Content uploaded by Konstantin Pelz
Author content
All content in this area was uploaded by Konstantin Pelz on Aug 11, 2020
Content may be subject to copyright.
The Vendor Lock-In Effect of Software: A
Case Study about LiMux and Microsoft
Open Source and Intellectual Property in the Digital
Society
Chair of Innovation Economics
Technical University of Berlin
Winter Term 2019/20
Konstantin Pelz 386557
Maximilian Schneider 367765
Supervised by Mirko B¨ohm
Contents
List of Figures II
1 Introduction 1
2 Why Microsoft Consistently Wins 2
2.1 Vendor Lock-In Definition . . . . . . . . . . . . . . . . . . . . 3
2.2 AvoidingLock-In ......................... 3
3 Switching Costs 5
3.1 Switching Costs: Differences Between UNIX, Linux and Win-
dows................................ 5
3.2 Network Effect of Microsoft Windows . . . . . . . . . . . . . . 6
4 LiMux: The IT Evolution 7
4.1 The End of Windows NT Server 4.0 and The Begin of LiMux 7
4.2 The Technical Barrier: The File Format for Office Applications 11
4.3 The Human Factor: The Problem Sits Between The Desktop
and The Back of The Chair . . . . . . . . . . . . . . . . . . . 14
4.4 TheEndofLiMux ........................ 16
5 Moving Into The Cloud 17
6 Conclusion 19
References i
I
List of Figures
1 LiMux logo and wordmark . . . . . . . . . . . . . . . . . . . . 7
2 WollMux logo and wordmark . . . . . . . . . . . . . . . . . . 12
3 Office 365 logo c
Microsoft ................... 17
4 Logo of Google’s ”G Suite” cloud service c
Google . . . . . . 18
5 Open-Xchange Logo with claim c
Open-Xchange . . . . . . . 18
II
Abstract
The paper researches Vendor-Lock-Ins by the technology company
Microsoft. In particular Windows and Microsoft Office will be the
target of this paper. We will find out if Microsoft is actively putting
mechanisms into their products to bind the user to their products.
To get a view on the effects in the real world, the paper will take a
closer look on the LiMux project, which dealt with the migration from
Microsoft products to open source software in the city of Munich.
In the first section we will directly dive into Microsoft and their
lock-in mechanisms and what Microsoft has done in the last years in
this field. We will look how a vendor-lock-in can be fulfilled and how
the user can prevent to be locked-in.
In the next section we will analyze the costs if a locked-in user
wants to switch to another product. Which costs does he have to
expect. In detail we will analyze the costs of an operating system
switch.
After the section about Microsoft we will dive into LiMux. First
the decision to switch away from Microsoft products will be analyzed,
especially the costs for the different solutions for their system. Addi-
tionally, we will take a look on the possible barriers. On the technical
side, these are the file formats of Microsoft Office. On the human side
it is the acceptance of the new system by the user. Finally we will
find out why the project failed and the city of Munich switched back
to Microsoft products.
Finally we will give a preview of the situation in the next years
by researching lock-ins for office applications in the cloud as Software-
as-a-Service. Microsoft plays an active role in this section by offering
their cloud solution Office 365.
III
1 Introduction
In March 2003 Microsoft’s vice president Steve Ballmer interrupted his ski-
ing trip in Switzerland to visit Christian Ude, who was mayor of Munich
at the time, and Peter Hofmann, the project leader of LiMux, and offered
them discounts of millions of euros to stay with Windows and Office. Nine
years later Bill Gates himself took the opportunity to give Ude a ride in his
limousine to the airport just to find out why they did not want to use their
products. It becomes evident, that a migration of more than 14.000 com-
puters has such a big impact that even the founder and the vice president
of one of the largest software companies in the world feel the need to talk
to someone they wouldn’t usually have spoken to, in this case the mayor of
Munich.
But Ude was sure to migrate away from Windows. He told Gates the
reason: ”Please note, it’s about independence. We don’t want to be de-
pendent.” Gates replied: ”That’s nonsense, depend on whom?” Ude stated:
”Because you are already here: from you, of course!” This brought Gates to
the conclusion: ”It’s incomprehensible to me, it’s ideology.” [1]
But is it only ideology to migrate to open source software instead of
staying with the de facto standard? Or are there also financial reasons to
switch to Linux?
Software companies are actively putting lock-ins into their products to
keep customers away from alternatives. The government of the city of Munich
tried to switch from Microsoft Windows to Linux and from Microsoft Office
to LibreOffice but in the end they switched back. We will find out the reasons
for the switch, for the roll back and the role of Microsoft in this.
1
2 Why Microsoft Consistently Wins
Competition has always been though in personal computing. Microsoft has
always been one of the strongest players on the market since the development
of the first operating systems, in which Microsoft played probably the most
prominent role [2, p.118ff]. There is still more to their Superiority than just
quality of software or infallible business decisions. One aspect called ven-
dor lock-in has had and still has a significant impact on the prosper market
position of Microsoft and might be the central reason for their excellent eco-
nomic development. As discussed extensively in the media, vendor lock-in
in proprietary software is a very common way of securing market share and
bind customers to your enterprise. Many big tech companies are known for
using such instruments. For example: Apple, Google and as researched in
this paper, Microsoft. Even the European Commission has researched this
topic and had to decide whether Microsoft has used methods to prevent other
companies from developing operating systems and features that would ques-
tion Microsoft Windows superiority on the market [2, p.124f]. One of these
methods and probably the most prominent one was vendor lock-in. Microsoft
used many techniques to prevent companies from closing the gap and gain-
ing market share. One of these methods that helped achieving this lock-in
effect was quite simple: it was just the absence of information. For integral
parts of their operating system, like Active Directory, Microsoft did not de-
liver extensive documentation [2]. That was one way that helped prevent
other companies from developing tools for better migration from Windows
to another operating system or to let software developed for Windows run
on other operating systems like Solaris as well.
