Conceptual Design of a Method to Support
IS Security Investment Decisions within the Context of
Critical Business Processes
Heinz Lothar Grob, Gereon Strauch and Jan Hermans
European Research Center for Information Systems, University of Muenster
Leonardo-Campus 3, 48149 Muenster, Germany
Abstract. In order to safeguard the compliance of information systems, private
enterprises and governmental organizations can implement a large variety of
distinct measures, ranging from technical measures to organizational measures.
Especially in the context of critical information system infrastructure e.g. data
centers, the decision for specific safeguards is complex. An appropriate method
for the profitability assessment of alternative IS security measures in the context
of critical business processes has not so far been developed. With this article we
propose a conceptual design for a method which enables the determination of
the success of alternative security investments on the basis of a process-oriented
perspective. Within the scope of a design science approach we combine
established artifacts of the field of IS security management with those of the
field of process management and controlling. On that basis we develop a
concept that allows decision-makers to prioritize the investments for dedicated
IS safeguards in the context of critical business processes.
1 Introduction
Information Systems (IS) security generally and critical business process specially
raised in the past more attention of budget responsible people – visible by the above-
average increase of IS security budgets compared to overall IS budgets and the
increaing relevance of this research community [1]. In the meantime though, more
recent works emphasis the elementary imperative of profitability analyses – despite
available findings this field of research are frequently characterized as being vague,
unusable or without reference to concrete recommendations for a course of action [2,
3]. In order to conceptualize a method for the decision support for IS security
investments in the context of critical business processes, these special challenges need
to be considered. The chosen research approach can be characterized as design-
oriented, where a conceptual-deductive research method has been applied [4, 5]. A
brief overview of the related work in this field shows that the suggested methods do
not fit the special requirements in the context of critical processes. Most approaches
in context suppose a linear exchange relationship between expected loss and the costs
of security measures [6]. This procedure does not apply for information systems,
which have a vital meaning for the organization [7]. Our main objective is to provide
Lothar Grob H., Strauch G. and Hermans J. (2008).
Conceptual Design of a Method to Support IS Security Investment Decisions within the Context of Critical Business Processes.
In Proceedings of the 6th International Workshop on Security in Information Systems, pages 113-121
DOI: 10.5220/0001741601130121
Copyright
c
SciTePress
a method for decision support for security investments within critical infrastructures
and to integrate this into an overall IS risk management procedure. So we define
requirements in this context and offer an outlook to an approach for controlling
security measures for critical business processes and information infrastructures (such
as data centers) based on a configurable criteria system. To support the
implementation of this method within a management information system, we develop
a conceptual model and specify the formal requirements for adequate portfolios of
safeguards afterwards. The article concludes with a brief summary and an outlook on
future research opportunities.
2 Decision Support for Security Investment within Critical
Infrastructures
2.1 Identifying Critical Processes and Aligned Information Systems
IS risk management deals with risks resulting from the usage of information systems
in a company. The procedure of tasks is oriented at the general process of risk
management as shown in figure 1. Focusing on business processes has been claimed
repeatedly for the IS security management to keep the security compliant to the
business goals [3, 8, 9]. Critical business processes should lead more the other
security management policies than within the scope of "normal" risk disposition and
in the same way to other instruments for decision support. Therefore, it is
recommended to fulfill different safeguard planning procedures for business processes
and associated information systems with normal risk disposition, critical business
processes and underlying critical IS infrastructures (e.g. data center). Criticality
analysis (CA) or business impact analysis (BIA) are usually carried out within the
bounds of risk identification and risk analysis to identify critical business processes
and the appropriate information systems (critical IS infrastructure) [10, 11]. We
recommend applying established approaches from investment theory to profitability
analysis based on process models [12], to information systems with normal risk
disposition according to the BSI IT-Grundschutz Methodology. The procedure
adjusted for regarding critical processes is shown in figure 1. The critical analysis is
an approach for identification of critical business processes [11]. The "Joint
Standards" are the accumulation of standards, which were published by the Business
Continuity Institute (BCI) and Disaster Recovery Institute International (DRII) [10]
for establishing a business impact analysis. Within this analysis, the relevance of
business processes is tried to identify. If there are fatal consequences appeared, this
process is considered to be critical. In case of critical analysis, business processes and
each information system that is to be applied for corresponding process are assigned
to different categories.
