software choice, as long as the ongoing cost
for maintenance are less than the cost of replacement, the vendor can continue to charge a premium
price.
Without open source software, Companies have been trying to protect their investments into
Research developing and testing with patents. And to maximize revenues, companies had to keep
the operational cycle as long as possible. With open source software, the situation looks different.
Development cost of OSS is significantly smaller, because OSS usually is built on top of other open
source frameworks or libraries. This is possible because open source communities give freely their
intellectual property away (and expect the same in return). As a result, if the software vendor want
to keep his position on the market, he needs to respond to the OSS attack with product innovation,
lower price or additional services which cannot be provided by the open source solution.
• It often takes takes less time to create a comparable solutions to a proprietary software, because
there a large number of developers involved.
• Additional features, which must be purchased in case of proprietary software (even if base
software is free or has a very low price) are included with OSS at no charge and no hassle to
purchase and activate a license.
• OSS will not eliminate the barrier to entry to the markets of proprietary software, but because
license costs are usually non-existent, that barrier will be significantly reduced. The size
remaining final barrier is tied to support services and depends on the end-user willingness to
engage with the open source community for further reduction of that barrier.
Another challenge for software vendors is predictability of market acceptance of OSS threatening
his product. If the vendor doesn't pay attention to community efforts currently underway, he can be
surprised by an open source version of his product. It is difficult to predict when an open source
product will reach the threshold of “good enough” in the majority of customer's mind.
52
take advantage of devaluation and the open source community rather than to react in a defensive
way. But to benefit from open source, it is not enough just to share the idea, it might be necessary to
change the traditional business model and such a structural change might be difficult to apply, or
does not fit into the general vision of a firm, therefore some companies prefer to react defensively to
open source.
According to Martin Fink92, closed software vendors have the following three technical possibilities
for defensive reactions to the threat from open source competition:
Distributors of open source software and open source communities are exposed to the threat of
patent attacks from closed software vendors. But software patents are topic of controversial
discussions. Risto Sarvas researched the costs and benefits of software patents93 and came to the
conclusion, that the processes to grant patents are very slow , the granted patents are of low quality,
but in some case it is a appropriate way to protect a software business.
An example of a software patent is the compression algorithm used for GIF images, which has been
patent protected by Unisys94. The open source community responded by simply using different
graphic formats like JPEG or PNG. More complicated is the situation when the patent holder is
threatening without revealing what exactly has been infringed. This is the case with current lawsuit
SCO / IBM. As soon as the community gets to know, what part of the code has been affected by a
patent enforcement, developers substitute the code with some other code, which provides the same
functionality, but uses a different algorithm. A major risk for aggressively enforcing patents is the
uncertainty about existing prior art. Software engineering is very complex and it is extremely
difficult to verify that absolutely no prior art exists when awarding or enforcing a patent. Aggressive
patent use can only be recommended for companies which have nothing to loose.
Today most software is written in a modular way as collection of functional layers. This is the most
flexible way to architect software. But it in case of proprietary software, this method opens the door
for replacing a specific layer with OSS. Such a replacement process usually starts with the bottom
layer. And then, OSS replacement of the next layer will follow. By creating monolithic software,
vendors hope to prevent commoditization from the bottom of the software stack. However,
92 Fink, M., The Business and Economics of Linux and Open Source, P.167-168
93 Sarvas, R., Costs and Benefits of Software Patents to Society, Helsinki University of Technology, 2002
94 More information about the Unisys GIF patent available at http://www.unisys.com/about__unisys/lzw/, accessed:
Aug. 28, 2003
53
monolithic architectures have several disadvantages like maintainability and scalability, and it is
very likely that the community creates better designed software solution and commoditizes the
monolithic software all at once.
Following two examples of OSS with modular architecture taking away significant market share
from monolithic closed source applications:
1. Java applications Servers – While open source JBoss95, which is based on a modular plugin-
architecture became more and popular, many other closed source monolithic solutions became
victims of the application server vendor concentration and disappeared from the market.
2. Integrated Development Environments: OSS Eclipse96 has such a modular architecture that it
is considered as a platform, not as an application. Eclipse is very well accepted in software
industry and it is integrated into commercial products from IBM and other vendors.
6.4.1.3 Competition
A common reaction is the attempt to compete with the open source community. But competing
against a community of first-rate developers who are not (or less) motivated by profit and not
constrained by time limits, is a difficult exercise, and it is common, that sooner or later, time and
resources will make such a move impossible to sustain.
The better option is to participate actively with the open source community and use devaluation to
one's own advantage. In most cases, Software vendors benefit from open source, if they engage with
the community and offer their customer open source software for the lower layers of the software
stack. With this approach, the software vendor does not need to apply resources on technology that
customers are considering base functionality and are not interested to pay for them. Figure 2
illustrates that software vendors can apply their resources on the upper layers of the software value
stack, where they generate more value and return to the company97.
However, Software vendors who intend to use OSS for the bottom layers of their software, should
keep in mind, that only by actively engaging with the community they have the chance to influence
the future direction of the bottom layer of their end product.