Governments, enterprises, private consumers and other stakeholders feel
the effects of this business practice. Governments are especially interested
from a macroeconomics perspective, since vendor lock-ins and the associated
effects can lead to a disequilibrium in the specific market.
And as shown in different publications [3] it can also inhibit the devel-
opment of strong, resilient technology and software which is fundamentally
important to our society now and even more in the future.
2
All these are reasons to bring light into the darkness, to investigate the
vendor lock-in and the resulting switching costs and to reflect the results by
means of the LiMux project.
2.1 Vendor Lock-In Definition
Vendor lock-in makes costumers dependent on a single vendor by creating
artificial obstacles that significantly increase the costs and effort needed to
switch to another vendor [4].
This is achieved through
1. incompatibility with the systems of other vendors
2. by using proprietary or closed sourced software and technology that is
not interoperable with those of other vendors
3. by licensing the software under exclusive conditions [5, 6] and thus by
creating a legal condition under which it is impossible for others to
reuse or get access to the source code or documentation
These factors lead to barriers that require substantial efforts to overcome
for customers. In summary vendor lock-in means being dependent on a
single provider and their technology and not being able to easily change or
migrate to another provider without facing substantial switching costs, legal
constraints or incompatibilities.
2.2 Avoiding Lock-In
In many areas, great attention is paid to avoiding lock-ins. Interoperability
and portability play a particularly important role here. Dependence on in-
dividual vendors can lead to a difficult situation for the customer in many
aspects, for example in price negotiations. Especially when it comes to buy-
ing upgrades in order to access files or get security updates or paying for a
monthly software subscription, you have a very weak negotiation position as
a customer: either you pay the price or you can stop operating your business.
3
Customers do not actively chose to become locked-in to one vendor. Most
of the times they are just unaware of the threat or sacrifice it for other impor-
tant key features like costs, software features et cetera. Another important
factor are non-functional requirements. Reliability, security and integrity
are important factors that need to be negotiated and can be very business
critical especially when you have to rely on a single supplier. So the more
you distribute these requirements or the more you use open standards and
open source software the less you are dependent on that single proprietary
provider that might be a very big risk for the whole organization.
There are also other ways of avoiding vendor lock-in. Screening the mar-
ket and deciding based on both entry and exit costs. Making sure that the
migration of data can be done without unforeseeable costs. Using software
supporting Open Standards and thus being vendor independent and interop-
erable with software from other suppliers. This seems to be one of the most
considered options. More and more organizations tend to use OSS and the
reputation of OSS is increasing as well [7].
Another aspect that needs consideration is the provision of OSS related
services like technical support. Actually everyone can offer services for OSS.
But some organizations only allow specific partners to offer such services.
That makes customer dependent especially when there is no competition on
a specific market. An example is Red Hat. They only allow official partners to
provide support for their software, trying to assure high-quality support [8].
So we can state, using Free Software is an is an excellent alternative because
it allows you to modify software on your own to enhance and tailor it to your
specific business use cases. It is not likely to experience vendor lock-in using
OSS. But you could be less free in your decisions as you might have initially
expected when switching to OSS.
In conclusion we can say that Free Software that uses Open Standards
helps reduce switching costs and provides vendor independent support as
well as significantly enhances flexibility which is fundamentally important in
a rapidly changing digital world. But as with almost everything, certain risks
must be taken into account and even in the area of OSS not all the glitter is
gold.
4
3 Switching Costs
Switching costs can be described as one time costs that occur when a cus-
tomer switches from one provider to another. They are directly related to
the switching process, but must not be incurred immediately upon switching.
Examples for switching costs are search costs, transaction costs, learn-
ing costs, customer habits or emotional costs.[PROOF: Consumer Switching
Costs: A Typology, Antecedents, and Consequences] These costs are often
not really calculated in when customers decide to buy software, but become
important when faced with a reason to consider switching.
Switching costs are an important part of the Total Costs of Owner-
ship(TCO). In short TCO covers the costs of deploying a product over its
whole life cycle [9]. Normally Software, Hardware and Training costs are
taken into account. Switching costs should be calculated as part of the
TCO. But the calculation is complex. Switching costs arise for example from
the migration of data, the purchase of software and licenses or the training
of personnel to new systems. Personnel costs increasingly include costs for
external consultants, as many processes related to personnel development are
lengthy, especially in the public sector. Switching costs are difficult to mea-
sure. However, the experience that employees have during such a transition
period is an integral part when calculating switching costs.
3.1 Switching Costs: Differences Between UNIX, Linux
and Windows
Studies showed that this is most notable when adopting Linux. While it
seemed to be no problem to migrate from UNIX to Linux, it is much more
difficult to switch from Microsoft Windows to Linux. One reason for that
discrepancy is that UNIX and Linux basically share the same foundation
and administration tools, whereas Windows is a completely different world.
Another reason is that Linux and the UNIX platform always tried to be as
open for other Open Source Software standards as possible. This makes the
migration much easier.