114
Fig. 1. Procedure model for IS risk management regarding critical infrastructures.
Seibold proposes the classification of processes in 3 to 6 groups [11], it complies
with established approaches from theory and practice like IT-Grundschutz-
Methodology and BIA [10, 13-17]. In his example he classifies the processes into four
classes A-D, where the processes of A class cause fatal consequences in one day,
class B – in 3 days and the processes of D class have no fatal consequences at all. The
information systems accompanying the critical business processes are called here as
critical IS infrastructures [7]. In this context, frameworks for decision support usually
do not support differentiated endangerment scenarios. So the selection of safeguards
should be supported by a configurable criteria system. The procedure of the
configuration, system analysis and the selection of adequate measure are introduced
in the following.
2.2 Procedure Model for Decision Support in the Context of Critical
IS Infrastructure
The core idea is to select different safeguards with a criteria system that can be
configured for the isolated case and its specific context. The criteria system should
help selecting necessary safeguards as well as controlling the compliance to a
required security level. The use of the criteria catalogue follows itself to a procedure
model. The procedural model follows the procedure model of the IT-Grundschutz
Methodology and the advanced risk analysis based on BSI standard 100-3 and state of
the art risk assessment approaches [14, 15, 18]. One very important extension to
common standards is that particular criteria can be defined as absolutely necessary
(so-called lethal criteria). In the case of non-fulfilment one of these criteria, the
necessary protection of the critical infrastructure as a whole is not guaranteed. At first
an analysis of requirements, technical context and specific endangerments should be
carried out, in order to adapt the criteria system. On this Base an as-is analysis of the
existing systems should be executed in order to identify the unaccomplished criteria
with a certain focus to lethal criteria. Thereafter it is possible to identify the possible
bundles of actions to fulfil all (necessary) lethal criteria. The procedure model is
shown in figure 2.
Having discussed how to identify critical IS infrastructure, we focus now on the
subsequent phases of the procedure model, especially the requirements and the
configuration processes that both deal with the adaptability of the criteria system. The
application of a uniform, monolithic criteria system would not ensure the
heterogeneity of the different application contexts. It is also not suitable for different
scenarios to define only a scale of varied levels: not the level of security can differ but
115
specific security requirements will result from application context. E.g. highly
availability can be archived by highest reliable systems or massive redundancy. The
first solution fits the needs of a core banking system, focusing the integrity of even
every transaction, the second approach of "peer production of suitable infrastructures"
[19] is e.g. used for the critical business process of Google [20]. So the criteria should
be adapted complying with the relevant environmental factors (requirements and
endangerments) and constitution parameters of the critical IS infrastructure. On this
occasion, threat scenarios classified as relevant and the enterprise-related application
context should be considered. Table 1 contains examples for different endangerments
that should lead to different (lethal) criteria for evaluating the critical IS
infrastructure.
Fig. 2. Procedure model for IS risk management regarding critical infrastructures.
After aligning the criteria system to a specific scenario, an as-is analysis should be
carried out as a weak point analysis. The criteria system contributes to identify weak
points to be repaired, in which it reproaches a huge number of measures for lethal
criteria, which are not fulfilled and have thus top priority. Every criterion also should
have a questionnaire to raise all important parameters. In the connection, all possible
action, which can be carried out to improve the level of the criteria, can be identified.
Through this, only such action portfolios fulfill the defined minimum requirements in
order to achieve all lethal criteria, are part of the allowed portfolios of the necessary
measures. The portfolio, which shows the slightest total cost of ownership (TCO), can
be selected. Other (more expensive) portfolios can be taken into consideration in the
frame of a "bargaining solution" if these fulfill more non lethal criteria in higher
measure to mention multiple objects [21].
3 Conceptual Design of a Decision Support System for Selecting
Optimal Action Sets
3.1 Design of a configurable Criteria System for Decision Support
By the development of a criteria system, for the assessment of critical infrastructures
three levels are to be regarded. In the core, the real criteria system is located itself on
the two essential fields for the availability, which refer to current enterprise and the
restart of the systems after an incident.