54
Community Software vendor
resources resources
Value Typical
software
solution
stack
Resources
Figure 4 Resource allocation correlated to value delivery
If OSS is used as part of a commercial product, development costs usually can drastically be
decreased. Of course this only works if the open source project has an active community outside the
company. On the other, if the company is operating in a competitive market, the OSS modules of
their commercial product is open for competitors also to be used. In such a case, it should be
analyzed whether the reduction of development costs are compensating the quantified competitive
disadvantage. Often there is no such a competitive disadvantage because OSS is playing a
significant role for product X of company Y but just because of the specific design of product X,
and not because OSS plays a significant role in the market the company is targeting. The software
vendor who plans to integrate OSS in his product, needs to think about licensing issues. Does the
license of the integrated OSS require to feed back changes made to the source code ? And if it is not
required, would it still have any benefit for the company to get engaged with the open source
community ?
For better illustration of the business model based on integration of OSS into commercial software,
55
an example will follow.
The current Apple Mac Operating System OS X is based on Free BSD, a Unix derivate which is
distributed under the BSD license. Apple modified the BSD code to run on the special
Microprocessors inside the Apple Mac. Apple called the modified version Darwin and released the
source code98. The company decided to feed back the modifications to the Free BSD community, if
this modification leads to an improvement for the Free BSD code itself. The BSD license does
require that. However, doing so, they facilitate synchronization between the two OS X and Free
BSD. Improvements at Free BSD can easier be ported to Darwin if the two operating system are
similar. The Apple Mac OS X itself is built on top of Darwin, but is not open source anymore. One
could argue now, that Apple is facilitating others to become their competitor, because the core of
their operating system is open source. But there does not exist any competitor in this market.
Microsoft can not benefit from that, they have their own universe. Linux is not a competitor for the
Mac, even if the Linux operating systems is Unix-like. The market for the Apple Mac is the segment
of people looking for a computer with extraordinary design and easy to use. The market for Apple
Mac is based on the consumer lifestyle. Before Mac OS X, Apple had its own proprietary operating
system, but they realized they can better compete on the market by concentrating on their core
competency, by delivering well designed boxes and cut development cost by having their operating
system based on an existing, stable and mature OSS.
In the past, Microsoft has ignored OSS100, at present Microsoft is considering OSS as competitive
challenge101. Before doing any predictions how Microsoft will deal with the OSS threat in the future,
it makes sense to take a look at how Microsoft has dealt in the past with devaluation of its software,
98 More information about Darwin can be found at http://developer.apple.com/darwin, accessed: Aug. 28, 2003
99 Vendor statistics available at http://www.theregister.co.uk/content/68/31880.html, accessed: Aug. 28, 2003
Authors note: Only smartphones from Orange are running on the Microsoft operating system, the smartphones
from Nokia and Sony Ericsson are running on the Symbian operating system.
100 See Halloween documents, chapter 3.4
101 See City of Munich decides for Linux, chapter 3.4
56
how they moved into new markets and it will be clearly visible that Microsoft flexibility and power
should not be underestimated.
After 1995, the Internet started to attract the masses. The opinion was widespread, that Microsoft
was a huge, inefficient, PC-bound company that couldn't adapt to the brave new world of the
Internet. Around that time, MS started to hire developers for the Internet Explorer team to react to
the competitive threat created by Netscape102. Netscape seemed to have a competitive advantage
because of its first mover role. We have a similar situation today with Google and Symbian.
The browser war was not the first time Microsoft had demonstrated that it can fight back. The
previous ten years were littered with once-powerful companies that had been literally squashed:
Ashton-Tate (the makers of “dBASE”), Borland, WordPerfect. The difference with Netscape was
that it all happened so publicly and visibly. Software markets had started to operate differently. But
it didn't make much difference, not in the end, except for one thing: no one can say they haven't seen
that Microsoft won't give up easily, or that, having missed a trend, they can't adapt.
The following examples show how Microsoft moved successfully into new markets and occupied a
dominant position within short time.
• In only a few years they have carved out a decent portion of the server market, which has been
previously dominated by Unix servers.
• They have carved out another portion of the market for Internet applications.
• Another example is the PDA market. PocketPC-based handhelds have been steadily growing,
while Palm handhelds seem not to be popular anymore.
• In only a couple of years, the xbox went from being a bunch of marketing documents to the
second gaming platform of the planet.
• Office, which everyone just seems to ignore, is a monopoly that Microsoft did not control at all
in the last years, and it is currently the biggest source of revenue and profits for the company.
Microsoft is the biggest software company in the world. They have managed software projects
bigger than anyone else, with bigger deployments than anyone else, and pulled it off. They have
moved successfully into new markets, even as many of their attempts have failed. They have tens of
thousands of really smart engineers, they have excellent management and a good software
development processes. With their operating system Windows and their office suite, they have two
strong monopolies and a few weaker ones like Encarta Multimedia encyclopedia for example. They
102 Rosenberg, Donald K., Open Source: The Unauthorized White Papers, 2000, P.34-35
57
have enough cash in the bank to underbid any proposal of their competitors.
History shows that only few companies managed to beat Microsoft on the long term. One of them is
Intuit103, they managed to become and to remain market leader for personal finance software with
their product “Quicken”.
There is one fundamental difference between the previous threats Microsoft managed to deal so
well with and the current threat from OSS. While previously competing companies had to make
profits with their proprietary product, OSS is not tied to company, is not a specific product, and
OSS as such does not generate profits. Only time will tell how Microsoft will deal with competition
from OSS, whether current market share will change in favor of OSS or Microsoft, and whether
Microsoft will get significant market share at new technologies like operating systems for
cellphones.