5
OSS is not always free. Sometimes the cost advantage over proprietary
software might be limited or even absent [8]. Therefore lower costs aren’t
necessarily a condition for the adoption of OSS or in this case Linux. We
still need to investigate further reasons that might be of interest.
3.2 Network Effect of Microsoft Windows
Another major factor worth considering when observing LiMux and related
scenarios is the networking effect [10]. This effect is created when a product
increases in value, the more people use it. A good example is the telephone.
The greater the number of users, the greater the value for each of them.
Network effect are notorious for creating vendor lock-in with the most
cited example being Microsoft products. Vendor lock-in can be mitigated
by opening the standards users depend on, thus allowing competition be-
tween different implementations. This is not sadly affecting industry wide
standards, as they are created by competition that decreased prices and in-
creased quality of the products thus leading to a superior solution which then
gets standardised and becomes the industry solution. But this is a lock-in
effect of another kind [11].
This effect is crucial to understand why LiMux failed and why the man-
agers in charge decided to roll back to Windows again.
6
4 LiMux: The IT Evolution
The term LiMux is a shortcut for Linuxin Munich.
Figure 1: LiMux logo and wordmark
The LiMux project initiated by the authorities of the Bavarian city of
Munich involved the switch from Windows software to Open Source Soft-
ware (OSS) for all their computers. This included server systems, operating
systems and desktop applications. All in all 14.000 computers had to be
migrated and the users had to be trained in the new environment. Nobody
managed an operating system switch of this size before. They were also the
first big administrative authority to use free and open source infrastructure.
In this section we will give an overview about the project, the reasons for
the switch, the technical barriers and the human factor. Additionally we will
research why the project failed in the end.
4.1 The End of Windows NT Server 4.0 and The Begin
of LiMux
In the end of 2004 the support for Windows NT Server 4.0 ended [12] and
customers were forced to upgrade to a newer version if they wanted to con-
tinue to profit from the support. The city of Munich operated more than
14.000 computers at the time and the upgrade would have cost them 7,5 mil-
lion euros, resulting in more than 500 euros per computer. Ude, the mayor of
Munich, was very upset that Microsoft used its monopoly, he commented Mi-
crosoft’s move with: ”Eat or die!” [1] That’s when the idea came up to look
for a cheaper alternative. All in all, the decision by Munich’s mayor Chris-
tian Ude can be summarized as an attempt to make a profitable decision and
free themselves of cost-intensive upgrades.
7
In the period between 2001 to 2003, Munich compared different solutions
for the time following the Windows NT Server 4.0 in form of a study [13].
A switch away from Microsoft this size had never been done before so they
had to figure out what was possible and, more importantly, what would be
effective. The study included alternatives for the client operating system
and the office application software, which was Microsoft Office 97 at the
time. They differentiated the technical feasibility and the economical and
qualitative-strategic consequences.
Concerning the technical feasibility the study analysed the systems and
applications to find out which ones needed to be migrated and which ones
could be kept. It was clear most of the computers needed a system change,
either from Windows 97 to Windows XP or to a Linux-based system. Al-
though the systems played are big role, the more important aspects were the
applications. In case of OSS office applications Munich would also have to
migrate the MS Office applications including the macros and templates. If
the city would stick to MS Office they could keep the old macros [13, p. 8].
To differentiate, IT experts of the study came up with five possible configu-
rations for the combination of the operating system and office applications.
The cheapest and easiest configuration was an upgrade from Windows 97
to Windows XP and an upgrade from MS Office 97 to MS Office XP [13,
p. 10]. The study estimated 34,2 million euros for this migration [13, p. 18].
The other ”Microsoft-friendly” configuration would also upgrade to Win-
dows XP as an operating system but would have switched to an OSS office
application like OpenOffice instead of MS Office XP [13, p. 11]. This config-
uration cost around five million euros more than the prior one, resulting in
39,7 million euros. This can be explained mostly because the macros would
need to be redeveloped, the new software would need to be migrated by
system administrators and the users would need training for the new office
applications [13, p. 18].
Additionally, three configurations with Linux as operating system and
OSS office applications were introduced. Although they technically work the
same way, they differ as follows: One way is to do a hard migrate but this
would mean a lot of technical work in a short period of time. Additionally
8
already purchased licenses for Microsoft products would not be used. That
is an economical loss. So the city thought of a Linux-based system where
these products could still be used.
One way is a Windows emulator on the clients computer. Another one
includes a server which offers Windows services over the terminal. Both
require extra work to implement this option but a clear advantage would be
that the migration could be done in a smooth way [13, p. 11]. In the end, all
three configurations would get rid of Microsoft products, the only difference
is the time frame in which the switch would be finished.
The hard migration would cost an estimated 45,8 million euros. It would
result in much higher cost than the Windows configurations. This results
mainly from the migration itself, migrating more than 14.000 computers from
one day to the another is a lot of work. The configuration with a Windows
Virtual Machine (VM) works a lot better. The migration can be done over
a longer period of time making it less cost-intensive. Another advantage are
the Microsoft licenses that have already been purchased and can be used
until they expire. The configuration with the terminal services for Windows
is the most expensive one, with an estimated cost of about fifty million euros.
The same work as in the prior configurations needs to be done but also some
servers for the Windows services have to be bought, set up and maintained.
That makes it very expensive [13, p. 19].
