Application of a Lightweight Enterprise Architecture Elicitation
Technique Using a Case Study Approach
Nicholas Rosasco and Josh Dehlinger
Department of Computer and Information Sciences, Towson University, Towson, MD, U.S.A.
Keywords: Enterprise Architecture, Case Study, Grounded Theory Method, Experiential Software Engineering.
Abstract: Enterprise architecture (EA) has demonstrated utility for improving overall Information System
(IS)/Information Technology (IT) outcomes for institutions, particularly those with large-scale or
integration-related needs. To achieve the core goal of an EA – integration, alignment and governance
between enterprise goals and the enterprise IS/IT portfolio – institutional vision, mission and objectives
must be elicited, analysed, understood and documented by the enterprise architects. This paper proposes and
evaluates a lightweight EA elicitation technique to gather the required information about enterprise mission
and objectives as a lightweight entry point to developing an EA. Specifically, we investigate the evaluation
results of our proposed technique in EA data elicitation and analysis stemming from a case study that was
conducted on an institution with a significant IS/IT asset portfolio. Our proposed EA elicitation technique
utilizes the VMOST elicitation question, a structured elicitation vehicle, and Grounded Theory Method as
the qualitative analysis technique to analyse elicited responses. Application of this approach in a real case
study garnered sufficient understanding the vision, mission and objectives of an enterprise to articulate
objectives in a way suitable to use as institutional goal as a part of the Zachman EA framework.
1 INTRODUCTION
An enterprise architecture (EA) seeks to enable the
creation of “organizational structure, business
processes, information systems and infrastructure”
from a “coherent whole of principles, methods, and
models” (Lankhorst, 2005). To aid in integrating,
aligning and governing an enterprise’s institutional
goals and its information systems (IS), a number of
EA frameworks (EAFs) have been created, including
the Zachman Framework (Zachman, 1987), The
Open Group’s Architecture Framework (TOGAF)
(Open Group, 2009) and the European Space
Agency Architectural Framework (ESAAF) (Gianni
et al., 2013). To support these ends, EAFs have
evolved to help align an enterprise’s vision, mission
and goals with the contents of their IS asset
portfolio. The effectiveness of this alignment is a
determinant for the overall utility and benefit of the
utilization of an EA and any misalignment is best
caught early in the crafting of an EA (Minoli, 2008).
The concept of an IS portfolio reflects the reality
faced by many enterprises which have, over time,
accumulated a variety of IS solutions to support their
business operations and architecture. By
encouraging a holistic, abstract view of both
enterprise goals and motivations, EA approaches can
provide critical context and scoping for IS needs and
decision-making. The immediate issue that occurs
for an institution when considering EA is the
approach, tools and techniques to use, as this is left
an open question by the various frameworks (Open
Group, 2009); (Zachman, 1987).
The body of existing research literature available
for the application of EA is largely focused on, and
shaped by, considerations derived from the original
enterprises that intended to use EA - namely very
large and well-staffed enterprises, both in terms of
IS and business staff (Zachman, 1987; Open Group,
2009); (Gianni et al., 2013). A lightweight approach
would be a logical alternative but requires evaluation
before such a path could be recommended. By
lowering the entry barrier created by the need for
expertise with suggested paths through the ecology
of available tools and techniques, a significant
opportunity for improved decision-making is placed
within reach of entities with IS assets but without the
substantial expertise available. To that end, this
paper is a part of a larger investigation addressing
the following two research questions:
Can a general lightweight elicitation technique
16
Rosasco N. and Dehlinger J..
Application of a Lightweight Enterprise Architecture Elicitation Technique Using a Case Study Approach.
DOI: 10.5220/0004868600160025
In Proceedings of the 9th International Conference on Evaluation of Novel Approaches to Software Engineering (ENASE-2014), pages 16-25
ISBN: 978-989-758-030-7
Copyright
c
2014 SCITEPRESS (Science and Technology Publications, Lda.)
Figure 1: Zachman Framework (as grid) (Zachman, 1987).
gather information suitable for conversion into
enterprise vision, and mission?
Does engaging in the elicitation provide data
sufficient to begin populating an enterprise
architecture framework?
