Attack Modelling and Security Evaluation for Security Information
and Event Management
Igor Kotenko, Andrey Chechulin and Evgenia Novikova
Laboratory of Computer Security Problems, St. Petersburg Institute for Informatics and Automation (SPIIRAS),
14th Liniya, Saint-Petersburg, Russia
Keywords: Attack Modelling, Security Evaluation, Security Information and Event Management Systems, Attack
Graph, Service Dependences Graph, Zero Day Vulnerabilities.
Abstract: The paper considers an approach to attack modelling in Security Information and Event Management
(SIEM) systems. The suggested approach incorporates usage of service dependency graphs and zero-day
vulnerabilities to produce attack graph, calculation of security metrics based on attack graph and service
dependencies and advanced any-time techniques for attack graph generation and security evaluation, etc.
1 INTRODUCTION
The complexity of computer network security
management causes the necessity to develop
powerful automated security analysis components
which can be important components of Security
Information and Event Management (SIEM) systems
(Miller et al., 2011); (MASSIF, 2012). These
components should allow finding and correcting
errors in the network configuration, reveal possible
assault actions for different security threats,
determine critical network resources and choose an
effective security policy and security mechanisms
appropriate to current threats.
The paper considers attack modelling security
evaluation processes, intended to be implemented
for the security analysis in SIEM systems.
There are a lot of papers, which consider
different approaches to attack modelling and
security evaluation taking into account various
classes of attacks. In (Huang and Wicks, 1998);
(Moore et al., 2001); (Dawkins et al., 2002) attacks
are described and modelled in a structured and
reusable “tree”-based form. Different approaches,
which use attack graphs and trees for security
analysis, have been suggested. Hariri et al. (Hariri et
al., 2003) calculate global metrics to analyze and
proactively manage the effects of complex network
faults and attacks. Noel et al. (Noel et al., 2003);
(Wang et al., 2006) propose a risk based technique
based on determining the minimum-cost network
hardening via exploit dependency graphs. Kotenko
and Stepashkin (Kotenko and Stepashkin, 2006);
(Kotenko et al., 2011) are focused on security
metrics computations based on attack graph
representation of malefactor behaviour. Lippmann
and Ingols (Lippmann and Ingols, 2006); (Ingols et
al., 2009) propose to use attack graphs to detect
firewall configuration defects and host critical
vulnerabilities. Wang et al. (Wang et al., 2008);
(Wang et al., 2011) propose an approach to calculate
attack resistance metrics based on probabilistic
scores by combining CVSS scores (CVSS, 2012).
Kheir et al. (Kheir et al., 2010) suggest an
implementation of confidentiality, integrity and
availability metrics using the notion of privilege,
which is inspired by access permissions within
access control policies.
We suggest an approach based on the following
main procedures: usage of comprehensive internal
security repository and open security databases;
generation of attack trees considering service
dependency graphs and zero-day vulnerabilities;
application of anytime algorithms to provide near
real-time attack modelling; usage of attack graphs to
predict possible malefactor’s actions; calculation of
a multitude of security metrics, attack and response
impacts; interactive decision support to select the
security solutions. The main difference of the
offered approach from the already suggested ones is
the integration of these functionalities in one
component to achieve better results in near real time
effective attack modelling and security evaluation.
391
Kotenko I., Chechulin A. and Novikova E..
Attack Modelling and Security Evaluation for Security Information and Event Management.
DOI: 10.5220/0004063403910394
In Proceedings of the International Conference on Security and Cryptography (SECRYPT-2012), pages 391-394
ISBN: 978-989-8565-24-2
Copyright
c
2012 SCITEPRESS (Science and Technology Publications, Lda.)
In the paper we discuss the architectural solution
of the proposed Attack Modelling and Security
Evaluation Component (AMSEC) as one of the
important SIEM subsystem and the techniques used
to realize main AMSEC functionality. To illustrate
these architecture and techniques we developed a
software prototype and carried out experiments for
different case-studies.