It is noticeable that all three Linux-based configurations need a lot of
training of personnel since they have all used Windows for the past years
and do not necessarily show affinity to technical improvements. Obviously
this costs a lot of money but also involves the risk that the users are not
accepting the new system. This point will be discussed later in more detail.
At this point of the study the combination of Windows XP and MS Office
XP seemed like the best option but the combination of Linux with OpenOffice
and a VM for Windows just followed shortly behind.
To not just get a monetary view on the configurations, the study applied a
value benefit analysis of different criteria for a qualitative-strategic view. All
the Linux-based systems had better scores than the Windows-based systems.
In detail, the compliance with laws and the impact on the IT organisation are
9
a big benefit for Linux. Contrarily, the negative impact on the personnel is a
big benefit for Windows [13, p. 21f]. In conclusion, the Linux-based system
should be preferred to the Windows-based system from a qualitative-strategic
point of view.
In the next chapter the study tried to find out which is the better configu-
ration. From the monetary view Microsoft is definitely the best and cheapest
solution. The Linux-based systems need a lot of training, which makes them
so expensive. The study estimated seventeen to twenty-six million euro only
for the instructional training [13, p. 26]. That is more than half of the costs
for a Linux-based system. The Windows-based systems do not need this
because the personnel had been using using their products for years. On
the other side they will have to pay license costs to Microsoft. The study
calculated that for a time frame of the next five years the costs for licenses
from Microsoft should be preferred to the costs for training and instructions.
But the study also declared that this cost could decrease over a time frame
of ten years or more [13, p. 26]. This means that the recommendation should
be used with care.
As noted earlier the Linux-based systems have a better score for the
qualitative-strategic view. The study took up the dependency on Microsoft,
especially the dependence of their software to Windows. Applications like
MS Office are so deeply integrated into Windows that it is very hard for users
to migrate away [13, p. 27f]. If this is a vendor lock-in and if this is done
on purpose will be discussed later in detail. Linux and OpenOffice do not
produce those dependencies.
To compare the economical view with the qualitative-strategic view the
capitalized value per score point was calculated. Results show that the Linux-
based system with a VM for Windows had the lowest costs per score point.
Windows XP with MS Office XP just follows shortly [13, p. 30].
The study recommended to migrate from Windows NT Server 4.0 to a
Linux based system, Microsoft Office 97 should be replaced by OSS office
application software such as OpenOffice. The clients computers should be
used with a Linux operating system and a Virtual Machine for Windows
to enable a smooth migration [13, p. 31]. This would give the users time to
10
familiarize themselves with the new environment which would overall increase
the acceptance of this environment. We guess that is the reason why it is
called an IT-evolution and not a IT-revolution.
In the year of 2003 plans were finished to switch to Linux and OpenOffice
and the project had its official kick off. In the year of 2004 preparations began
and in the year of 2005 the first computers were migrated. Over the next
ten years they migrated all clients computers and servers to Linux. From
the year 2009 to 2013 they also migrated all computers to an open source
office application, but instead of OpenOffice they changed their decision to
LibreOffice [14]. The biggest difference was not the migration of the office
application itself but the migration of the macros, templates and forms for
their daily work [15, p. 5f]. Therefore they developed WollMux as an open
source software. More details about WollMux and why it was so difficult to
migrate the macros, templates and forms will be given in the next section.
4.2 The Technical Barrier: The File Format for Office
Applications
The city of Munich is a public administration with many technical procedures
(in German ’Fachverfahren’). For these technical procedures they need very
specialised documents. To manage and automate those they used macros
and templates under Microsoft Office 97. But they could not take them
to OpenOffice or LibreOffice because of incompatibility. Back then the file
format for Word was the Word Binary File Format with the identifier .doc.
It was not human-readable and the first documentation of this format was
released by Microsoft in the year of 2008 [16, p. 2]. In the years before it was
not possible for other companies and organisations than Microsoft to build
a software which could parse .doc files. The same applies for other office
applications by Microsoft like Excel and PowerPoint. Microsoft forced users
to stay with their products once they started to save their files in Microsoft’s
file formats. At a certain point the network effect sets in. In the year of
2000 Microsoft Office had a market share of 96 % [17]. This makes it almost
impossible for a user who wants to use alternatives to interact with most
11
of the other users because they use the file format from Microsoft which
is incompatible to other office applications. Of course this did not come
out of nowhere, back in those times Windows and Microsoft Office were the
industrial standard because there were good products. Today people see it
differently.
This illustrates the situation of Munich. They were using Microsoft Office
and their files were saved in the file formats by Microsoft. Then they wanted
to migrate to OpenOffice and later LibreOffice. Back in the year of 2004 the
city of Munich had over 21.000 templates and 900 macros in use [14]. And
all these files had to be migrated. But they could not just import them into
OpenOffice or LibreOffice. This forced them to rewrite all of them. They
used this chance to reorganise all their templates and macros to reduce the
amount of them. For the management of these the city of Munich developed
their own open source software in lack of a public available product. They
named their application ”Eierlegende WollMux” which is a term in German
to describe something that symbolically is able to do everything. With this
application they can manage, verify, test and document their macros, forms
and templates. It is licensed under the European Union Public License 1.1
(EUPL), the EUPL is a free and open source software license created by
the European Commission [18]. It is accessible in most of the European
languages and compatible to European Union law.
Figure 2: WollMux logo and wordmark
In the end they had 9000 templates and 800 macros less than before, most
of the forms are now web based [14]. And, most important, there were all
saved in the Open Document Format for Office Applications (ODF) which
became an Open Standard in 2006 [19, p. 1].