To address these research questions, this paper
presents the overall investigative structure and
research methods employed to study of the
application of a lightweight effort in EA data
elicitation and analysis through a case study. Our
lightweight approach for initial EA elicitation
(Rosasco and Dehlinger, 2011a); (Rosasco and
Dehlinger, 2011b) utilizes an interview vehicle from
business strategy, VMOST (Sondhi, 1999), which
has been employed and extended in this work to
specifically enable the elicitation of the input to a
simple EA framework, the Zachman Framework
(Zachman. 1987). To achieve the necessary
sensemaking (Weick et al., 2005) of qualitative data
and assist in analysing elicited EA information, the
Grounded Theory Method (GTM) has been used
(Strauss and Corbin, 1989).
To apply and evaluate our proposed approach,
this research utilizes a case study approach (Yin,
2003); (Eisenhardt 1989) as the research
methodology. A case study approach affords the
opportunity to increase the utility of EA methods
and tools for smaller or less-resourced institutions by
demonstrating the ease of use of this approach, as it
is being tested and evaluated. This, in turn, reduces
the need for costly changes, potential maintenance
pitfalls and possible stakeholder frustration that will
improve overall effectiveness and help avoid wastes
of time and resources on inappropriate, unnecessary,
or unworkable solutions. By encouraging internal
discussion on strategic planning, overall goals and
objectives within an enterprise, a greater sense of
mission is also likely to result from the employment
of EA techniques (Sondhi, 1999; Minoli, 2008).
The contribution of this work is multipart. The
case study will demonstrate and, primarily, evaluate
our lightweight EA elicitation technique as a way to
solicit the initial input for an EAF. This effort will
also further work undertaken as part of a larger, on-
going study of guided and structured approaches for
EA and investigating how qualitative techniques can
be utilized in a software engineering context.
Specifically, this paper describes the research
methodology and evaluation of our proposed
lightweight EA elicitation technique conducted with
a real, representative enterprise.
The remainder of this paper is organized as
follows. Section 2 reviews background and related
work on the GTM, the VMOST strategy approach,
and EA. Section 3 details our approach and
evaluation within the case study approach. Section 4
provides evaluation results, discussion and caveats
of our evaluation. Finally, Section 5 provides
concluding remarks and future research directions.
2 RELATED WORK
The work for this case study combines techniques
from several areas, including enterprise architecture
(EA), business strategy studies, the VMOST
elicitation vehicle (Sondhi, 1999; Bleistein et al.,
2005) and the Grounded Theory Method (Strauss
and Corbin, 1998).
ApplicationofaLightweightEnterpriseArchitectureElicitationTechniqueUsingaCaseStudyApproach
17
Figure 2: VMOST questions, as adapted by Bleistein (Bleistein et al., 2005); (Bleistein et al., 2006).
2.1 Enterprise Architecture
Frameworks
Developed in 1987 and later elaborated into a
framework for EA, the Zachman Framework is used
as a way to organize and analyse various
presentations of an enterprise (Zachman, 1987). This
EA framework (EAF) was employed to prepare the
elicited EA data into a structure suitable for use in
EA artefact creation. The Zachman Framework
provides taxonomy, in the form of a simple grid, to
document the “building blocks of enterprises”
(Luftman, 1993). It is based on six questions (i.e.,
what, how, where, who, when, and why) and six
general stakeholder-derived perspectives (i.e.,
planner, owner, designer, builder, implementer and
worker) (Zachman, 1987). The resulting framework,
typically depicted as a grid as shown in Figure 1,
provides an overall view of the enterprise that is
extremely flexible but provides little guidance on
how to elicit the necessary information to populate
the schema.
The lack of systematic procedures to guide
enterprise architects in developing an EA is not
unique to the Zachman Framework. For example,
the 780 pages of The Open Group Architectural
Framework (TOGAF) core document poses a similar
problem for the user looking for answers about
structured procedures for the enterprise architect to
elicit and analyse an enterprise’s mission, goals and
objectives (Open Group, 2009). This absence creates
a significant barrier to the application of the
processes and methods, forestalling the use of these
powerful EA concepts outside arenas similar to the
original adopting communities. This barrier, in turn,
could deprive an institution of the insights that come
from viewing IS/IT assets in the context of items in
a portfolio.
2.2 The Vision-Mission-Objectives-
Strategy-Tactics Approach
EA depends on having an enterprise’s vision,
mission and business strategy in hand as the
beginning of integrating, aligning and governing it’s
IT infrastructure with its business architecture using
an EAF (Minoli, 2008). The Vision-Mission-
Objectives-Strategy-Tactics (VMOST) approach
created by Sondhi (Sondhi, 1999) is used by the
business community to understand and improve
strategic comprehension within commercial
enterprises. Sondhi’s work is built for the needs of
the business community and enables those in
management to assess institutional situations so as to
rigorously define and explore options in complex
situations. While providing a hierarchy of
considerations – ranging from tactics, classed as day
to day, to strategy, comprising major objectives –
this tool is only a scaffold for understanding the
layout of an enterprise (Sondhi, 1999).