For decision support a business intelligence layer should provide different kinds of
indexes, reports and dashboards and drill-down functionality to the different level of
criteria aggregation. The adaptation will occur through the choice of the criteria, their
scaling and the way of their settlement into an index. The base for the criteria system
116
is the form level. A form repository should support the evaluation of criteria with
questionnaires for each one. The contents of the criteria system should be descended
from the relevant standards. Criteria will be affiliated during the as-is-analysis to real
existing entities to control their compliance to the security policy. Most frameworks
in this context focus on technical aspects and regard organizational considerations
only on the brink. Given the importance of these questions we suggest a multi-
perspective design, inclosing an organizational and a process oriented dimension of
every technical criterion facing the questions, how to observe the criterion and who is
responsible for that [22]. The criteria simultaneously should be divided into different
classes according to best practice standards to improve the transparency of this system
[15, 23]. The connections or cause-effect relations between the specific criteria within
the whole criteria system can be visualized analogously to the strategy map of a
balanced scorecard in an area map [24]. By this multidimensional view, detailed
evaluations within the particular areas are possible. So every criterion contains
necessary value for all these dimensions and is associated with actions to ensure these
values, which cause defined costs to implement them. The evaluation of the criteria
bases on the questionnaires of the form level. At this level, it is defined, how the
single criteria should be raised. In addition to the elevation way, the elevation time
should also be defined at this level. In dependence of the single criterion, suitable
methods should be identified and should be specified in a discipline-conceptual draft.
The forms can serve as templates for the concrete arrangement of a specific criteria
system. The forms should be raised and their contents should be adapted accordingly
to the modeled context. This design should lead to the construction of a management
information system which offers support for risk assessment and governance within
critical IS infrastructures by identifying adequate portfolios of actions. To gain a
deeper understanding of meeting these requirements and computing of adequate
portfolios, we introduce a conceptual model in the following section.
3.2 Conceptual Model for the Computation of Action Bundles
Figure 3 depicts an entity relationship model [25, 26] that provides the basis for the
computation of suitable action bundles. In the model, the entity type »Scenario« is
used to describe the environment, in which security-related actions (»Action«) are to
be applied. Each scenario consists of a set of criteria, that the actions in a bundle
(»Action Bundle«) have to satisfy. Accordingly, instances of the entity type
»Criterion« embody the requirements of scenarios. As stated above, we differentiate
between lethal and non-lethal criteria by introducing the two entity types »Lethal
Criterion« and »Non-lethal Criterion«. The entity type »Value« is used to create
presets for the different criteria. Thereby, a value is assigned to its criterion via the
relationship type »VALCRIS«. By connecting values with a scenario through
»VALSCE«, we express which criteria different actions have to satisfy in order to
make up a suitable bundle. We describe actions in the same way as we describe
scenarios by assigning instances of the entity type »Value«. A comparison of the
values that a scenario requires, and the values that an action exhibits, serves as a
starting point for the computation of action bundles.
117
Fig. 3. Conceptual Model for Computing Action Bundles.
Due to space restrictions however, the conceptual model cannot be elaborated in
greater detail at this point, especially in respect to the trivial modellizing of the
attributes like different perspectives and the costs of the affiliated actions for a single
criterion. The relationship type »Interdependency« allows combining different
actions. As different types of interdependencies, we distinguish between (1) the
simple dependency (»Dependency«), (2) the extension (»Extension«), and (3) the
exclusion (»Exclusion«):
1. If action a depends on action b, a bundle covering action a also has to
contain action b.
2. If action a extends action b, action a also satisfies the criteria that are
fulfilled by action b.
3. If action a excludes action b, a bundle covering action a must not contain
action b.
3.3 Formalizing the Computation of Action Bundles
In order to explain how to compute suitable action bundles, we express some of the
conceptual model elements by sets. In the following, let ACT be the set of all actions,
SCE the set of all scenarios, and VAL the set of all values. The subset
VALSCE VAL × SCE expresses the values which describe a scenario, while the
subset VALSCE VAL × SCE depicts the values which characterize an action.