This was an important part of the migration with expected costs of
1.376.468 euros [13, p. 18]. In August 2012 a calculation by the city of Mu-
nich came to the conclusion that the disuse of Microsoft Office already saved
12
them 4.200.000 euros [14] which is far more than the costs of the migration
of the macros, templates and forms. They saved money and are independent
from their office application suite. On the other hand most of their partners
and externals were using Microsoft Office with the restrictive file format. It
got more complicated to share files with each other.
In the last years the situation got better. Other governmental institutions
have shown proof that is is possible to not use Microsoft Office and their file
formats. The Italian military services are currently migrating to LibreOffice.
Their plan is to migrate the last systems by the year of 2020. One of the
reasons for LibreOffice was the ”the open standard ODF format, [...] [which]
guarantee interoperability, legibility of digital documents over time, as well
as making the Administration independent of supplier and software”, says
General Camillo Sileo [20].
Despite all rumours, even the government of the United Kingdom is a
good example for open standards. In the year of 2013 they started to use
only open standards for their technology. The then Cabinet Office Minister
Francis Maude sad: ”Open standards will give us interoperable software,
information and data in government and will reduce costs by encouraging
competition, avoiding lock-in to suppliers or products and providing more
efficient services.” [21]
Another example is the French Gendarmerie. In the year of 2002 they es-
tablished a strict open standard policy for themselves. The first results could
be seen in 2005, they migrated from Microsoft Office to OpenOffice. Later
they also migrated from Internet Explorer and Microsoft Outlook to Mozilla
Firefox and Mozilla Thunderbird, the technical architect Pascal Denke de-
scribes this as the ”next step towards the freedom of our computer system”.
Finally they decided to migrate the operating system itself. Starting in the
year of 2008 they migrated the first computers to a Linux-based OS [22].
There are many more examples, these were just the big ones. For all of
them the most important reason is the independence from single products
and the interoperability to other products in the future.
Another important key for the success of open source office applications
was a decision by the European Commission in the year of 2004. The Euro-
13
pean Commission found that the interoperability of the Microsoft file formats
is an abuse of their powerful position [2, p. 295] (as mentioned earlier they had
a market share of 96 % around that time). They put a fine on Microsoft about
nearly 500 million euros. After this Microsoft agreed to standardise their file
formats. In the year of 2006 their Office Open XML file format was stan-
dardised by the European Computer Manufacturers Association (ECMA)
International [23].
From then on other software developers were able to implement new fea-
tures to read write files in the Office Open XML file format and make their
software more interoperable. But the specification was more than 6000 pages
long (the specification for ODF had less than 1000 pages) [24] and the file
format was very complicated. It was tough work for developers to make it
possible to import such files. So by now interoperability is given on paper
but it is not easy to implement it.
The next big step for LiMux was to make the new environment well-
known to the employees. The problems with this and more details are given
in the next section.
4.3 The Human Factor: The Problem Sits Between
The Desktop and The Back of The Chair
Dieter Reiter, who became the mayor of the city of Munich in 2014, sad
in a meeting with other members of the city council: ”All the crutches [of
LiMux] we have created here are simply not what I have known for 10 or
15 years from my private business dealings and therefore I was personally of
the opinion that a change [to Microsoft products] makes sense.” [25, min. 19]
This shows the dilemma of many users: Allegedly Linux and LibreOffice
or OpenOffice can not deliver the same functionality as Microsoft products.
It have to be noticed that Reiter helped with the movement of Microsoft’s
German headquarter to Munich and he outed himself as a fan of Microsoft
products [26]. His statement must be taken with caution.
A major point for the success of a new system is the acceptance by the
user. Based on the Technology Acceptance Model (TMA) by Davis [27,
14
p. 985-989] the user needs to perceive the ease of use and the usefulness
of the new system. If a higher position dictates the user the new system
it is important to show them the reasons for this decision. They need to
understand that this system will be useful for their daily work. To fulfill
this, training and time is required. A user will not see more usefulness and
ease of use in the first moment. He is used to the old system and he know
how to handle a workflow there. In the new system everything is unknown
in the beginning.
The people behind LiMux were aware of this. Back in the beginnings
of LiMux, Windows was the primary operating system. The typical user
in a governmental institution probably never heard of Linux before. That
is one of the reasons they have choosen the soft migration and why they
estimated more than twenty million euros for training [13, p. 18] which is
more than half of the overall calculated costs. They used the money to
train the administrators on who to deploy software and similar things. The
employees where trained on who to use their daily programs. Besides classical
trainings they implemented the ”LiMux Lernwelt” which was an e-learning
platform. Instead of teaching a software as one big component, the software
was split into sub modules and the users could decide which ones they want
to learn. This made the teaching process more effective and the users also
could redo the trainings if it was necessary. It was so successful they won
the eureleA Eurpean e-learning award [14].
Not everything went well. The expected problems have occurred. For ex-
ample the theatre of Munich (M¨unchener Kammerspiele) complained about
Linux and OpenOffice. They had problems with the document exchanges
with partners outside of Munich because of incompatibility of the files. They
also had problems maintaining their website because the browsers were too
old to display the site properly [28, p. 11f]. The department of labour and
economics and the department of health and environment also had issues with
the exchange of files because of the incompatibility [28, p. 14f]. On the other
side many users liked the features of WollMux. It made it easier to quickly
create and deploy qualitative macros, forms and templates [28, p. 26]. Un-
fortunately nobody did a study about the user acceptance of LiMux. These
15
were just some voices but there is no general picture available. We can say
that LiMux must be usable otherwise the governmental institutions could
not work. There are users who can handle the new environment better and
there are users who can not. But all in all it seems to work, no information
on the contrary could be found.