The VMOST approach, created for strategic
assessment, was adapted by Bleistein to a question-
based format, and is shown in Figure 2. This
adaptation was then combined with specific tools
(Bleistein et al., 2005; Bleistein et al., 2006) for use
within the information systems, technical
management, and decision-making realms. While
valuable, this work was done with static material,
published by the entity being studied, and therefore
is restricted to extant and available material. The
study under discussion, however, interacts with a
live enterprise, accepting the complications that
come from study of a working institution and
operational staffers.
The VMOST questions built for these needs are
taken for this study as a foundation, a tested and
workable basis for a series of closed interviews,
rather than starting from the beginning and risking
ENASE2014-9thInternationalConferenceonEvaluationofNovelSoftwareApproachestoSoftwareEngineering
18
Figure 3: Elicitation and Analysis Methodology, set within the Case Study approach and employing GTM.
omission or oversight. In contrast to the earlier work
the questions are employed as become a direct
elicitation tool, rather than an organizational
structure for existing textual materials.
2.3 Grounded Theory Method
The conduct of this study called for an analytic
method that could provide sensemaking (Weick et
al., 2005) utility of qualitative data. The Grounded
Theory Method (GTM) is a technique created to
allow creation and discovery of theory directly from
qualitative data analysis (Strauss and Corbin, 1998).
The method, authored as a joint effort by
sociologists Glaser and Strauss, employs multiple
stages of data coding to enable full comprehension
of information (Strauss and Corbin, 1998). GTM
provides a rigorous method for analysis of semi-
structured, conversation-derived data, allowing for
significant concepts to be discovered and theory to
be developed and evolved from a variety of sources.
This method is uniquely suited to working with the
broad variety of data-types and inputs accumulated
and solicited in the process of working across a
broad spectrum of disciplines, objectives, and
environments typical of a modern multiple role
enterprise. By allowing for this data to be processed
and analysed, a fuller and more comprehensive
conception of the enterprise and conformance to its
strategic goals is likely to result (Strauss and Corbin,
1998).
GTM has been employed previously in a variety
of software engineering contexts. For example,
GTM has been utilized to examine requirements as a
means to develop UML class diagrams (Chakraborty
and Dehlinger, 2009), to provide sensemaking utility
in analysing non-functional requirements with the
goal of developing non-functional goal trees
(Chakraborty et al., 2012) and to understand the
software maintenance practices of small software
development organizations (Hasan et al., 2011). The
work presented here differs in that it utilizes GTM as
a sensemaking (Weick et al., 2005) procedure
specifically in the analysis of qualitative data
resulting from the elicitation of EA mission, goals
and objectives with the target being the initial EA
artefacts needed in the Zachman Framework.
3 RESEARCH METHODOLOGY
This section elaborates on the configuration of the
case study conducted for investigation of these
concepts, beginning with the selection of the
enterprise to be studied, as this was the primary
driver for the research process. The case study
documented here is informed by the work of
Eisenhardt (Eisenhardt, 1989) on design and theory
creation and Yin (Yin, 2003); (Yin, 2009) on the
design and conduct of the study itself. Their
foundational scholarship addresses a variety of
concerns and considerations surrounding the
specificity and reproducibility of case study acquired
data, and allows for the application of these methods
across a broad range of disciplines and areas of
study, so long as known limitations are
acknowledged.
The results of this approach have the benefit of
providing actual usage feedback, based in concrete
examples and become illustrative to other interested
practitioners and scholars. Yin’s work also posits
that a case study can cope with a technically
distinctive and intricate situation in which there will
be more variables of interest than data points, and
provide one result even while relying on multiple
evidence sources. It also provides overall guidance
on how to engage in and discuss a case study,
including addressing research design, and planning,
methodological questions, and overall rigor.
The approach used in this work, shown in Figure
ApplicationofaLightweightEnterpriseArchitectureElicitationTechniqueUsingaCaseStudyApproach
19
3, to apply and evaluate our proposed lightweight
EA elicitation technique utilizes Yin’s case study
approach (Yin, 2003); (Yin, 2009) and the GTM
coding steps (Strauss and Corbin, 1998). Each of the
steps of our research methodology is discussed in
the following subsections.