Furthermore, a certain action bundle acb consists of a set of actions. Hence, the power
set of ACT describes the set of all (theoretically) possible action bundles viz. ACB.
The subset INT ACT × ACT of the Cartesian product of two action sets expresses
interdependencies. Dependencies, extensions, and exclusions are defined as subsets of
INT, so that INT = DEP EXT EXC holds. Next, we introduce the function
118
dep: ACT (ACT). By this function, we compute the set of all actions required by
a certain action act in order to satisfy the dependency relation:
()
(
)
{
}
dep act dep ACT | act,dep DEP=∈
(1)
With ACBdep we define the set, which contains all action bundles, whose actions
satisfy all dependencies required within the bundle:
()
(
)
{
}
dep dep dep
ACB acb ACT | act acb, act dep act :act acb=∈
(2)
Next, we introduce the function exc: ACT (ACT), by which we compute the
set of all actions excluded by a certain action act:
()
(
)
{
}
exc act exc ACT | act,exc EXC=∈
(3)
With ACBexc, we denote the set, which contains all action bundles, whose actions
do not violate any exclusion required within the bundle:
()
(
)
{
}
exc exc exc
ACB acb ACT | act acb, act exc act :act acb=∈
(4)
By the function ext: act (ACT), we compute all actions that extend a certain
action act:
()
(
)
{
}
ext act ext ACT | act, ext EXT=∈
(5)
Based on eq. 5, we introduce the function actval: ACT (VAL) to calculate all
values covered by a certain action act:
(
)
(
)
{
()( )
}
ext ext
actval act val VAL | val,act VALACT
act ext act : val,act VALACT
=∈
∨∃
(6)
In order to compute all values of lethal criteria for a scenario sce, we use the
function letval:
()
(
)
{
()
}
letval sce val VAL | val,sce VALSCE
val,crit VALCRIT crit LCRIT
=∈
∧∈
(7)
Based on eq. 6 and eq. 7, we define the function ACBval: SCE (ACT), by
which we compute all action bundles that satisfy the requirements of a certain
scenario sce:
()
(
)
{
(
)
()
}
val
ACB sce acb ACT | val letval sce
act acb : val actval act
=∈
∃∈
(8)
By computing the intersection of the sets defined in eq. 2, eq. 4, and eq. 8, we
establish all action bundles which are suitable for a certain scenario sce:
()
dep ext val
ACB sce ACB ACB ACB=∩
(9)
119
In order to select the optimal bundle, we take the cost of the action bundles into
account. Therefore, we assume, that the cost caused by a certain action act is given by
the function cost: ACT . Based on this assumption, we can compute the cost of a
certain action bundle acb:
()
(
)
act acb
cost acb cost act
=
(10)
By applying eq. 10 to each element of eq. 9, we can compute the cost optimal
action bundle.
4 Summary and Outlook
With this paper, a procedure model for the decision support of IS safeguards in the
context of critical business processes has been introduced. Since then IS security
investments have been primarily exhibiting a direct impact on the organizational
processes, the latter are in the focus of the suggested method. Therefore, we suggested
identifying the most important processes, the critical processes by a criticality
analysis. Existing approaches were integrated into a generic proceeding model for IS
risk management in the case of regular risk disposition. In addition, the necessity of a
distinction between such methods for regular and critical business processes was
shown. After a refinement of the procedure for critical business process requirements
for decision support in this context were developed. For that purpose we recommend
the development of a management information system to handle different criteria
resulting from the multitude of relevant standards e.g. for data center security.