4.4 The End of LiMux
In the begin of the year 2017 the city council decided to migrate back to a
Windows-based-client. One of the reasons given is the incompatibility of the
current products. A stated goal is the highest possible compatibility. Another
stated goal is the transfer of local applications into web based applications.
This is based on the desire for independence of the operating system [29].
This might seem astonishing because in 2012 already 12.000 systems were
migrated to Linux, that were 80 % of the computers. In late 2013 all com-
puters were migrated to Linux and Libre Office [14]. Shortly before Peter
Hofmann, the project leader of LiMux declared the project as successfully
finished and recommended to accept the project [28, 30]. The system was
runnning and working.
Remigrating back to Windows only four years later seems ridiculous. The
costs for the LiMux projects (PROOF) were not much higher than initially
calculated. The system was running properly. The only reason they provide
is the incompatibility, that sometimes occurred but is getting better with
open standards. If the removal of the incompatibility is actually worth the
estimated costs of tens of million euros (PROOF) is not clear. It seems more
like a political decision.
Moving back to a closed and proprietary system is the opposite of what
many public institutions are doing as the examples from Italy, France and
the United Kingdom have shown in the section about the file format.
16
5 Moving Into The Cloud
LiMux started in 2003 and ended in 2013. Since then a lot has changed. Less
software is shipped as CD, a user just downloads it from the web. Today more
and more applications moved into cloud environments. Users can access these
Software-as-a-Services via their web browser regardless of their operating
system. Microsoft jumped on the train and launched in the year of 2011
Office 365 [31] which is the equivalent to Microsoft Office, just in the cloud.
How is the situation in the cloud where the user does not have the files stored
locally on his device? What if the user want to switch from his local office
suite to a cloud based office suite? What if he wants to switch from one cloud
office suite to another one?
Figure 3: Office 365 logo c
Microsoft
In fourth quarter of the financial year of 2017 Office 365 revenue surpassed
revenue from Microsoft Office for the first time [32]. This demonstrates
that cloud services are winning against standard software distribution and
licenses. But this is only the beginning. Most areas of Microsoft Cloud
Services were able to record significant increases in revenue. Outperforming
all other major business units. One major aspect of this positive development
is that organizations that had relied on MS Office products in the past do
now have a very easy time migrating all their data and business critical
additional software to Office 365. Of course Microsoft cared a lot about
compatibility with their previous products and Office 365. Another aspect
that increases the growth of Office 365 is the fact that Microsoft stops support
for older versions of Office and recommends switching to the cloud. All these
developments show again that Microsoft still benefits a lot from their vendor-
lock in and network effects. Even now with their switch to the cloud and
more subscription based business models it is difficult to deny the impact it
has on their operational results. Since the successful project phase during
which Microsoft won the tender for the JEDI platform[33], Microsoft is also
17
on the way to becoming the market leader in general cloud computing. The
company is gaining momentum in many cloud-related business areas. As the
US alone is predicted to spend around 120 billion dollar on cloud services
this year this is one if not the most important market for Microsoft [34].
Microsoft is not alone on the cloud office suite market. Google launched
G Suite in the year of 2006 [35].
Figure 4: Logo of Google’s ”G Suite” cloud service c
Google
Besides those two proprietary solutions there is an open source solution
called Open-Xchange App Suite, the backend is licensed under the GNU Gen-
eral Public License version 2 and their frontend under the Creative Commons
Attribution-NonCommercial-ShareAlike 2.5 License [36].
Figure 5: Open-Xchange Logo with claim c
Open-Xchange
Unfortunately each of these services stands for its own. A G Suite user can
not access shared documents from a Office 365 user and vica versa. Same
applies to Open-Xchange. All three offer a migration service but only for
email [37, 38, 39]. If a user wants to migrate to another office suite he can
not take his documents to the other service. The user has to export each
document manually in a compatible format and import it into the new suite.
If the user wants to migrate from Microsoft Office or LibreOffice to a cloud
solution, he also have to import each file individually.
We can conclude the situation for switches between the solutions is not
getting better, it is getting worse. Earlier it was only necessary to have the
files stored in compatible format and the software installed on the device.
Now it is necessary to first export all the data, store it locally and import it
to the new cloud application. Vendor lock-in may be not the right term for
it, cloud lock-in or supplier lock-in is better.
18
6 Conclusion
It can be concluded that Microsoft is using several techniques to lock-in
users to their products. Microsoft provides only little documentation for their
products and tries to keep a low profile in order to attract little attention from
the European Commission. Microsoft has learned from incidents like 2004.
From then on they changed their approach and started documenting their
file formats and API’s. But the documentation remained very complicated.
Today OpenOffice, LibreOffice and others are able to read and write files in
Microsoft formats but it was a long and hard way.
Microsoft still benefits from the network effect. The normal citizen uses
Windows on his personal computer and is not used to other operating sys-
tems. And because most application software is written for Windows and
not for Linux, it is difficult for developers of Linux to sustain their position
in the market. Microsoft is working hard to maintain this status quo.