103 More information about Intuit and Quicken can be found at http://www.intuit.com, accessed: Aug. 28, 2003
58
7 OSS and usability
OSS has gained a reputation for stability, security, reliability, efficiency and functionality that has
surprised many people in the IT world. But most of this software has been used by highly skilled IT
experts. What is most criticized by average users, when they makes their first experience with OSS,
is the poor usability.
Usability is typically described in terms of five characteristics: ease of learning, efficiency of use,
memorability, error frequency and severity, and subjective satisfaction.104 Usability is separate from
the functionality of software and from other characteristics such as reliability and cost. That there
are usability problems with open source software is not significant by itself; all interactive software
has problems. To understand the usability of OSS we need to consider the following points of the
open source development process:
• Developers are not users - This has been pointed out by Nielsen105. For some advanced OSS
products targeted to expert users, developers are indeed users, and these software products often
come with user interfaces that would be unusable by a less technically skilled group of users are,
perfectly adequate for their intended elite audience. However when designing products for less
technical users, all the traditional usability problems arise.
Nichols and Twidale analyzed usability problems of OSS in 106. They compared open source
development with commercial systems development in the early years of computing, when
software was designed by computing experts for other computing experts, but over time an
increasing proportion of systems development was aimed at non-experts and usability problems
emerged. Nichols and Twidale claim, that the transition to non-expert applications in OSS
products is following a similar trajectory, just some years later.
• OSS developers are more motivated to improve functionality than usability – voluntary
developers work on the topics that interest them and this usually does not include features for
novice users. Adding functionality or optimizing code provides opportunities for showing off
one’s talents as a hacker to other hackers. If OSS participants perceive improvements to usability
as less high status, less challenging or just less interesting, then they are less likely to choose to
work on this area.
59
• Usability experts do not get involved in OSS projects – OSS draws its origins from a culture
which is extremely welcoming to other computer freaks, comfortably spanning nations,
organizations and time zones via the Internet, but less welcoming to people with intellectual
cultures focused on topics as graphic design, ergonomics, psychology and sociology107. But this
interests and skills are required to improve usability. To sum up: on the one hand, usability
experts are not interested to participate in OSS projects, on the other hand ,OSS projects do not
have the financial means to hire in missing skill sets to ensure that user-centered design expertise
is present in the development team.
• Solving usability problems does not fit well into the open source development model - open
source developers have skills and tools to solve functional problems. And they have
communication and collaboration tools to to benefit from the network. But the common OSS
collaboration tools do not offer any help for improving usability in a collaborative manner.
• Encouraging end user involvement – the possibility for end users to test, report bugs and
requesting new features at any point of the development cycle should be facilitated. Just the
availability of the source code does not help end users to install and test the software. Early
versions of new software releases should come as easy installable package and documentation
targeted to end users should be included. In the lifecycle of commercial software there exists
often a public beta phase. Everybody knows that a beta release of a software contains bugs, but
the release comes with documentation and can be installed comparable to the the final version.
Alpha and Beta releases of new OSS should ship with an integrated, Internet based bug reporting
tool. Such tools encourage end users to report bugs and usability problems. Most OSS which ship
with such a bug reporting tool, like the beta releases of the Mozilla web browser with talkback108
option enabled, focus only on functional problems and reporting of software crashes, but not on
usability problems.
• Commercial approaches - there exist a lot of software products, which use OSS for providing
107 Nichols, D. M., Twidale, M.B., Usability and Open Source Software, University of Waikato, 2002, P.5
108 For more information about the Mozilla web browser and talkback enabled installers for Mozilla beta releases
visit http://www.mozilla.org, accessed: Aug. 28, 2003
60
the core functionality and have wrapped a commercial layer on top of the OSS framework to
provide a graphical user interface (GUI). With this, approach the end user software is not open
source anymore, but if the software vendor actively participates at the development of the OSS
core layers, the final software product benefits form the best of both worlds. Apple Mac OS X109
serves as a prime example, the core is an open source project maintained by Apple and derived
from BSD, an open source Unix variant, which provides security, stability, performance and
scalability, the top layer is the Apple GUI, designed by world class graphical artists, and famous
for being easy to use.
109 Chapter 6.4.2.1 describes the open source relation of the Apple Mac OS X
61
8 OSS examples and case studies
Linux servers are dominating the racks at ISPs and Linux servers have found their way into
enterprises as Intranet web servers, mail servers, database servers file servers and print servers.
8.1.1.1 Strength
• Reduced license cost – If downloaded form Internet repositories, then there is no license cost at
all. This definitely is a point where Linux can't be beaten by MS Windows.
• Security - Because of its open source nature, there are more “eyes” watching for security
problems. Linux is based on a strong security design and concept.
• Stability – The Linux Kernel has been designed for high stability. Properly configured Linux
servers never need to be rebooted, expect for hardware upgrades.
• Manageability - Linux server can run headless,112 this makes remote administration of server
farms very easy.
• Possibilities to customize - There are unlimited possibilities to customize and optimize a Linux
server to fulfill specific customer requirements.
• No vendor-lock-in - There are many different suppliers for Linux based technology and services
110 Gartner, Linux poised for the Desktop failure,
http://www.zdnet.com.au/newstech/os/story/0,2000048630,20269971,00.htm, Nov. 2002, accessed: Aug. 28,
2003
111 Summary of the IDC server market share study available at: http://news.com.com/2100-1010-5069581.html,
Aug. 2003, accessed: Aug. 28, 2003
112 “headless” means that there exists no requirement to attach a monitor and a keyboard to the server,
administration of the server can be done remotely from another computer.