Thereafter, the necessary structure of a criteria system for critical infrastructure and
the procedure model to apply this to existing business information systems were
shown. Afterwards, we formalized the necessary conditions to identify cost optimal
bundles of actions to provide a basis for our proof of concept. Further research
demand lies in parameterizing the criteria system with relevant norms. At this time
several countries in Europe joined their efforts conforming their standards and
frameworks to a common base. So we expect in the near future detailed common
criteria catalogues for critical information system, which can be applied with the
presented approach
References
1. Anderson, R., Moore, T.: The Economics of Information Security. Science 314 (2006) 610–
613
2. Sonnenreich, W., Albanese, J., Stout, B.: Return on security investment (ROSI). A practical
quantitative model. In: Fernández-Medina, E., Hernández, J.C., García, L.J. (eds.): Security
in Information Systems, 3rd Int. Workshop on Security in Information Systems
(WOSIS'05), In conjunction with ICEIS'05 (2005), New York (2005) 239-252
3. Neubauer, T., Klemen, M., Biffl, S.: Business process-based valuation of IT-security. In:
Sullivan, K. (ed.): Seventh international workshop on Economics-driven software
engineering research, St. Louis (2005) 1- 5
120
4. Hevner, A.R., March, S.T., Park, J., Ram, S.: Design Science in Information Systems
Research. MIS Quarterly 28 (2004) 75-105
5. Becker, J., Niehaves, B.: Epistemological Perspectives on IS Research - A Framework for
Analyzing and Systematizing Epistemological Assumptions. Information Systems Journal
17 (2007) 197-214
6. Le Veque, V.: Information Security - a Strategic Approach. Wiley, Hoboken (2006)
7. Hyslop, M.: Critical Information Infrastructures: Resilience and Protection. Springer, New
York (2007)
8. Röhrig, S.: Using Process Models to Analyse IT Security Requirements. Zürich (2003)
9. Jakoubi, S., Tjoa, S., Quirchmayr, G.: Rope: A Methodology for Enabling the Risk-Aware
Modelling and Simulation of Business Processes. In: Österle, H., Schelp, J., Winter, R.
(eds.): Fifteenth European Conference on Information Systems, St. Gallen (2007) 1596-
1607
10. Business Continuity Institute: Business Continuity Management - Good Practice. In:
Institute, T.B.C. (ed.): (2005)
11. Seibold, H.: It-Risikomanagement, München (2006)
12. Grob, H.L., Strauch, G., Buddendick, C.: Conceptual Design of a Method to Support IS
Security Investment Decisions. In: Kop, C., Kaschek, R. (eds.): International Conference on
Information Systems Technology and its Applications (ISTA 08), Klagenfurt (2008)
13. BSI: BSI-Standard 100-2: IT-Baseline Protection Methodology. (2005)
14. BSI: BSI-Standards 100-3: Risk Analysis based on IT-Baseline Protection. (2005) 19
15. BSI: IT-Baseline Protection Catalogues. Bonn (2007)
16. von Rössing, R.: Betriebliches Kontinuitätsmanagement. mitp Verlag, Bonn (2005)
17. Kairab, S.: A Practical Guide to Security Assesments. Auerbach, Boca Raton (2005)
18. Vidalis, S., Blyth, A.: Understanding and Developing a Threat Assessment Model. 2nd
European Conference on Information Warefare, London (2002)
19. Benkler, Y.: Peer Production of Survivable Critical Infrastructures. In: Grady, M.F., Parisi,
F. (eds.): The Law and Economics of Cybersecurity. Cambridge University Press,
Cambridge (2006) 73-114
20. Barroso, L.A., Dean, J., Hölzle, U.: Web Search for a Planet: The Google Cluster
Architecture. IEEE Micro 23 (2003) 22-28
21. Neubauer, T., Heurix, J.: Defining Secure Business Processes with Respect to Multiple
Targets. In: Jakoubi, S., Tjoa, S., Weipel, E.R. (eds.): Third International Conference on
Availability, Reliability and Security (ARES 2008). IEEE Computer Society, Barcelona,
Spain (2008) 758-764
22. Asnar, Y., Giorgini, P.: Modelling Risk and Indentifiying Countermeasure in
Organisations. In: Lopez, J. (ed.): Critical Information Infrastructures Security: First
International Workshop, Critis 2006, Vol. 4347. Springer, LNCS, Samos Island, Greece
(2006 ) 79-90
23. BITKOM: Reliable Data Centers Guideline. (2006)
24. Kaplan, R.S., Norton, D.P.: The Balanced Scorecard-Measures that Drive Performance.
Harvard Business Review 70 (1992) 71-79
25. Chen, P.P.: Entity-Relationship Model: Towards a Unified View of Data. ACM
Transactions on Database Systems 1 (1976) 9-36
26. Scheer, A.-W.: ARIS - Business Process Modeling. Springer, Berlin (1998)
121