2013 IEEE International Conference on Business Informatics

Using ArchiMate to Represent ITIL Metamodel

Marco Vicente∗ , Nelson Gama∗† and Miguel Mira da Silva∗

∗ Instituto
Superior Tecnico, Av. Rovisco Pais, 1049-001 Lisboa, Portugal
† CINAV-PT Navy Research Center, Escola Naval, 2810-001 Almada, Portugal
Email: {marco.vicente, nelsongama, mms}@ist.utl.pt

Abstract—Enterprise Architecture (EA) and ITIL, two dis- and general EA principles, methods and models to the design
tinct governance approaches with different perspectives, have and realization of the remaining organizational structure.
become recently dominant between practitioners. However, In early work, we have already defined this architecture’s
parallel EA and ITIL projects can lead to wasted resources
and a duplication of costs and efforts. This paper proposes motivation model [7], and now we will focus on the architec-
an integration by approaching ITIL from an EA perspective ture representation, on modeling ITIL components according
and proposes a mapping of ITIL concepts to EA, and a set of to the EA approach and the ITIL principles, constraining
models representing the ITIL metamodel using the ArchiMate the organization’s freedom of design. The goal is to give
modeling language. Our goal is twofold: on one hand to give the the architect the elements and models to design specific
architect the elements, relationships and models that represent
best practices in IT service management, and, on the other, organizations according to best practices, which in this case
to formally model ITIL for knowledge sharing, stakeholder is ITIL and ITSM, but could as well be extended to other
communication and to contribute to ITIL discussion and domains, like, for instance, logistics or purchasing. This
validation. For evaluation we shall use interviews and the Wand approach may also be a tool to check for compliance and
and Weber ontological method. maturity levels in these particular organizations.
Keywords-ITIL, ArchiMate, Modeling, Enterprise Architec-
ture. IT Service Management II. R ELATED W ORK
There have been attempts to integrate EA and ITIL.
I. I NTRODUCTION Braun and Winter [8] proposed an EA expansion to ITIL v2
Enterprise Architecture (EA) is a coherent whole of and Service Oriented Architectures (SOA), having EA as a
principles, methods, and models used in the design and re- pivotal concept. EA provided an overview of the IT archi-
alization of an enterprise’s organizational structure, business tecture, while ITIL was assigned to management processes
processes, information systems, and infrastructure [1]. IT for services delivery [9]. Nabiollahi [10] provides a service
Service Management (ITSM) evolved naturally as services based framework for EA to meet the ITSM requirements
became underpinned in time by the developing technology of ITIL v3, suggesting an EA extension to involve service
[2]. ITIL [3] is the de facto standard for implementing architecture layer from ITIL Service Design [11].
ITSM [4]. It is a practical, no-nonsense approach to the Thorn [12] addresses the relation between ITIL and TO-
identification, planning, delivery and support of IT services GAF, regarding EA as a fundamental concept for organiza-
to the business [5]. It is a set of five books with the Service tional engineering, in which ITIL is included as a framework
Lifecycle processes that are described in the book in which to an operation model for IT delivered services. He argues
they find their key application [6]. that both frameworks can be used together by mapping them,
Unfortunately, having two different approaches to IT TOGAF covers the development of EA, and is involved in
governance can lead to several setbacks. In a time when the products conception lifecycle whereas ITIL ensures the
organizations strive to be efficient and effective, resources delivery and management of IT services to users [9], [12].
should not be wasted by having different organizational de- In the same note, Sante [13] argues that the recent versions
partments or teams handling both approaches independently. of ITIL and TOGAF keep converging to integration. In fact,
This paper is part of a wider effort to join both of these in ITIL v3 references are made to architectural concepts,
approaches by defining a specific EA for organizations that hitherto only found in publications on architecture. The
need to manage IT services. Enterprise architectures do not same, although to a much lesser extent, applies to TOGAF
focus on specific issues because their goal is to be able to 8: where references are made to IT management [13]. The
design and realize every organization. On the contrary, our author relates the five ITIL books to TOGAFs ADM cycle,
goal is to narrow it down, and restrict the architecture to showing there are indeed several similarities, but identifying
organizations that have the management of IT services as two main differences: a) developing business architecture is
an architectural driver. Therefore, the main motivation of our part of the TOGAF framework while the scope of ITIL is
work is to define an EA that uses ITIL principles, methods, limited to developing an effective and efficient IT depart-
processes and concepts to perform IT service management, ment, whilst developing business architecture is out of scope