62
(Redhat, Suse, IBM, HP and many smaller companies)
• Quick release rate - Fixes and patches are released quickly, an order of magnitude faster than
commercial software. This is extremely useful for security patches. A consumer does not need to
wait until a particular vendor has released a patch. The Linux community, which has hundred
thousands of developers, usually provides security fixes only hours after the detection of a
security whole.
• Long term accessibility for upgrades - While Windows NT is a product and now has reached
end-of-life and Microsoft does not provide support and security patches anymore, Linux is more
a platform and can gradually be upgraded to newer releases if desired. There is no vendor
dictating when the servers needs to be upgraded to a new release.
8.1.1.2 Weaknesses:
• Usability - Linux, and open source software in general, are usually focused on providing
functionality, and their developers often do not pay much attention to usability aspects. Graphical
user interfaces (if they exist at all) are poorly designed, and less intuitive to use, compared to
Microsoft's offerings.
• Documentation - While Windows offers easy built-in access to user documentation, at Linux the
huge amounts of documentation are available, but the documentation is not structured in a way
for easy centralized access.
• Support Staff - Linux experts, which often have a Unix background, are more expensive and
more difficult to find. The selection of the right people is also more difficult, because
certifications comparable to Microsoft Certified System Engineer (MCSE) are not common and
only available for specific Linux Implementation from Red Hat, IBM or other vendors.
• Lack of Ownership - Users want accountability and want to make someone responsible for
possible problems with their operating system.
8.1.1.3 Opportunities
• General business trends - The recent adaption of Linux by big companies like IBM, Oracle and
HP has significantly increased the credibility and acceptance of an operating system which
started a decade ago as spare time activity of a student from Finland.
• Many distributors - While Windows is coming from a single vendor, there are hundreds of
Linux distributors, each one targeting different customer needs, different hardware platforms or
63
different niche markets. The high diversification of distributors might be confusing for some
users.
8.1.1.4 Threats
• Risk of fragmentation - So far, there exists only one Linux kernel. But code fragmentation
happens very often in OSS. Typical reasons are project administrator doing a poor job or
personal conflicts between developers. This often results in a code fork and project split. A
typical example are the different open source BSD Unix versions: Open BSD, Net BSD, Free
BSD, and others.
• Microsoft's marketing machinery - Linux companies have very limited marketing possibilities.
Microsoft however has financially nearly unlimited possibilities for any kind of sponsorship
which can be interpreted as purchase of mind share and political power. This is a threat for
Linux, especially when Microsoft is sponsoring research institutes. The results those sponsored
research institutes present, are not independent anymore. The same applies to Microsoft
sponsoring of educational institutions.
• Microsoft's monopolistic power - Microsoft has enough financial resources for lowering the
price of its services and products for specific projects just to underbid any Linux based offering.
64
8.2 J2EE Application Server: JBoss
The market for Java based Enterprise Application servers is dominated by Websphere from IBM
and Weblogic from Bea Systems. Only for simple application, which only use a small part of the
J2EE specification, OSS like Tomcat has been used on large scale. In the upper league, for a long
time, Open Source application servers did not even play a marginal role. Recently the Bea Systems /
IBM equilibrium of power has been attacked by JBoss, a J2EE compliant OSS application server
with one of the most professional OSS organizations behind it. The market share of JBoss is
difficult to estimate, because there is no licensing involved, which could reveal the number of
production sites running JBoss. And because enterprise applications usually are running inside
company intranets, netcraft113 server statistics are not representative as for web servers. However,
JBoss has been downloaded over two millions of times in 2002 and it is very popular among
software developers, because it is easy to use and has some technical features which not even its
commercial commercial competitors have.
JBoss is OSS project with LGPL license. The core developers of JBoss are all employed at JBoss
Group, a service company founded by JBoss creator Marc Fleruy.
According to JBoss creator Marc Fleurys, the visions of JBoss are highly profit oriented and very
differentiated in comparison to Linux114: “The difference is that Linux is not really structured
commercially, meaning (Linux creator) Linus Torvalds is off doing something else. And there is
Red Hat as a third-party packager. Whereas we are very much for profit. There is a lot of services
in middleware, as compared to Linux, where you do not have so much that you do and consult
around. Middleware is very consulting intensive. We have a profitable consulting operation that
has all the developers in it. We are borne from inside the group, and the company grows from
inside the group, which is a different model. We're highly structured and commercially focused. We
just use open source as R&D and recruitment if you will”
8.2.1.1 Strengths
• No license cost - Depending on the configuration, application servers from Bea Systems and
113 Netcraft web server statistics show on what operating system Internet web servers are running, For more
information visit http://www.netcraft.com, accessed: Aug. 28, 2003
114 LaMonica, M., Behind the story at JBoss, Internet: http://news.com.com/2008-1082_3-994819.html,
accessed: Aug. 28, 2003
65
IBM have been around 10,000 U$ per CPU. For a clustered server farm, which consist of several
or even dozens or hundreds of single computers, a lot of money can be saved. And there is
another issue involved with licenses. While activating a license technically usually straight
forward and done within minutes, at large companies the work flow to obtain license involves
many steps, and developers don't like to loose their time with such time consuming tasks.
• Customization options - The purpose of the J2EE standard is to provide a platform, where the
components of the platform should be vendor-independent and interchangeable. J2EE propagates
best of breed approaches. JBoss goes even a step beyond, because of its modular structure, it
allows to use only the required parts. And if even more customization is required, than every
module of JBoss can be modified independently of the others.