3.1 Selecting the Case
The case study approach begins with a definition of
the entities and population under consideration for
the study, so that the appropriate controls, scoping
and eventual limitations are apparent and identified.
This selection defines the overall context and
therefore, the generalizations from and applicability
of the overall results and any theory based on the
data.
Eisenhardt’s synthesis of the various
considerations in executing of a case study
emphasizes the centrality of selecting an appropriate
group for study. This definition is a central factor in
reducing the opportunities for “extraneous
variations” and allows for generalization of limits
(Eisenhardt, 1989). Within that model, two rounds
of population choice would be necessary. The first
would be driven by the theoretical objectives, the
second by statistical and coverage considerations.
3.1.1 Selecting the Enterprise
After identification of the chief question(s) for the
research and adoption of the case study approach,
the next step, according to Yin (Yin, 2009), is the
identification of and engagement with an enterprise.
Concerns at this point in the sequence included
willingness to participate, toleration of the presence
of outsiders in the form of the research staff, and
permissibility of using data. The latter, with
commercial enterprises, sometimes limits the later
publication of results with full examples.
In this instance, a representative of the staff of a
U. S. public University’s library, during an unrelated
discussion, expressed an interest during the initial
planning stages for this study. The use of an
institution of this sort for this research afforded
numerous advantages, chief among them the core of
the staff, which numbered in excess of 50, are all
generally accustomed to research and researchers.
Local conditions also included a complex set of
committees for internal management, a diverse set of
roles and functions, responsibilities spanning
everything from the curricular to legally mandated
records retention, frequently including off-hours and
off-site expectations.
The local information systems environment was
also of significant complexity, involving multiple
public and internal web presences and mechanisms,
a complex set of standards for data interchange,
local support mandates for multiple sets of systems,
and stewardship responsibility for the licensing of a
variety of digital and physical data sources.
Investigating an enterprise with an IS asset portfolio
but limited access to specialized staff, the research
team was presented with a set of circumstances that
is not unusual for an academic sector entity,
including the absence of the profit/loss metric,
which offered an opportunity to distinguish this
work from various prior studies (Bleistein et al.,
2005); (Bleistein et al., 2006).
The diversity of both the stakeholder population,
including a primary user community of 23,000, and
the varied objectives and needs they present, as well
as relationship to external and parent entities,
combined to make this a rich environment for
application and evaluation of our approach. These
connections included involvement in the collective
university and library system created by the state
government, plus interaction with the larger
academic library community. The relative ease of
scheduling and access, and willingness of the staff to
participate, also proved a significant asset.
Overall, the library is comparable to a large
number of peer institutions, and shares a profile
generally similar to even more entities. It exists as
an IS “heavy” institution, confronted with an
evolving spectrum of social media systems and
choices, and is constrained – compared with the
original enterprises employing EA – by limited
staffing and technology expertise. It also has an
evolving mission and, like any enterprise, a need to
budget both time and resources.
3.1.2 Choosing Participants
With a subject enterprise for our case study
evaluation identified and willing to participate,
thenext consideration in the case study approach is
to select which staff would be requested to
participate. In order to cast as wide a net as possible,
the second round of selection included consideration
of the library as a system and the multiple roles and
functions it and the staff performed.
This second stage established which staff within
the enterprise under consideration would be
interviewed as a part of the data collection process
by starting with an existing organization chart and
staff directory of the library. To maximize coverage
and cast as wide a net as possible, participation by as
broad swath of the staff was targeted; ultimately
ENASE2014-9thInternationalConferenceonEvaluationofNovelSoftwareApproachestoSoftwareEngineering
20
Table 1: Example data subset, in Grounded Theory Method three-stage coding.
ultimately over 40% of the full-time/professional
staff would become participants. Additionally,
involvement by staffers representing the operating
groups of the four major functional divisions was
achieved, allowing consideration across multiple
roles, viewpoints and sets of responsibilities.
3.2 Instruments and Protocols
To avoid unnecessary variables and to ensure
consistent behaviours, the instruments and protocols
for a case study must be defined. In this instance,
that required a standardized set of questions as an
elicitation instrument, and a well-specified interview
strategy. These in turn, are fitted within the overall
case study construction, as shown in Figure 3.