978-0-7695-5072-5/13 $26.00 © 2013 IEEE 270

DOI 10.1109/CBI.2013.45
in ITIL; and b) running IT operations and delivering actual solution. In fact, Braun’s [8] and Thorn’s [12] work is limited
IT services are within the scope of ITIL, while TOGAF to ITIL v2 and Nabiollahi [10] proposes a new service
does not cover the development and maintenance of a run architecture layer, but does not clarify the architectures’
time environment, neither the way how services are actually relationship. As for Sante’s [13] work, although we agree
produced and delivered [13]. with his approach and conclusions, the result is not a unique
On the other hand, ITIL is a collection of five books body of knowledge, but two different frameworks linked by
with the best practices related to the effective and efficient a mapping.
management of IT [14]. These books illustrate concepts, Moreover, none of these approaches provide models or
the relationships between them, framework lifecycle, pro- a formal representation for the proposed solutions. In fact,
cesses, information management, information systems and what we are looking for is a holistic solution, following the
databases through visual representations. However, it is EA approach but using ITIL best practices. A body based
mainly in process modeling that we see a formal representa- on a set of principles, methods and models to underlie the
tion, with a known symbolic and semantic model. However, design and change of these organizations. And since we have
to describe other ITIL domains it seems to lack a common, already defined the motivations and principles that constraint
clear and formal notation and semantic. this architecture’s design [7], this paper will address the
architecture representation that realizes those motivations.
III. R ESEARCH P ROBLEM
IV. P ROPOSAL
Today, there is no fully complete framework to be used
as a comprehensive off-the-shelf solution to ensure the A. EA and ITIL Relationship
alignment between ITSM and the organization’s concepts We soon realized that it would be harder to integrate two
and artifacts. In fact, different frameworks are often used approaches if they did not speak exactly the same language,
and, most of the times, simultaneously too. Parallel EA and so we needed a uniform representation, a common frame
ITIL projects imply a duplication of investments and costs. of reference. Our intention was to find graphical languages
In effect, even with shared infrastructures we cannot avoid that best described each approach and map them according
a duplication of data repositories, procedures and human to similar concepts.
resources, being hard to define a way for teams not to For EA, Lankhorst [1] enumerates several languages.
compete together or maintain different efforts aligned [9]. There is IDEF, BPMN, ARIS and finally UML. However,
Although EA and ITIL describe areas of common interest, Lankhorst also identifies common issues among them, like
they do it from different perspectives. ITIL was developed poorly defined relations between domains, models not in-
to support Service Management and EA to support an tegrated, weak formal basis and lack of clearly defined
holistic organization view. However, since services have semantics, missing most of them the overall architecture
become part of fast-changing organizations, the prediction vision being confined to either business or application and
of what will be needed tomorrow is of growing interest technology domains. ArchiMate, on the other hand, provides
to the people that deliver them. Conversely, architecture a uniform representation for diagrams that describe EAs.
has changed from a rather static design discipline to an It offers an integrated architectural approach that describes
organization-encompassing one, and is only useful if the rest and visualizes the different architecture domains and their
of the organization is using it to enable alignment [13]. underlying relations and dependencies [17], so ArchiMate
There are several common benefits and components which stood out as the language we were after for modeling EA.
raise the issues of relationship and integration of EA and As for ITIL, it is often depicted as just a process and infor-
ITIL [10]. EA is regarded as a pivotal concept for organiza- mation architecture, lacking a formal representation on other
tional engineering, while ITSM is regarded as the dominant domains. Although we realize that most of ITIL content
operations model for IT [8]. EA guarantees consistency in is about describing best practice processes, we believe that
building new products or services and addresses business limiting ITIL to these only two domains is one of the factors
requirements, while ITSM guarantees the consistency of that turns its integration with EA so difficult.
services, through the use of standard processes [15]. In fact, as Sante [13] points out, the earliest versions
Accordingly, Radhakrishnan [16] identifies several ben- of ITIL hardly contained any references to architecture as
efits of collaboration, as organizational learning, avoiding a concept, method or framework. However, in ITIL v3
duplication of effort, re-use of documentation, cross training, references are made to architectural concepts, while showing
and planning and implementing the target EA and ITSM that the main structural difference is that ITIL does not
architectures with a coordinated and integrated method. change the organization’s own business processes while,
Therefore, all integration attempts that we presented ear- on the other hand, it runs IT operations and delivers IT
lier tried to answer a problem that should not be taken services. Additionally, Gama [9] also related the core EA
lightly. However, although they all seemed close to an artifacts and the EA five architecture layers to ITIL artifacts
integration, they do not propose a definitive and holistic and management processes, showing there is in fact a link