• Ease of use – This is one of the main reasons why JBoss is so popular among software
developers.
• Professional Support and Training - JBoss Group provides globally and for competitive
conditions consulting, support and training. All members of JBoss Group are core developers of
the JBoss software.
8.2.1.2 Weaknesses
• Usability - OSS developers usually concentrate on the core functionality. This is valid for JBoss.
The administration console is very simple and limited, compared to the products from Bea
Systems and IBM.
• Not officially J2EE certified - SUN is the authority which licenses the J2EE certification kit to
interested parties. All major closed source J2EE vendors have their products J2EE certified.
Theoretically, JBoss Group, the professional services company which is staffed with the JBoss
core developers the could also license the certification kit, and if the JBoss server passes the
J2EE test, they could call their server J2EE certified. However, this has not happened so far,
because of the following reasons:
• the J2EE certification kit is closed source and an NDA must be signed to use is, and this
doesn't fit into the distributed, collaborative JBoss approach of Software development.
• Project too much under control of one commercial company – Marc Fleury and his JBoss
Group company decide which contributions will be accepted for inclusion into the main software
66
releases. In case there are conflicts of interests, they will decide for the benefit of JBoss Group.
8.2.1.3 Opportunities
• Current economic situation, where IT investments with high license costs are difficult to get
approved.
8.2.1.4 Threats
• IBM and Bea Systems have been reducing the price of their licenses over the last years.
Therefore the argument of reduced license cost, which often only play a marginal role at large IT
integration projects, is loosing even more of its importance.
67
8.3 Office Suites: OpenOffice
OpenOffice offers all the basic and most of the advanced features that Microsoft Office has, but for
free. Solveig Haugland, author of several books dedicated to OpenOffice beliefs that the future
looks bright for OpenOffice: “... use will continue to grow, perhaps not from commercial centers
first, but from the public sectors, school and home use. But, eventually, because students will have
used it in high school and college, or because governments require an open-source file format,
OpenOffice.org and StarOffice will become as widely used or more so than Microsoft Office is
now.”115
8.3.1.1 Strengths
• Zero license cost - While this point is less relevant at most typical Open Source areas like server
software and frameworks for software development, OpenOffice actually is a end user product
for desktop., an area where license costs have a strong impact on the total cost if many licenses
are required. The license issue is also important for home users. Whether they have to pay several
hundred dollars for the legal use of an office suite or not, is a big difference.
• File Format Compatibility - Beside of reading all important Microsoft File formats, OpenOffice
has also support for direct output to PDF or SWF. PDF is the most important format for viewing
text documents on the Internet. SWF is the vector file format of the Macromedia flashplayer,
which is currently installed on 98% of all web browsers worldwide. SWF output allows to view
presentations created with OpenOffice (or Microsoft PowerPoint and then imported into Open
Office) on any web browser.
• Cross platform compatibility - While MS Office is only available for MS Windows and for
The Mac, Open office is also available for Linux and Solaris. And beside of English, 27 other
languages are supported.
8.3.1.2 Weaknesses
• Performance - MS Office is faster. Not much but enough for a user to recognize the difference.
68
• Help System - OpenOffice doesn't have popup wizards (“agents”) which help the unexperienced
user to perform actions.
• No Visual Basic – Many companies and organizations have developed with MS Visual Basic
applications and integrated them into MS Office. These application cannot be migrated to
OpenOffice.
• Support - There is no globally organized, professional support for Open Office. Only for
StarOffice, the commercial distribution of OpenOffice, SUN provides support.
8.3.1.3 Opportunities
• Microsoft has announced to release new versions of the Internet Explorer only bundled with the
Windows operating system and anymore and not to release anymore any new IE Web browser for
the Apple Operating system. Whether this also will happen to the MS office suite for the Apple
Mac, purely depends on Microsoft and its business strategies, which are directed to push users
from upgrading their system to the newest Microsoft Windows operating system, otherwise they
not receive any support and software updates like security fixes anymore. This threat is a good
opportunity to get the mind share of the Apple users. If they want to stay on the safe side for the
future, they will stick to Open Office.
8.3.1.4 Threats
• MS Office popularity - Everybody knows to some extent MS Office. And if somebody does
not, it is easy to find somebody who has basic or even profound knowledge of MS office. At
workplace, just asking the guy next to you, is enough to resolve basic MS office user problems.
The situation is different with Open Office. The typical Linux user, which uses Open Office
sporadically to create and print out some paperwork is more technically interested and is usually
not the corporate user which works with Open Office or MS Office on a daily base.
69
8.4 Web browser: Mozilla
When Netscape released the sources of their browser, they called the browser project Mozilla. After
having lost the browser war against Microsoft, Netscape was bought by AOL. Mozilla continued to
be the code base for the Netscape Browser, which was the AOL default browser. Then AOL made a
deal with Microsoft and used Internet Explorer (IE ) as the default browser browser for their
customers. In July 2003, AOL stopped their engagement with the Mozilla community 116, after
having pledge two million dollar into a new independent organization called Mozilla Foundation, .
Mitch Kapor, the creator of Lotus 1-2-3, one of the first killer-applications for PCs, is Chairman of
the Mozilla foundation. The separation from AOL has been well received at the Mozillla
community, and with the new independent organization, the future looks bright for Mozilla.