The Vision-Mission-Objectives-Strategy-Tactics
(VMOST) method, adapted as questions in
(Bleistein et al., 2005) from (Sondhi, 1999), was the
initial instrument used for this study; this instrument
was later extended as a result of the initial round of
interviews based on the initial data. The original
VMOST questions (shown in Figure 2), designed as
a generic mechanism for developing an enterprise
architecture (EA), are phrased in such a way as to
make them meaningful outside the specialist realms
of business strategy or software engineering. The
VMOST questions are also not tailored to a specific
industry and have been used previously in both
business strategy (Sondhi, 1999) and information
technology scenarios (Bleistein et al., 2005), giving
a reasonable confidence to their reuse in this new
application.
3.3 Fieldwork
After the institution, personnel and study
instruments had been selected using the case study
approach (Eisenhardt, 1989), the next step called for
by the methodology is the fieldwork. This is the
actual execution of the designed study upon the
selected subjects and with the chosen instruments.
In this work, this was the actual conduct of the
interviews. This was followed by transcription of the
collected recordings and some post processing of the
transcripts. The processed transcripts were then
united with the notes taken during the interviews.
It became apparent as the fieldwork application
and evaluation continued that the one-on-one
interview format exercised the questions in such a
way as to capture nuanced reaction. The
combination of notes and recordings provided the
required mechanism, one that permitted interaction
and feedback. This approach also appeared to
reassure the participants on any concerns they may
have had concerning the use of the resulting data and
the objectives of the experiment, a critical
consideration with human subjects. Multiple
subjects were curious about both the goals and the
ultimate outcomes of the experiment, and the in-
person encounters allowed a much more fluid
discussion. The interview format proved that it could
provide for the need to acquire both the responses to
and about the questions.
The interviews conducted using the VMOST
questions generated a set of seven and a half hours
of recordings from the 23 participants that form a
broad base of responses to the VMOST questions,
made broader still by allowing participants to
ApplicationofaLightweightEnterpriseArchitectureElicitationTechniqueUsingaCaseStudyApproach
21
Figure 4: Sample of data presentation artefact, after GTM applied to data.
request clarification of the question and otherwise
comment on the phrasing of and impression made by
the question. These recordings, more than 100 pages
once transcribed, are combined with the concurrent
notes for each interview for asynchronous analysis
of the responses by a second investigator, which the
literature (Eisenhardt, 1989) posits as a step likely to
increase complementary insight. Such additional
insights lead to richer data and identification of
several additional insights. The additional
consideration of the information also, generally,
leads to greater confidence in the analysis, as
differing perception adds to the empirical grounding
of the hypotheses derived. When this method was
joined to a qualitative data collection form, and
applied in a recorded close interview format with
open-ended responses, a multi-format data collection
strategy was created.
The resulting transcripts required some
additional manual processing before further use.
The transcripts were produced by an external,
independent transcribing firm to ensure accuracy
and clarity, and then are redacted to remove certain
types of identifying information, including names,
for protection of the identity of participants.
Extraneous capture, including occasional anomalies
and interruptions, acknowledgement of interview
parameters, and some conversational pauses and
digressions were also removed.
By going beyond a routine pigeonholing of
answers-to-questions, it is possible to have a holistic
interpretation of a full interview. This is in
accordance with the Mintzberg synergy concept
(Mintzberg, 1979), discussed in (Eisenhardt, 1989),
that says that theory can be constructed from “rich
description”. It also led to greater notice of subtle
shadings in phrasing and vocabulary and of
differences arising from distinct specialties and
responsibilities amongst interview participants. It
also distinguished areas where “backtracking” to
answer prior questions by respondents and other,
conversational interaction “moments” occurred.
3.4 Data Analysis
Following the data collection and initial processing
in the fieldwork stage, the case study sequence
moves into the analysis portion of the investigation.
This entails attempting to identify trends, repeated
themes, and other constructs and concepts in order to
understand the overall nature of and information
contained in the data.
The interview sessions and subsequent
processing and editing yielded a set of transcripts
based on the primary interviews of staff. Grounded
Theory Method (GTM) was used to analyse the
qualitative data (Strauss and Corbin, 1998). This
entailed using the three phase coding technique
(open, axial and selective coding).
A set of sample intermediate data is shown in
Table 1Table 1, with associated codes and
demonstrates the stages of the GTM coding
sequence used in this research as a part of the
sensemaking process of analysing the data elicited
from conducted interviews using VMOST questions.