271
between ITIL and EA in all these domains and not only on Table I
M APPING CONCEPTS
the business and information ones.
Thus, we propose that like EA, we can also look at ITIL ITIL concept ArchiMate concept
as a composition of other architectures, namely business, in- Role, Department, Business Role
formation, application and infrastructure. Hence, on business Business unit, Service Owner
we have actors, roles, ITIL processes and functions, events; Channel, Interface Business Interface
on application the information systems, like the Configura- Collaboration Business collaboration
tion Management System (CMS); on infrastructure we have Process Business process
databases like the Configuration Management Databases Service Business service
(CMDBs); and on the information we have business objects, Application service
data objects and database artifacts. All these linked by a Function Business function
service oriented approach, where functionality is available Information Business Object
to the next layer in the form of services. Thus, if one looks Data Object
at ITIL from this point of view, we begin to realize that Artifact
by representing and splitting it across EA domains, we can Meaning
actually integrate them by integrating each of its layers. Representation
Hence, we propose that if an organization is represented Service Level Agreement Contract
by an EA, with all its layers, components and relationships, Operational Level Agreement
and if it has implemented ITIL, then ITIL components and Underpinning Contracts
relationships will be a subset (in every layer) of the EA ones. Business Service Product
Therefore, if ITIL can be regarded as part of EA, sharing Value Value
the same domains, components and relationships, and in the Software system Application component
absence of a formal ITIL graphical language we can then Information system
model the ITIL meta-model with EA elements, using the Application assets relationship Application collaboration
language we had already chosen for EA: ArchiMate. Application interaction
Software function Application function
B. ITIL Meta-model Using ArchiMate Concepts
Databases (CMDB, KEDB,...) System software
Before starting to model we needed to map ITIL con- Database access Infrastructure service
cepts in the ArchiMate metamodel, which has five generic Infrastructure interface
concepts: passive and active structure, behavior, service and Database function Infrastructure function
interface. Those are then instantiated on several others on Console, server, mainframe Node, Device
each of the three layers (business, application and infras- Network Network
tructure). It is these latter concepts that we will bridge with Communication path
the ITIL ones to show how closely they relate (Table I). Key performance indicators Value
It should also be noted that these concepts describe the Critical success factors
architecture of systems that support the enterprise, not cov- Metrics
ering the elements which motivate its design and operation Requirements Requirements
[17]. These correspond to the “Why” column of Zachman’s Product
framework [18], while the core concepts correspond to the ITIL relationship ArchiMate relationship
“How, What, Who, Where and When”. ArchiMate also has Has, uses, creates, Association
a Motivation extension which adds concepts such as goal, writes, reads and updates information Accesses
principle, and requirement [17]. Likewise, there is also a Used by Used by
yet to be released Valuation extension with concepts such Makes possible Implements Realization
as value, risk, constraint and resource [19]. Implements
In fact, we have already modeled ITIL’s motivation [7] Process/function ownership Assignment
and in future work we will also do it for its valuation. Flow Flow
Therefore, we will only show here the ITIL main con- Triggering Triggering
cepts and relationships. The mapping is based on ITIL and Element grouping (eg CMS and SKMS) Grouping
ArchiMate’s own definitions, and while it is often quite Process decision Junction
straightforward as both share many concepts’ semantics, Tool specialization Specialization
there are some exceptions which we will address now.
First, ArchiMate distinguishes between a business object
as a business concept; a data object as its logical repre-
sees it all as information. Second, we mapped KPIs and
sentation and an artifact as a physical piece of data. ITIL
CSFs as value, following Iacobs [19] proposal. Finally, we