The rest of the paper is organized as follows.
Section 2 discusses the AMSEC framework. In
Section 3, we describe the AMSEC implementation
and an example of experiments. Conclusion analyzes
the paper results and future research.
2 FRAMEWORK
According to the analysis of state-of-the-art in attack
modelling we selected the following key elements to
be included in the architectural solution of AMSEC
as part of the SIEM-system:
Comprehensive security data repository;
Effective attack tree and service dependencies
generation techniques based on TVA (Topological
Vulnerability Analysis) approach which enumerates
potential sequences of exploits of known
vulnerabilities to build attack graphs;
Attack graph generation considering both known
and zero-day vulnerabilities;
Usage of anytime algorithms for near-real time
attack sub-graph (re)generation and analytical
modelling;
Stochastic analytical modelling;
Combined usage of attack graphs and service
dependency graphs;
Security metric calculation, including attack
impact, response efficiency, response collateral
damages, attack potentiality, attacker skill level
assessment, etc.;
Interactive decision support to select the
solutions on security measures/tools by defining
their preferences regarding different types of
requirements (risks, costs, benefits) and setting
trade-offs between high-level security objectives.
To bind the key elements we developed the
following generalized architecture of AMSEC. The
brief descriptions of the AMSEC’s modules and
their functions are given below.
Network interface supports interaction with
external environment (sending requests to external
databases and communicating with data sources).
Interactive decision support module provides the
user (decision maker) with the ability to select the
solutions on countermeasures by defining their
preferences regarding different types of
requirements and setting trade-offs between objects.
Decision support can include three phases: setting
feasible security solutions (security measures/tools);
identification of efficient (Pareto-optimal) security
solutions; selection (generation) of the final solution.
Generator of system and security policy
specification converts the information about network
configuration and security policy received from the
data collection and correlation components or user
into internal representation. It is supposed, that at the
design stage, this information is specified on special
System Description Language and Security Policy
Language. Used specifications of the analyzed
network (system) and the security policy should
describe network components with the necessary
degree of detail, for example, the used software
should be set in the form of names and versions.
Data repository updater downloads the open
databases of vulnerabilities, attacks, configuration,
weaknesses, platforms, countermeasures, etc., for
example, National Vulnerability Database (NVD)
(NVD, 2012), Common Vulnerabilities and
Exposures (CVE) (CVE, 2012), Common Platform
Enumeration (CPE) (CPE, 2012), and then translates
them into the AMSEC security data repository.
Reports generator shows vulnerabilities detected
by AMSEC, represents “weak” places, generates
recommendations on strengthening the security
level, etc.
Security repository is a hybrid (relational, XML-
based and triplet-based) data storage which contains
information necessary for attack graph generation
and analysis. We suggest to use a set of MSM
related standards (MSM, 2012) or other related
standards for the common enumeration, expression
and reporting of cyber-security-related information
as the basis for the design of the common security
repository.
Malefactor Modeller is responsible for
malefactor modelling and is used on both design and
exploitation stage of the AMSEC operation. On the
first phase it is used to build the set of all possible
attack graphs using preset characteristics of
malefactor (the malefactor profile) which are
determined by the user. Later on the second phase it
allows predicting the possible characteristics of the
malefactor according to the actions fulfilled.
Attack Graph Generator is responsible for attack
graph building. In our approach we use hierarchical
model of attack scenarios which consists of three
levels: integrated, script and actions levels.
SECRYPT2012-InternationalConferenceonSecurityandCryptography
392
The algorithm of generating the common attack
graph is based on the attack scenarios model. The
result of the algorithm is attack graph which
describes the possible routes of attack actions in the
view of malefactor’s initial position, skill level,
network configuration and used security policy.