The “Interview excerpt” (shown in the second
column of Table 1) cells contain key words or
phrases, with fuller context beneath, that are used to
inspire the initial, “open” codes that identify
concepts, vocabulary and ideas for consideration
ENASE2014-9thInternationalConferenceonEvaluationofNovelSoftwareApproachestoSoftwareEngineering
22
(Strauss and Corbin, 1998). The open codes (shown
in the third column of Table 1) are words and
phrases driving understanding and comprehension of
the overall enterprise, generated from both the
phrases and contextualization of the interviews
(Strauss and Corbin, 1998). With the given
examples, the open codes are annotated based on
impression and initial reflection(s) of the coder. The
open codes are then grouped and ordered with axial
codes (shown in the fourth column of Table 1),
which typically emerge as the set of open codes are
considered and studied. Selective coding entails
choosing one of the axial codes as the central or
primary idea (not shown in figure), which is what is
perceived to be the touchstone to which the others
can be related for overall consideration of the
situation (Strauss and Corbin, 1998).
3.5 Shaping Hypotheses
In both a theory development case study and an
application of the GTM, one objective is to discover
theories that are grounded in the data. In this
research, the initial review of the coded data has led
to several preliminary conclusions. The initial
review of the data demonstrated remarkable
coverage of the target EA framework (in our case
the Zachman Framework (Zachman, 1987)),
indicating a general validity to the initial concepts
behind the study (Rosasco and Dehlinger, 2011a).
Additionally, after the first set of interviews was
conducted an opportunity for additional data
collection presented itself. The organization under
consideration had been conducting an internal
strategy capture exercise in response to an
institution-wide effort, and the concluding document
was finished and made available. This allowed a
comparison to be made between the library’s
internal efforts and the research results (Rosasco and
Dehlinger, 2011b).
The results of this comparison indicated an
improvement in areas of coverage for the VMOST
methodology versus the more conventional exercise
conducted by the library. This sequence of results
also tracked with the overlap between collection and
analysis, observed by Eisenhardt in several studies
(Eisenhardt, 1989), and permitted some perspective
on the overall responses and increased the awareness
of the investigative staff to the additional contexts
associated with a number of specialist terms and
constructs.
After these two rounds of analysis with the initial
data, a full application of the GTM was conducted
once the complete data was available. The results of
the coding and review resulted in a number of
artefacts and perceptions. A sample of one of these
artefacts is shown in Figure 4, demonstrating the
identification of groups and concepts. Code 2
(“students, faculty”) from the first row in Table 1,
referring to the different types of customer is, for
example, reflected in the classifications being
presented as part of the larger group
“Stakeholders”). Similarly, row 2’s code 2,
“scoring” finds a home – along with a wide variety
of other metrics and tracking related concepts –
under “Assessment/Metrics”. Several of these
artefacts were delivered to the institution’s liaison to
the project team, who responded positively to the
overall understanding and modelling they
represented.
As expected, several syntax and phrasing issues
in the VMOST queries were identified, along with
several general considerations for application of this
process. Enhancements and modifications to the
original VMOST questions are currently being
considered for adaptation specifically in an EA
context. The coverage of the Zachman framework,
while generally good for the relatively low time and
effort impact on the subjects, could be improved by
relatively brief additions to the VMOST questions.
These additional VMOST questions were drafted,
and further interview sessions were conducted with a
subset of the original members of the staff of the
enterprise. This additional result set shows promise
for even higher coverage levels of the framework
and overall interest as a future departure point for
deeper investigation.
3.6 Enfolding Literature
The Eisenhardt strategy for case study execution
allows for consideration of “enfolding literature” –
data text and artefacts that inform the understanding
of the research team (Eisenhardt, 1989). In a
commercial, governmental or other structured entity,
these will often include various external and internal
documents. This sort of input can include
organizational charts, process diagrams and asset
inventories, for example, as well as regulatory and
oversight data.
The enterprise in this study participates in several
larger institutional constructs and possesses a well-
defined internal structure. Consideration of various
governing, strategic and planning materials related
to these larger and complex entities may yet prove
relevant as feedstock for theory and context capture.
This investigation and research, as a part of the
longer-term project, is still underway. Successful
ApplicationofaLightweightEnterpriseArchitectureElicitationTechniqueUsingaCaseStudyApproach
23
evaluation and identification of these items, if any,
may prove useful for other entities, in terms of
determining either inputs into an EA process or
overall consideration of questions of operational
control, scope, stakeholder expectations and general
accountability.