8.4.1.1 Strengths
• Zero license cost - One could argue, that IE also can be downloaded for free. This is true for the
moment, but it is likely that Microsoft will charge for the browser in the future, because
Microsoft only offered IE for free to win the browser war against Netscape. It does not make
economical sense for Microsoft to offer a product for free, if they have the monopoly on that
market.
• Innovation - Mozilla offers a lot of advanced features like tabbed browsing117, integrated
Google-access, e-mail client with Spam filtering and much more.
8.4.1.2 Weaknesses
• Many websites only optimized for IE - While simple HTML content looks identical on Internet
Explorer and Mozilla, websites which are using Cascading Stylesheets CSS118, Javascript119 and
other advanced features might look different on different Browsers, because some of those
features haven't been properly standardized or the browser developers did not implement those
features properly. And because IE has about 90 % market share, Website developers often
optimize their content only for Internet Explorer.
70
• Too many features - Mozilla is more than just a browser, it is a suite including e-mail client
and web page development tools. The download installation file is more than 10 MB large. Most
people just want a good browser and don't want to download additional tool they never use.
• Slower than IE – IE is tight integrated with Windows and therefore slightly faster to start up and
to render web pages.
8.4.1.3 Opportunities
• No more free stand alone version of IE - Microsoft announced in June 2003 to deliver the next
version of the Internet Explorer only bundled with the Windows operating system and not to
offer free downloads anymore. If this plan will turn into reality, it is likely that not everyone will
agree to buy a new operating system just for having a new browser application and might switch
to Mozilla.
8.4.1.4 Threats
• IE Integrated into Windows - Internet Explorer is bundled with Window, this is a very big
competitive advantage fro Microsoft. If a new version IE, which is planned to be shipped with
“Longhorn”, the next Windows Operating system in 2005, will be significantly quality improved,
there is no reason for Windows users to download and install another browser application.
71
8.5 ERP/CRM Software: Compiere
Compiere120 is an open source ERP software with integrated CRM solution for streamline order
processes on the web and customer centric point-of-sales (POS), for inventory management,
automated accounting, customer service and more. Compiere is targeted to small to medium sized
businesses in distribution and service, and is designed for the global marketplace. There is a
worldwide partner network providing support for Compiere in more than 20 countries.
Compiere is based around standard business processes, the software is written in Java and requires
an Oracle database. Jörg Janke, the project founder, lead developer and owner of a company which
provides services for Compiere, has been working as an Oracle consultant before, and this is the
reason, why Compiere is tightly integrated with proprietary functions of the oracle database. This
makes Compiere less attractive for small companies which do not want to deal with the complexity
of administrating an Oracle database. But there is hope for database independence and support of
open source databases for the next major release, scheduled for the end of 2003.
8.5.1.1 Strengths
• Zero license cost - While there is a general trend today away form orbitant prices for licenses,
ERP / SCM systems are the exception. The Microsoft Business solution121, derived form the
acquisition of Great Plain and Navision, which is targeting the same market as Compiere, comes
with license cost higher than 100,000 U$, depending on the selected functionality, features and
number of users.
8.5.1.2 Weaknesses
• No vertical modules - There do not exist different versions of Compiere modules targeting
different industries.
• No local support - Implementing and supporting ERP/ SCM solutions requires a lot of
knowhow and it is very unlikely that the next local IT services company is a Compiere expert.
120 More information about Compiere is available at http://compiere.org, accessed: Aug. 28, 2003
121 More information about Microsoft Business Solutions is available at
http://www.microsoft.com/BusinessSolutions/default.mspx, accessed: Aug. 28, 2003
72
8.5.1.3 Opportunities
• Open standards reduce integration costs - ERP / CRM always needs to be integrated into an
existing environment. While integration of closed source software usually is getting very
expensive, because support is only available from the vendor (vendor-lock-in) and often the
closed source software is not supporting open standards. Compiere is fully based on open
standards.
8.5.1.4 Threats
• No OSS history in ERP - OSS has been used successfully in software development and for any
kind of Internet software. ERP is related to neither of them. Therefore the barriers are higher to
convince management for selecting a relatively unknown product for a business critical
application. Outside the world of open source enthusiasts, not many people have ever heard of
Compiere.
73
8.6 Content Management Software (CMS): ez Publish
Norwegian independent software vendor ez Systems122 is successfully running a service business
around their highly popular open source and PHP-based content management system ez Publish.
With advanced functionality like integrated workflow engine, role based access control, discussion
forums, content versioning, multi language possibilities, search engine, template engine and more,
this software is competing against expensive proprietary CMS software.
8.6.1.1 Strengths
• Zero license cost – Proprietary CMS licenses with functionality comparable to ez Publish, are
usually very expensive (> 10K $)
• Fully Integrated OSS Stack - ez Publish uses the Apache Webserver, the MySql database and
the PHP script language as its core engine. Installation packages for Windows and Linux are
available, which contain everything to run the content management system.
8.6.1.2 Weaknesses
• Poor Documentation - Like many other OSS projects, even the ones supported by commercial
organizations, ez Publish does not offer much documentation for the novice user. While getting
up and running the preconfigured examples is very simple, any more
• No local support – There does not exist exist a globally organized partner network of companies
which provide support for ez Publish.
8.6.1.3 Opportunities
• Only few OSS CMS competitors – Today users can choose between more than 50 OSS content
management systems which are actively developed. But most of them are not comparable in
terms of features and workflow integration with ez Publish.