272
 Figure 2. Detail of Service Operation model. Full model available in
Figure 1. Detail of ITIL overview model. Full model available in http:
http://db.tt/wmrivNof
//db.tt/TF0Ycdoh

have the Product concept, which in ArchiMate is seen as

a collection of services with a contract, while in ITIL it
often refers to something tangible. In fact, the ArchiMate
Product is closely related to ITIL’s Business Service. As for
the ITIL Product, ArchiMate is information-centric and is
not designed for physical products or logistics.
Moreover, we also want to distinguish two kinds of map-
pings: clear specializations (eg SLA for Contract; Database
access for Infrastructure service) and synonyms (Process
for Business Process; Role for Business Role). This also
demonstrates the relationship between ITIL and EA: while
the latter show equivalent concepts, the former show that
EA concepts are indeed generalizations of ITIL ones.
Using the above mapping, we shall present some models. Figure 3. Detail of Incident Management model. Full model available in
First we have an ITIL book overview (Figure 1). Its utility http://db.tt/7MlcqXvR
is to understand in a glance which are the services that
ITIL provides to its external environment (however, those
are not IT services, but business ones, since they represent In this latter model we also placed the requirements from
a general behavior and not its implementation). We can also our early ITIL motivation models [7] that are realized by
see the applications that ITIL uses to support its processes, each of the IM activities. This allows us to show how our
and also the infrastructure components that support those models can be used to achieve business/IT alignment.
applications. It provides a view with ITIL processes as a Also, another ArchiMate advantage is the existence of
black box system that provides services to the environment views that address a set of related concerns and are addressed
while using application and infrastructure ones. Actually, to a set of stakeholders [17]. Views are specified by means of
this model gives a meta Service Catalog: ITIL’s service list. viewpoints. In fact, these models were chosen to demonstrate
Next, we zoom into one of the ITIL books, the Service how ArchiMate can be used to show different ITIL views,
Operation one (Figure 2). Here we can see all its processes, directed to different stakeholders with different concerns.
events, functions, business objects. Finally, we aim for a Yet, the three models remain consistent, since the processes
deeper fine-grained representation and focus on the Incident inputs and outputs, business objects, business events and
Management (IM) process (Figure 3). This allows us to look business application and infrastructure services are the same
inside this process and see all its individual activities, which but on different granularity levels.
objects they access and what services they use and expose. However, the viewpoint we used was mainly the introduc-