4 DISCUSSION
The application of the VMOST questions (Sondhi,
1999) as a mechanism for EA elicitation and the
utilization of the Grounded Theory Method (GTM)
(Strauss and Corbin, 1998) for sensemaking of the
qualitative data shows promise for enabling
alignment of the IS/IT portfolio with strategic
institutional goals. Specifically, the application of
the lightweight EA elicitation mechanism proposed
in this work has garnered a remarkable body of data
at a cost to the institution that, so far, has proven to
be an easily acceptable additional burden. The
combination of tools employed (i.e., VMOST
(Sondhi, 1999) and the GTM (Strauss and Corbin,
1998)) has been within acceptable scope of effort,
and has created what appears to be a high level of
overall comprehension regarding institutional
objectives, context and goals. Similarly, the case
study approach has proved an effective structure for
handling numerous variables. This can be said with
some confidence since, when using the Zachman
Framework (Figure 1) as a measure, execution of the
initial interview round alone produced inputs for all
but one of the columns of the grid, when the
responses to the two additional questions were used,
the final column (“When”) had input data as well.
While informal, the agreement by enterprise staff
that the diagrammatic artefacts (Figure 4) derived
from the GTM-coded data do correctly present a
working understanding of the institution also affords
an attestation of the validity of the results.
However, as with any experiential investigation,
with a case study it is necessary to consider to the
limitations of the data and process. As this particular
study generated qualitative data, and in using the
GTM applied a qualitative analytic method, there are
potential risks of subjectivity present. There are also
the customary challenges of results replication
where working institutions are concerned. For
example, having “been studied,” the researcher is
unlikely to get the same response to the same
question, if only because the subject has had an
opportunity to consider their initial response.
Additionally, as the institution was not static during
the study routine evolutions and changes, including
changes to the roles of some of those interviewed,
limits the chances of exact duplication later.
Given the origination of the project as a reaction
to possible barriers to adoption of EA, it is worth
noting that the particular analytic method applied
could itself be a hurdle for an enterprise embracing
the precise approach discussed in this work to date.
While it is possible to the GTM could be applied by
this specific enterprise without outside personnel,
given the background and levels of education and
expertise represented, it could prove impractical for
some possible users. Further investigation and
evaluation of other approaches and options to
achieve similar ends are certainly merited.
The lightweight EA elicitation approach
proposed and evaluated here, with these caveats,
could be reused by another, similarly professionally-
staffed institution. No limitations in this study have
surfaced that would prohibit a comparable academic,
non-profit, or governmental institution from
successfully using as a guide the approach
demonstrated by this case study. This study helps
open the way for such an institution to apply EA
and, thus, improve its IS/IT management. In
providing a pattern, this work is a contribution to the
IS artefacts available to practitioners.
5 CONCLUSIONS
The strategy of applying lightweight guidance for
the initial data elicitation and analysis of the
artefacts necessary for an enterprise architecture
(EA) plan shows promise. With a locally-informed
EA plan in hand, there appears to be a high
likelihood of delivering the expected benefits of this
area of practice including clear objectives and good
requirements context information. The use of the
case study method has provided useful results for the
exercise of the approaches being examined and has
successfully documented the use of EA outside of
the large institution and technical-specialist contexts.
In terms of the overall project questions,
lightweight elicitation techniques have exceeded the
initial expectations for collecting data to populate an
EAF, as exemplified by the easily identifiable
information contained in the interview results even
before processing. Additionally, the lightweight
techniques have succeeded in fulfilling the need for
vision and mission capture, as demonstrated by the
richness of the resulting summary artefacts. The
suitability of using this data in conjunction with
other techniques, including the Business Motivation
Model (BMM) (Object Management Group, 2010)
ENASE2014-9thInternationalConferenceonEvaluationofNovelSoftwareApproachestoSoftwareEngineering
24
and Archimate (Open Group, 2012) should also be
investigated as alternative mechanisms. Future work
for this project will include soliciting additional
feedback on those constructed artefacts and
assembly of the various stages into a fuller
presentation of the process, capabilities and
operational considerations for the application of this
lightweight methodology. Further investigation of
the generalizations regarding scalability of the
specific suite of techniques and the overall process
will be undertaken as a part of the overall project.
ACKNOWLEDGEMENTS
This work was partially supported by Towson
University. The authors would like to thank the staff
and administration of Towson University’s Cook
Library for their help and ongoing support and the
reviewers for their comments and insights. This
research was conducted under Towson University’s
Institutional Review Board for the Protection of
Human Subjects exemption number 11-0X14.