8.6.1.4 Threats
• Prices of proprietary CMS are dropping – As side effect of a market consolidation of CMS
122 More information about ez Systems and ez Publish is available at http://ez.no, accessed: Aug. 28, 2003
74
vendors, the consumer gets more features and a better product for less money than lets say a year
before. Compared to OSS, the commercial CMS vendors are developing their products faster and
better than the majority of the CMS OSS communities. Ez Publish has not reached yet the critical
mass to play a dominant role on the global CMS market.
75
9 OSS survey in Switzerland
To find out, to which extent companies and governmental organizations in Switzerland are already
using OSS, whether their IT strategies consider Integration of or Migration to OSS, how cost
analysis are performed and whether the cost of OSS solutions are taken properly into consideration,
more than 20 IT decision makers of small companies, large corporations and governmental
organizations have been interviewed.
The business approaches of OSS integration / migration among the interviewed parties was very
diversified. The interviews have been realized by phone, and the duration of one interview varied
between 10 minutes and 30 minutes. The results only show trends.
40
35
30
25
%
20
15
10
0
Yes, for some No Yes, for most Yes, but only No, but we
systems / ap- systems /ap- to collect have plans for
plications plications experiences OSS
At first sight, the popularity of OSS seems to be surprisingly high. However, the companies using
only OSS are all ISPs, with Linux powered web servers.
76
For which sytems / applications is OSS used ?
40
35
30
25
% 20
15
10
0
Webserver Mailserver DNS Software Linux Thin
Server develop- Clients
ment tools
Figure 6 OSS at Swiss companies and organizations
OSS is most used for Internet related services, systems and applications. Mailservers are the
preferred pilot projects to gain experiences with OSS.
% 25
20
15
10
5
0
No li- Higher Better No licens- Ideology
cense stability security ing pa- behind
cost perwork OSS
Figure 7 OSS advantages
Most of the interviewees who selected license costs savings as main advantage of OSS did not do an
IT cost analysis based on a model.
77
Main disadvantage of OSS
70
60
50
40
%
30
20
10
One interviewee was disappointed about the expected cost savings of only five percent, for planned
Linux migration. He had expected much more.
60
50
40
%
30
20
10
0
More Equal Less
Figure 9 OSS trends
Some interviewees claimed to have only migrated to OSS, because they were unsatisfied with
licensing and upgrade issues of proprietary software. Nobody regretted the move to OSS or to have
gained experiences with a OSS pilot project.
78
Are Linux clients planned for the dektop ?
90
80
70
60
50
%
40
30
20
10
0
No Yes Linux Thin Clients
already in use
Figure 10 Linux on the desktop
Linux on the desktop as client operating system seems to be far away. However, thin clients123 are a
viable option for some governmental organizations.
35
30
25
% 20
15
10
0
No strategy yet Internal training Self study External train-
ing
Figure 11 OSS knowhow
Smaller IT companies had wrong expectations about the required skill level to administrate Linux
servers. They made the experience that Windows NT knowledge does not help much to administrate
123 With thin clients, applications run on the server and display is redirected to client computers
79
a Linux Server.
% 50
40
30
20
10
0
No influence Influence
Figure 12 SCO / IBM lawsuit
40
35
30
25
%
20
15
10
0
Only marginal Significant costs No costs savings
costs savings savings
Figure 13 OSS and IT costs savings
Most of the interviewees, who claimed that the costs savings of the migrated systems were only
marginal (smaller 10%) had unrealistic expectations.
80
How are IT costs analyzed, which costs are considered ?
30
27.5
25
22.5
20
17.5
% 15
12.5
10
7.5
5
2.5
0
Only Hard- No cost Purely based Inhouse Simplified
ware and li- analaysis on experi- developed Gartner TCO
cence-costs ences model analysis
Figure 14 IT cost analysis
81
involved with migrant ions and changes. And a migration to OSS never is straight forward for
complex systems. Second, there are no real OSS alternatives for enterprise software like ERP,
expect of Compiere124 , which provide partial functionality of SAP, but which is more or less
unknown in the commercial wold expect of some special interest groups.
Because license and support cost for operating system an other basic software is only marginal
compared to cost of the enterprise software, many large enterprises are not considering migration to
OSS worth the effort in regard to cost savings.
However, OSS is moving into large corporations, but it is less visible, sometimes not even to the
CIO. And if large enterprises are moving to OSS, they often do not make this information public.,
this is particularly true for financial institutes. Beside of the visible move to OSS, which is very rare
yet at large companies in Switzerland, there exists something like a quite revolution. On the one
hand, more and more department intranet servers are running o Linux, on the other hand, Linux is
more and more embedded into devices like hardware for any kind of automatize or network and
communication equipment.
Another area where OSS has a strong position in Switzerland, is enterprise software development,
especially if the programming language is Java. There are man tools or frameworks which highly
popular among developers not only because of their excellent quality, but also because there is no
paperwork needed for ordering, budget and licensing issues.
82
Linux as operating system, but for the benefit of the user, this desktop clients are running a Citrix126
powered MS Windows desktop. There was never such a media hype involved, comparable to the
city of Munich move to Linux, but Switzerland has governmental organizations, with years of
experience in running large part of their systems on Linux.
The next big visible wave of OSS replacing proprietary software might be represented by
governmental organizations migrating to Linux.