273
tory, which provides an overview of the system of interest,
but other models could (and should!) be produced to repre-
sent other ArchiMate viewpoints. That way, different ITIL
viewpoints could be built to introduce and communicate
ITIL to the people that will play distinct roles.
Overall we produced a full set of models, for all the ITIL
books. Together, our ITIL core representation are the models
for the whole ITIL 26 processes and 4 functions.
V. E VALUATION Figure 4. Form answers
In this section we will evaluate our work following two
distinct approaches. First, we will analyze the proposed
mapping using a grammar evaluation method, and afterwards Node, Device and Business object. On these cases it could be
we will use interviews to evaluate our proposed models. a problem if we ever wanted to do the opposite process: go
from an ArchiMate ITIL model back to ITIL again. Here,
A. Mapping Analysis one should have to rely on the correctness of the object
According to the Wand and Weber method for the on- attributes to properly do the reverse mapping.
tological evaluation of grammars [20], we can compare
two languages’ mapping by identifying four ontological B. Models Evaluation
deficiencies: To assert the models’ utility and correction we used inter-
Incompleteness if it is not defined in total: if there are el- views since it allows asking questions that are open-ended
ements of the first language without a mapping; redundancy and explore emotions, experiences or feelings that cannot
if the mapping is ambiguous: if a first language element is easily be observed or described via pre-defined questionnaire
mapped in more than one on the second; excess if in the responses [22]. However, we also wanted to have some
second language exist any element without a mapping from quantitative data analysis, so, at the end of the interviews,
the first one; overload if two distinct elements of the first we also asked our guests to fill out a small survey regarding
language map onto the same element on the second one. our models.
Therefore, we can not say our mapping is complete, Therefore, we interviewed 13 specialists, from different
because there is not in ArchiMate the ITIL Product concept areas but all with a strong ITIL background, with dis-
(a tangible item). Moreover, on one hand, we left for future tinct occupations, and from diverse nationalities, including
work the motivation model and on the other, sometimes Phd students, university teachers, researchers, enterprise
the completeness stems from the fact that in some cases architects, managers and process owners at different sized
(KPIs, CSFs, and metrics) ArchiMate core has elements organizations.
generic enough to accommodate these, so our mapping Along the interviews, the same vision of ITIL as just a
does not reflect exactly the actual element meaning, but process architecture was very much present amongst the
its generic meaning. Therefore, any extension to specialize majority of our interviewees. In fact, when introduced to
and accurately represent these concepts would be much the suggestion that the ITIL books also mentioned another
welcomed. In fact, there is already a recent proposal [21] three dimensions that could be represented and modeled,
to represent KPIs in ArchiMate. our subjects would frequently turn skeptical and doubt our
As for redundancy sometimes there is more than one claim. However, when we finally showed them the models,
ArchiMate element to represent a ITIL concept. This hap- their opinions promptly changed, as they all agreed that this
pens in information, application relationship and database overall architecture vision would benefit ITIL implementa-
access, because ITIL is not much specific on application tion.
and infrastructure layers’ descriptions. We also find excess, The remainder of the interviews served to present our
as ArchiMate has concepts that are not defined on ITIL as motivation, explain our models, our mapping method, the
meaning or representation. One could argue that implicitly reasoning process behind it and gather ideas and suggestions
they actually exist with their ArchiMate definitions, where for further work. At the end of the interviews we asked the
meaning is “the knowledge or expertise present in a busi- subjects to fill a six question survey about our work. The
ness object” and representation “a perceptible form of the questions were: 1 - How do you classify the models’ cor-
information carried by a business object”, but the concepts rection?, 2 - How do you classify its utility for stakeholder
themselves are not mentioned in ITIL. communication?, 3 - Comparing with other ITIL graphic
Finally, we also have overload, when there are several models you know, how do you rate this one?, 4 - How
ITIL concepts to only one from ArchiMate, like Business do you classify its utility for ITIL validation?, 5 - How do
Role, Business Interface, Contract, Application component, you classify its utility for someone who is leading the ITIL