Attack Graph Generator operates in conjunction
with Manager of Service Dependencies and
Generator of Attack Graph Based on Zero-day
Vulnerabilities to obtain more precise results in
attack modelling. The usage of service dependency
graphs makes it possible to exclude information
about attack impacts from the attack graph and to
use the dependency graph in order to simulate
impacts and obtain a dynamic evaluation of an
attack impact. In our approach we take to into
account zero-day vulnerabilities to generate attack
graph. To do this we modified the approach
suggested in (Ingols et al., 2009) by adding
additional characteristics which define probability of
existence of the zero day vulnerability. The main
idea is to automate process of selection of hosts
which are likely to have zero day vulnerabilities then
others (instead of manual search).
The attack graph is based on the network model
and the probabilities of vulnerabilities defined as
weight coefficients. The net of the connected
information sensors that are able to detect attacks is
formed in the real network. The monitoring system
allows constructing general view about the events
that take place in the network according to the
information sensors.
Security Evaluator is responsible for qualitative
and quantitative assessment of the system security.
For qualitative express assessment of the network
security we planning to use several approaches
which are based on different security metrics, risk
analysis and security evaluation techniques.
3 EXPERIMENTS
By now a prototype of AMSEC, which can generate
possible attack trees for a predefined network and
evaluate the network security level, was
implemented. It contains three basic components:
VDBUpdater, Network Constructor and Security
Level Evaluator. Additionally the prototype includes
the MySQL database as a common repository.
VDBUpdater allows updating the internal database
of known vulnerabilities, using information obtained
from National Vulnerability Database (NVD, 2012).
Network Constructor allows creating models of
tested computer networks and checks selected
software and hardware to match NVD dictionary.
Security Level Evaluator makes topological
vulnerability analysis and evaluates the security
level of the network.
A set of experiments with the prototype of
AMSEC was conducted. The prototype makes use of
two scenarios (MASSIF, 2012): “Managed
Enterprise Service Infrastructures” and “Critical
Infrastructure Process Control (Dam)”.
Let us consider the experiments where we chose
a malefactor located outside the controlled network
of the dam infrastructure. After constructing the
attack graph, the AMSEC provides the following
information: the malefactor knowledge after all
possible attacks, the attack tree in the graphic form
and the log of the malefactor’s actions. Figure 1
illustrates different attacks traces that the malefactor
can perform in the tested network. The malefactor,
carrying out attack actions, is located in the centre of
the spherical representation. The other icons are as
follows: “A” – an attack action, “S” – a scenario
which does not use vulnerabilities (for example, host
discovery (PING)), “V” – an attack action which
exploits some vulnerability. According to the
suggested metrics the security level of the tested
network is evaluated.
Figure 1: Attack graph.
4 CONCLUSIONS
In the paper we presented our approach to the attack
AttackModellingandSecurityEvaluationforSecurityInformationandEventManagement
393
modelling and security evaluation. We also
described the developed prototype of the AMSEC,
which can generate a possible attack tree for a
predefined network and a simple experiment was
considered.
The future steps of the research will be devoted
to detailed elaboration of all AMSEC components.
One of the important research issues is development
of techniques which can cope with large networks,
such as those in enterprise infrastructure.
Also it is planned to optimize the generation of
attack trees through the use of the ontology based
repository, to expand the list of parameters,
characterizing the hosts and the network, to improve
the malefactor model, and to add currently
unrealized components.
ACKNOWLEDGEMENTS
This research is being supported by grant of the
Russian Foundation of Basic Research (projects 10-
01-00826), the Program of fundamental research of
the Department for Nanotechnologies and
Informational Technologies of the Russian Academy
of Sciences, the State contract #11.519.11.4008 and
by the EU as part of the SecFutur and MASSIF
projects.
REFERENCES
CPE, 2012. Common Platform Enumeration, viewed 01
March 2012, <http://cpe.mitre.org/>.
CVE,2012. Common Vulnerabilities and Exposures.
viewed 01 March 2012, <http://cve.mitre.org/>.
CVSS,2012. Common Vulnerability Scoring System,
viewed 01 March 2012, <http://www.first.org/cvss/>.