126 Windows is installed on the CITRIX server, and just the display is redirected to the end user
83
10 Conclusions
The objective of this thesis was to identify possibilities and limits of OSS. The current trend is an
increase of the possibilities and a reduction of the limits. However, the limits will never be
eliminated, because proprietary software is better suited for niches, and a healthy software
ecosystem, where innovation is driven by competition, requires open source and proprietary
software.
• reduced or no vendor-lock-in
• more flexibility
• increased security
• better stability
• usability: user interfaces and documentation often only targeted to expert users
84
10.3 Possibilities for OSS
10.3.1 Governmental organizations
A drastic increase of OSS at governmental organizations can be expected for the next years.
Governmental organizations usually have the necessary size and resources to implement streamlined
and cost-optimized IT processes, which benefit on the long term from open source, therefore OSS
for governmental organizations can only be recommended.
10.3.3 ISP
For ISPs and other companies with strong relation to the Internet the use of Linux and other OSS is
recommended without any restrictions.
• Customer specific user interface, high quality documentation and commercial support from
traditional software engineering
Smaller companies often do not have the same possibilities for analyzing their costs. Their IT
processes are not streamlined, and everything related to IT depends on single person. Switching to
Linux just to save licensing costs does usually not make sense for small companies.
85
10.4.1 Linux for the desktop
In general, Linux on the desktop for the corporate user should be avoided, only if a company has
already several years of Linux experience or in other special cases, where user acceptance will not
be a problem, Linux for the desktop is recommended.
86
10.5 First Steps for migrating to OSS
The process of deciding whether or not to deploy OSS, then possibly planning for and adopting it
follows these high-level milestones:
2. Assess existing IT systems and functions to see where they meet objectives and where gaps exist.
3. Identify areas of business process complexity or high IT costs out of proportion to enterprise
value.
4. Look for appropriate opportunities for Open Source software and methodologies to help meet
enterprise IT objectives.
5. Estimate costs and projected returns of migrating specific platforms and functions to Open
Source.
6. Perform initial financial Total Cost of Ownership (TCO), Return On Investment (ROI) and
Business Value analysis to evaluate the feasibility and financial impact on enterprise IT budgets
from Open Source migrations.
7. Obtain executive management approval and sponsorship for planning Open Source initiatives.
8. Create an Open Source Steering Committee ideally composed of executives, IT stuff and
business representatives.
9. Identify key executives and technologists to lead open source efforts, and obtain training for
them in open source concepts, resources and methods.
Pilot projects
10.Plan a series of projects to integrate Open Source software and methods into the IT
infrastructure.
12.Develop an Open Source Engagement Plan including guidelines for using Open Source software
and engaging with the community.
87
13.Refine and confirm the preliminary TCO, ROI and Business Values analysis of financial returns
from OSS
14.Gain executive approval and sponsorship for conducting one or more OSS migrations.
Production projects
15.Execute OSS migrations and review the results in a progressive sequence of project cycles.
88
11 Bibliography
• Rosenberg, Donald K., Open Source: The Unauthorized White Papers. John Wiley & Sons,
2000. ISBN 0-7645-4660-0
• Sandred, Jan, Managing Open Source Projects: A Wiley Tech Brief. John Wiley & Sons, 2001.
ISBN 0- 471-40396-2
• Fink, Martin, The Business and Economics of Linux and Open Source. Prentice Hall PTR, 2002.
ISBN 0-13-04-7677-3
• Hahn, Robert William (Editor), Government Policy Toward Open Source Software. The
Brookings Institution, 2003. ISBN 0-8157-3393-3
• Afuah, Allan, Internet Business Models and Strategies. McGraw Hill, 2003.
ISBN 0-07-251166-4
• Raymond, Eric S., The Cathedral and the Bazaar, O'Reilly, Sebastopol, CA, 1999,
ISBN: 0596001088
• Nielsen, Jakob Usability Engineering. Boston, MA: Academic Press, 1993, ISBN: 0125184069
• Hawkins, Richard E, The Economics of Open Source Software for a Competitive Firm,
Pennsylvania State University, Dubois, November 2002.
• Lee, Samuel; Moisa, Nina; Weiss, Marco, Open Source as a Signalling Device - An Economic
Analysis, JEL Classification: D82, L14, L86, O3, Goethe-University, Frankfurt/Main, March
2003.
• Nichols, David M; Twidale, Michael B, Usability and Open Source Software, Working Paper
Series ISSN 1170-487X, The University of Waikato, December 2002.
• Bonaccorsi, Andrea; Rossia, Cristina, Why Open Source software can succeed, Laboratory of
Economics and Management, Sant’Anna School of Advanced Studies, Pisa, Italy, 2003.
89
• Commission of the European Communities, Linking up Europe, the importance of
interoperability for e-government services. Brussels, July 2003
• West, Joel, How Open is Open enough ? Melding Proprietary and Open Source Platform
Strategies, San Jose State University, 2002
• Fleury, Marc, Why I love Professional Open Source, JBoss Group LLC, 2003
• http://newsforge.com/newsforge/03/07/11/2056234.shtml
(Frederick Noronha: Open Asia: Japan and Korea embrace Open Source. Jul. 11, 2003)
• http://www.washingtonpost.com/wp-dyn/articles/A59197-2002Nov2.html
• http://seattletimes.nwsource.com/html/businesstechnology/134908115_ballmer05.html
• http://news.com.com/2008-1082-994819.html
• http://www.opengroup.org/tech/open-source/opengroup-os-strategy.htm
90
11.3.1 Expert Interviews:
IT managers from the following companies or organizations shared their experiences and opinions:
91