REFERENCES
Bleistein, S. J., Cox, K. & Verner, J., 2005. Strategic
alignment in requirements analysis for organizational
IT. In Proceedings of the 2005 ACM Symposium on
Applied Computing - SAC ’05. New York, New
York, USA: ACM Press, p. 1300.
Bleistein, S., Cox, K. & Verner, J., 2006. Validating
strategic alignment of organizational IT requirements
using goal modeling and problem diagrams. Journal of
Systems and Software, 79(3), pp.362–378.
Chakraborty, S. and Dehlinger, J., 2009. Applying the
Grounded Theory Method to Derive Enterprise
System Requirements. 2009 10th ACIS International
Conference on Software Engineering, Artificial
Intelligences, Networking and Parallel/Distributed
Computing, pp.333–338.
Chakraborty, S., Rosenkranz, C., and Dehlinger, J. 2012.
"A Grounded Theoretical and Linguistic Analysis
Approach for Non-Functional Requirements
Analysis." Proceedings of the International
Conference on Information Systems ICIS 2012.
Eisenhardt, M., 1989. Building Theories from Case Study
Research. Academy of Management Review, 14(4),
pp.532–550.
Gianni, D., Lindman, N., Fuchs, J., & Suzic, R. (2012).
Introducing the European Space Agency architectural
framework for space-based systems of systems
engineering. In Complex Systems Design &
Management (pp. 335-346). Springer: Berlin.
Hasan, R., Chakraborty, S. & Dehlinger, J., 2011.
Examining Software Maintenance Processes in Small
Organizations: Findings from a Case Study. In R. Y.
Lee, ed. Software Engineering and Research (SERA),
Springer, pp. 129–143.
Jarvis, R. Enterprise Architecture: Understanding the
Bigger Picture – A Best Practice Guide for Decision
Makers in IT, The UK National Computing Centre,
Manchester, UK.
Lankhorst, M., 2005. Enterprise Architecture at Work:
Modeling Communications and Analysis, Springer-
Verlag, Berlin, Germany.
Luftman, J., Lewis, P. R., and Oldach, S. H., 1993.
“Transforming the Enterprise: The Alignment of
Business and Information Technology Strategies”, In
IBM Systems Journal, 32(1):198-221.
Minoli, D., 2008. Enterprise Architecture A to Z (1st ed.),
Boca Raton, FL: Auerbach/CRC Press.
Mintzberg, H., 1979. An emerging strategy of “direct”
research. Administrative Science Quarterly, 24, pp.
580-589.
Object Management Group, 2010. Business Motivation
Model (BMM) Version 1.1. At: http://www.omg.org/
spec/BMM/1.1/
The Open Group, 2012. Archimate 2.1 Specification, At:
http://pubs.opengroup.org/architecture/archimate2-doc
/toc.html
The Open Group Architecture Forum - Forde, C., 2009.
The Open Group Architecture Framework (TOGAF)
9th ed., Reading, Berkshire, UK.
Rosasco, N. and Dehlinger, J., 2011. “Eliciting Business
Architecture Information in Enterprise Architecture
Frameworks Using VMOST.” 1
st
ACIS International
Conference on Computers, Networks, Systems and
Industrial Engineering, pp. 474–478.
Rosasco, N. and Dehlinger, J., 2011. "Business
Architecture Elicitation for Enterprise Architecture:
VMOST versus Conventional Strategy Capture."
Ninth International Conference on Software
Engineering Research, Management and Applications,
pp. 153-157.
Sondhi, R.K., 1999. Total Strategy, Airworthy
Publications International.
Strauss A. and Corbin J., 1998. Grounded Theory
Methodology: An Overview, Sage Publications,
Thousand Oaks, CA.
Weick, K. E., Sutcliffe, K. M. & Obstfeld, D., 2013.
Organization Science and the Process of Sensemaking.
Organization Science, 16(4), pp.409–421.
Yin, R. K., 2003. Applications of Case Study Research
(2nd ed., Vol. 34). Thousand Oaks, CA: SAGE
Publications.
Yin, R. K., 2009. Case Study Research Design and
Methods (4th ed., Vol. 5). Thousand Oaks, CA: SAGE
Publications.
Zachman, J. A., 1987. “A framework for information
systems architecture,” In IBM Systems Journal, vol.
26, pp. 276-292.
ApplicationofaLightweightEnterpriseArchitectureElicitationTechniqueUsingaCaseStudyApproach
25