Dawkins, J., Campbell, C., Hale. J., 2002. Modeling
network attacks: Extending the attack tree paradigm.
In Workshop on Statistical and Machine Learning
Techniques in Computer Intrusion Detection, Johns
Hopkins University.
Hariri, S., Qu, G., Dharmagadda, T., Ramkishore, M.,
Raghavendra C. S., 2003. Impact Analysis of Faults
and Attacks in Large-Scale Networks. In IEEE
Security and Privacy, vol.1 pp.49-54.
Huang, M.-Y., Wicks, T. M., 1998. A Large-scale
Distributed Intrusion Detection Framework Based on
Attack Strategy Analysis. In First International
Workshop on the Recent Advances in Intrusion
Detection, Raid’98, Louvain-la-Neuve, Belgium.
Ingols, K., Chu, M., Lippmann, R., Webster, S., Boyer, S.,
2009. Modeling modern network attacks and
countermeasures using attack graphs. In Proceedings
of the 2009 Annual Computer Security Applications
Conference (ACSAC ’09), Washington, D.C., USA,
IEEE Computer Society.
Kheir, N., Cuppens-Boulahia, N., Cuppens F. and Debar
H., 2010. A service dependency model for cost-
sensitive intrusion response. In Proceedings of the
15th European Symposium on Research in Computer
Security (ESORICS), Athens, Greece.
Kotenko, I., Stepashkin, M., 2006. Attack Graph based
Evaluation of Network Security. In Lecture Notes in
Computer Science, Vol. 4237, pp.216-227.
Kotenko, I., Stepashkin, M., Doynikova, E., 2011.
Security Analysis of Computer-aided Systems taking
into account Social Engineering Attacks In
Proceedings of the 19th Euromicro International
Conference on Parallel, Distributed and network-
based Processing (PDP 2011), Los Alamitos,
California. IEEE Computer Society, pp.611-618.
Lippmann, R., Ingols, K., 2006. Validating and Restoring
Defense in Depth Using Attack Graphs. In
Proceedings of MILCOM 2006. Washington, DC.
MSM, 2012. Making Security Measurable, viewed 01
March 2012, http://measurablesecurity.mitre.org/
index.html>.
MASSIF, 2012. Massif project, viewed 01 March 2012,
<http://www.massif-project.eu>
Miller, D. R., Harris, Sh., Harper, A. A., VanDyke, S.,
Black, Ch. 2011. Security Information and Event
Management (SIEM) Implementation. McGraw–Hill
Companies. 2011. 430 p.
Moore, A. P., Ellison, R. J., Linger, R. C., 2001. Attack
Modeling for Information Security and Survivability.
Technical Note CMU/SEI-2001-TN-001. Survivable
Systems.
Noel, S., Jajodia, S., O’Berry, B., Jacobs, M., 2003.
Efficient minimum-cost network hardening via exploit
dependency graphs. In Proceedings of the 19th Annual
Computer Security Applications Conference
(ACSAC’03).
NVD, 2012. National Vulnerability Database viewed 01
March 2012, <http://nvd.nist.gov/>
Wang, L., Islam, T., Long, T., Singhal, A., Jajodia, S.,
2008. An attack graph-based probabilistic security
metric. In Proceedings of the 22nd Annual IFIP WG
11.3 Working Conference on Data and Applications
Security (DBSec’08).
Wang, L., Jajodia, S., Singhal, A., Noel, S., 2010. k-Zero
Day Safety: Measuring the Security Risk of Networks
against Unknown Attacks. In ESORICS'10
Proceedings of the 15th European conference on
Research in computer security. Springer-Verlag,
Berlin, Heidelberg, pp.573-587.
Wang, L., Whitley, J. N., Phan, R. C. W., Parish, D. J.,
2011. Unified Parametrizable Attack Tree. In
International Journal for Information Security
Research, Vol.1 (1), pp. 20-26.
SECRYPT2012-InternationalConferenceonSecurityandCryptography
394