LiMux initially started because Microsoft increased prices for their prod-
ucts especially for their Windows and Office licenses and for the support as
well, which is still a current problem [40]. To get independent from Microsoft
the City Council of Munich initiated the migration to a Linux-based client
and LibreOffice. The switching costs for this project have been considered
significantly higher than to continue the prevailing situation. Nevertheless,
independence from Microsoft was rated as a higher good. They expected
the most freedom and independence with open source software. And It took
them ten years to accomplish the migration of all computers to the new
operating system.
However, it was by no means easy to get to this point. Several problems
occurred. There have been problems with the file formats from Microsoft,
which forced them to rewrite all documents. This also had a positive side
effect. It enabled them to reorganise all their documents. Furthermore they
used the situation to develop a centralized system which is managing all
the macros and templates for each department. The results were a massive
reduction of the amount of documents and a new tool called WollMux for
the management which is now used by several other institutions outside of
19
Munich as well. Another problem they faced had to do with their employees.
A lot of training was necessary to introduce them to the new software and to
convince them how useful and easy to use Linux is. This difficulty was also
mastered by the responsible team. Of course, among thousands of employees
there are always some who have had problems. As in almost every situation
that involves major changes. But the number of this group could be neglected
on the horizon of this great change.
It can be concluded that the switch from Microsoft to Linux allowed the
Council of the City of Munich to emancipate from the negative consequences
that mainly occurred due to the lock-in effect. It can also be noted that the
city was no longer in a state of dependence after it had completed the switch.
The new mayor, however, seems to have different ideals according to
which he makes his decisions. His proximity to Microsoft is hard to deny.
It is hard to understand why they have chosen to migrate back to Microsoft
and its products and services only a few years later. This makes Linux and
open source software in general appear in a bad light.
The people in charge of the strategic orientation are integral in the deci-
sion making process. Unfortunately this often has nothing to do with techni-
cal arguments. The previous generation thought much of Linux and the new
independence from software companies. The new generation did not. In any
case, the open source software is not to blame.
20
References
[1] Jan Kleinert. M¨unchens Ex-OB Christian Ude im Interview. Linux-
Magazin, 10, October 2019.
[2] Commission Decision of 24.03.2004 relating to a proceeding under Arti-
cle 82 of the EC Treaty(CaseCOMP/C-3/37.792 Microsoft), April 2004.
[3] Justice Opara-Martins, Reza Sahandi, and Feng Tian. Critical review
of vendor lock-in and its impact on adoption of cloud computing. IEEE,
November 2014.
[4] Eric Albers. Modernising public infrastructure with free software.
”https://download.fsfe.org/campaigns/pmpc/PMPC-Modernising-
with-Free-Software.pdf”, 2019.
[5] Microsoft. Microsoft software license terms of-
fice 2019 desktop. ”https://www.microsoft.com/en-
us/Useterms/Retail/Office/2019HomeandBusiness/
Useterms Retail Office 2019HomeandBusiness English.htm”,
2018.
[6] Microsoft. Supplement to microsoft services agreement mi-
crosoft office 365 consumer subscription service and software.
”https://www.microsoft.com/en-us/Useterms/Retail/Office365/
Home/Useterms Retail Office365 Home English.htm”, 2018.
[7] Linux adoption trends: A survey of enterprise end users, 2010.
[8] K. Ven, J. Verelst, and H. Mannaert. Should you adopt open source
software? IEEE Software, 25(3):54–59, May 2008.
[9] Mogens K¨uhn Pedersen Michael Holm Larsen, Jesper Holck. The
challenges of open source software in it adoption: Enterprise architec-
ture versus total cost of ownership. ”https://openarchive.cbs.dk/
bitstream/handle/10398/6490/wp%2011-2004.pdf?sequence=1,
2004.
i
[10] (2006) network effect theory. in: The network(ed) economy. duv. https:
//doi.org/10.1007/978-3-8350-9213-6 3.
[11] Robert M. Grant. Contemporary Strategy Analysis. John Wiley & Sons,
2016.
[12] Sean Michael Kerner. Windows NT 4.0 Support Ends To-
morrow. https://www.serverwatch.com/news/article.php/3453131/
Windows-NT-40-Support-Ends-Tomorrow.htm, December 2004. [On-
line, accessed 2019-12-13].
[13] Unilog Integra Unternehmensberatung GmbH. Projekt - Client Studie
der Landeshauptstadt M¨unchen - Kurzfassung des Abschlussbereichts
inklusive Nachtrag. https://web.archive.org/web/20060630023127/
http://www.muenchen.info/pia/clientstudie kurz.pdf, June 2003.
[Online, accessed 2019-12-13].
[14] Markus Feilner. LiMux - the IT evolution - An open source success story
like never before. https://joinup.ec.europa.eu/collection/open-
source-observatory-osor/document/limux-it-evolution-open-
source-success-story-never, December 2013. [Online, accessed
2019-12-13].
[15] Mario Silic and Andreas K. Back. Open Source Software Adoption:
Lessons from Linux in Munich. IT Professional, 19(1):42–47, 2017.
[16] Microsoft Corporation. [MS-DOC]: Word (.doc) Binary File Format.
Technical report, Microsoft Corporation, November 2019. [Online, ac-
cessed 2020-01-05].
[17] Harry McCracken. A Peek at Office Upgrade. https:
//web.archive.org/web/20050907045510/http://www.pcworld.com/
news/article/0%2Caid%2C18462%2Cpg%2C1%2C00.asp, September
2000. [Online, accessed 2020-01-06].
ii
[18] European union public license. https://joinup.ec.europa.eu/sites/
default/files/custom-page/attachment/eupl1.1.-licence-
en 0.pdf, 2007.