274
implementation on an organization? and 6 - If all ITIL books [5] V. Arraj, ITIL: The Basics White Paper. The Stationary
and processes were modelled this way, would you use it in Office, 2010.
your organization? [6] J. van Bon et al., Foundations of IT Service Management
The multiple choice answers had 5 levels and ranged from Based on ITIL v3. Van Haren Publishing, 2007.
None/Poor/No (1) to Very Useful/Very Good/Always (5). On
[7] M. Vicente, N. Gama, and M. M. da Silva, “Modeling itil
Figure 4 we present for each question its average rating. business motivation model in archimate,” in International
Conference on Exploring Service Science 1.3 (IESS), vol.
VI. C ONCLUSION LNBIP 143. Springer, 2013, p. 8699.
Along the years, several governance frameworks were [8] C. Braun and R. Winter, “Integration of it service man-
developed. Two of them, EA and ITIL, have grown to agement into enterprise architecture,” in ACM (ed.) ACM
be worldwide standards, having thousands of practitioners Symposium on Applied Computing, New York, 2007, pp.
today. However, having two distinct approaches results on 1215–1219.
duplication of investments, costs and wasted resources. [9] N. Gama, P. Sousa, and M. Mira da Silva, “Integrating
This paper contributes with ITIL’s 26 processes and 4 enterprise architecture and IT service management,” in 21st
functions modeled in ArchiMate. Our main motivation is to International Conference on Information Systems Develop-
ment, Prado, Italy, August 2012.
use these models to design ITSM organizations, and to check
for best practices’ compliance and maturity, by building as- [10] A. Nabiollahi, R. Alias, and S. Sahibuddin, “A service based
is models with the current organization’s ITIL processes and framework for integration of itil v3 and enterprise archi-
to-be models representing the ITIL maturity level where the tecture,” in 2010 International Symposium in Information
Technology (ITSim), vol. 1, Kuala Lumpur, 2010, pp. 1–5.
organization plans to stand in the near future. This would
allow to use EA methods (as TOGAF ADM) to perform [11] S. Taylor, V. Lloyd, and C. Rudd, ITIL: Service Design.
ITIL implementations like any other architecture change. Norwich, UK: TSO, 2007.
In times where cost and value generation are such im- [12] S. Thorn, “Togaf and itil. in: The open group (ed.),” vol.
portant drivers, IT governance, more than ever, should turn Catalog number W071, San Francisco, 2007, p. 26.
organizations more effective and efficient. EA does not tell
us how to design specific organizations that have IT service [13] T. V. Sante and J. Ermersj, “Togaf 9 and itil v3,” White Paper,
www.best- management-practice.com, 2009.
management as a main concern, and ITIL cannot help on
the overall organizational engineering. [14] N. Gama, P. Sousa, and M. M. da Silva, “Integrating en-
Therefore, our goal is twofold: on one hand to contribute terprise architecture and it service management,” in 21st In-
ternational Conference on Information Systems Development
with a formal definition of ITIL for knowledge sharing,
(ISD2012), Prado, Italy, 2012.
stakeholder communication and to help and promote ITIL
discussion and validation by the ITIL community itself; and, [15] A. Correia and F. Abreu, “Integrating it service management
on the other, to give to enterprise architects a set of principles within the enterprise architecture,” in 4th ICSEA, 2009, pp.
553 – 558.
and models, modeled according to the EA approach, to
restrict and guide the design of organizations according to [16] R. Radhakrishnan, “Enterprise architecture & it service man-
best practices in specific domains, which is in this case is agement - itsm frameworks and processes and their relation-
IT (ITSM/ITIL), but could as well be applied to any other ship to ea frameworks and processes.” The Open Group,
2008.
domains.
[17] The Open Group, Archimate 2.0 Specification. The Open
R EFERENCES Group, 2012.

[1] M. Lankhorst et al., Enterprise Architecture at Work. Berlin: [18] J. Zachman, “A framework for information systems architec-
Springer, 2009. ture,” IBM Systems Journal, vol. 26, pp. 276–292, 1987.
[19] M. E. Iacob et al., “Capturing business strategy and value
[2] The Stationery Office, The Official Introduction to the ITIL
in enterprise architecture to support portfolio valuation,” in
Service Lifecycle, 2007.
EDOC 2012, 2012.
[3] A. Hanna, J. Windebank, S. Adams, J. Sowerby, S. Rance, [20] Y. Wand and R. Weber, “On the ontological expressiveness
and A. Cartlidge, ITIL V3 Foundation Handbook. Norwich, of information systems analysis and design grammars,” Infor-
UK: The Stationary Office, 2008. mation Systems Journal, vol. 3, no. 4, pp. 217–237, 1993.

[4] A. Hochstein, R. Zarnekow, and W. Brenner, “Itil as common [21] A. Reis, Key Performance Indicators Representation in Archi-
practice reference model for it service management: formal Mate Framework. Instituto Superior Tecnico, May 2012.
assessment and implications for practice,” in 2005 IEEE
International Conference on eTechnology eCommerce and [22] B. Oates, Researching Information Systems and Computing.
eService, vol. 21, Nagoya, Japan, 2005, pp. 704–710. Sage Publications Ltd, 2006.

275