[19] Patrick Durusau and Michael Brauer. Open Document Format for Of-
ficeApplications (OpenDocument) v1.0(Second Edition). Technical re-
port, OASIS Open, July 2006.
[20] La Difesa risparmia milioni di euro scegliendo software libero: parla
il generale Camillo Sileo. https://www.techeconomy2030.it/2016/
04/28/difesa-risparmia-milioni-euro-con-software-libero-
parla-generale-sileo/, April 2016. [Online, accessed 2020-01-09].
[21] Cabinet Office and The Rt Hon Lord Maude of Horsham. First
open standards for government technology endorsed: The Open Stan-
dards Board has endorsed the first set of open standards for gov-
ernment technology. https://www.gov.uk/government/news/first-
open-standards-for-government-technology-endorsed, September
2013. [Online, accessed 2020-01-09].
[22] Gregor Bierhals. Towards the freedom of the operating system: the
French Gendarmerie goes for Ubuntu. https://joinup.ec.europa.eu/
collection/open-source-observatory-osor/document/towards-
freedom-operating-system-french-gendarmerie-goes-ubuntu,
February 2009. [Online, accessed 2020-01-09].
[23] Jan van den Beld. Ecma International approves Office Open XML
standard: The new open standard safeguards the continued use
of billions of existing documents and promotes document process-
ing interoperability. https://www.ecma-international.org/news/
PressReleases/PR TC45 Dec2006.htm, December 2006. [Online, ac-
cessed 2020-01-09].
[24] Google’s Position on OOXML as a Proposed ISO Standard. https://
www.csun.edu/~hcmth008/odf/google ooxml.pdf, February 2008. [On-
line, accessed 2020-01-09].
iii
[25] Harald Schumann and ´
Arp´ad Bondy. Das Microsoft Dilemma - Europa
als Softwarekolonie, February 2018. [documentary film].
[26] Markus Feilner. ”Microsoft-Fan“: M¨unchens neuer OB Re-
iter will in Sachen Limux ”neue L¨osung finden“. https:
//www.linux-magazin.de/news/microsoft-fan-muenchens-neuer-
ob-reiter-will-in-sachen-limux-neue-loesung-finden/, 2014.
[Online, accessed 2020-01-13].
[27] Fred D. Davis, Richard P. Bagozzi, and Paul R. Warshaw. User accep-
tance of computer technology: A comparison of two theoretical models,
August 1989.
[28] Peter Hofmann. Abschlussbericht: Projektbezeichnung LiMux it
M@ITV, September 2013.
[29] Stefan Hauf. Stadtrat beschließt Neuorganisation der st¨adtischen IT.
Rathaus Umschau, 032:4–5, February 2017.
[30] IABG. LiMux Abschlussbericht des externen Projektcontrollings,
September 2013.
[31] Mary Jo Foley. Microsoft launches Office 365: Here’s what you need
to know. https://www.zdnet.com/article/microsoft-launches-
office-365-heres-what-you-need-to-know/, June 2011. [Online, ac-
cessed 2020-01-16].
[32] Peter Bright. Microsoft 4q17: Office 365 revenue surpasses traditional
licenses. https://arstechnica.com/information-technology/
2017/07/microsoft-4q17-office-365-revenue-surpasses-
traditional-licenses/, 2017.
[33] Tom Simonite. Microsoft is the surprise winner of a $10b
pentagon contract. ”https://www.wired.com/story/microsoft-
surprise-winner-dollar10b-pentagon-contract/”, 2019.
iv
[34] Sergei Klebnikov. Microsoft is winning the ‘cloud war’ against amazon:
Report. https://www.forbes.com/sites/sergeiklebnikov/2020/
01/07/microsoft-is-winning-the-cloud-war-against-amazon-
report/#54df1e613bec, 2020.
[35] Courtney Hohne. Google Launches Hosted Communications Services:
New Google Apps for Your Domain Available Now for Easy Email,
Calendaring, and More. https://googlepress.blogspot.com/2006/
08/google-launches-hosted-communications 28.html. [Online, ac-
cessed 2020-01-16].
[36] Open-Xchange. Licenses for Open-Xchange App Suite. https://
www.open-xchange.com/licenses-ox-app-suite/. [Online, accessed
2020-01-16].
[37] Google. Migrate your organization’s data to G Suite. https:
//support.google.com/a/answer/6251069?hl=en. [Online, accessed
2020-01-16].
[38] Microsoft. Migrate email and contacts to Office 365. https:
//docs.microsoft.com/en-us/office365/admin/setup/migrate-
email-and-contacts-admin?view=o365-worldwide. [Online, ac-
cessed 2020-01-16].
[39] Open-Xchange. A Hassle-Free Move to OX Solutions Migration. https:
//www.open-xchange.com/services/ox-services/migration/. [On-
line, accessed 2020-01-16].
[40] Christof Kerkmann and Dietmar Neuerer. Bundesregierung zahlt
fast eine Million Euro f¨ur veraltetes Microsoft-Betriebssystem.
https://www.handelsblatt.com/politik/deutschland/windows-
7-bundesregierung-zahlt-fast-eine-million-euro-fuer-
veraltetes-microsoft-betriebssystem/25452158.html, January
2020. [Online, accessed 2020-01-22].
v