Cyber Threat Information Classification and Life Cycle Management
using Smart Contracts
Roman Graf and Ross King
AIT Austrian Institute of Technology, Vienna, Austria
Keywords:
Cyber Security, Situational Awareness, Access and Usage Control, Content Protection, Smart Contracts,
Software Security Assurance.
Abstract:
Nowadays, cyber critical infrastructures (CIs) are increasingly targeted by highly sophisticated cyber attacks
and should be protected. Advances in cyber situational awareness technology lead to the creation of increas-
ingly complex tools. Human analysts face challenges finding relevant information in large, complex data sets,
when exploring data to discover patterns and insights. To be effective in identifying and defeating future
cyber-attacks, cyber analysts require novel tools for incident report classification and life cycle management
that can automatically analyse and share result in secure way between CI stakeholders to achieve better sit-
uation comprehension. Our goal is to provide solutions in realtime that could replace human input for cyber
incident classification and management tasks to eliminate irrelevant information and to focus on important
information to promptly adopt suitable countermeasures in case of an attack. Another contribution relates to
the provided support for document life cycle management that should reduce the number of manual operations
and save storage space. In this paper we evaluate the application of so-called “smart contracts” to an incident
classification system and assess its accuracy and performance. We demonstrate how the presented techniques
can be applied to support incident handling tasks performed by security operation centers (SOCs).
1 INTRODUCTION
The widespread use of cyber security (CS) informa-
tion technologies is considerably increasing the num-
ber of electronic documents. Automated methods for
organizing and improving the access to the informa-
tion contained in these documents become essential to
cyber security information management. This paper
describes a methodology developed to improve infor-
mation organization and access in cyber security in-
formation systems based on automatic classification
of cyber security documents according to their ex-
pected threat level.
Cyber Situational Awareness (SA) (Barford et al.,
2010) provides and overview of a security and threat
situation as well as a current and future impact
assessment. Speed of events, data overload, and
meaning underload (Kott and Wang, 2014) make
real-time situational awareness of cyber operations
very difficult to evaluate. Report data are often
imprecise, which makes it difficult to find relevant
information in large, complex data sets. Novel
techniques that can automatically make obvious or
predefined decisions by means of smart contracts can
help by identifying and defeating of cyberattacks.
Smart contract is a piece of software that fixes
and verifyies negotiated behavior and can not be
manipulated, because it is distributed among multiple
nodes on a blockchain. Another value of using “smart
contracts” is that once uploaded on blockchain
it is working automatically, without the need for
interaction with human. We hypothesize that the
application of “smart contracts” based on existing
blockchain technology (Ethereum (Wood, 2014))
can solve some SA problems. The main purpose
of designing smart contracts for SA is to enable
rapid and trusted cyber incident classification and
management, without the need for a large centralized
authority. We propose that smart contracts based on
decentralised assets such as Ethereum can reduce
effort for securing report transfer, manual analysis
costs, and increase speed of severe information shar-
ing. With raising number of advanced CS tools for
situational awareness, SOC analysts receive a huge
amount of threat reports daily. These experts face
challenges finding relevant information in large, com-
plex data sets and following organisation business
processes, such as proper acquisition, use, archival
and disposal of threat reports. The management
system based on smart contracts and blockchain tech-
nology is aiming at automatic management of threat
reports provided by threat analysis tools, such as
304
Graf, R. and King, R.
Cyber Threat Information Classification and Life Cycle Management using Smart Contracts.
DOI: 10.5220/0006605203040311
In Proceedings of the 4th International Conference on Information Systems Security and Privacy (ICISSP 2018), pages 304-311
ISBN: 978-989-758-282-0
Copyright © 2018 by SCITEPRESS – Science and Technology Publications, Lda. All rights reserved
CAESAIR
1
, IntelMQ
2
or MISP
3
and should provide
effective decision support for SOC operator. Com-
pared to manual classification, automatic classifica-
tion by threat level can significantly facilitate and ac-
celerate reaction time of a SOC analyst. For exam-
ple CAESAIR tool (Settanni et al., 2016) supports
various security information correlation techniques
and provides customizable import capabilities from a
multitude of security-relevant sources. These sources
include a custom repository, open source intelligence
(OSINT) feeds, and IT-security bulletins, as well as a
standardized vulnerability library (Common Vulnera-
bilities and Exposures - CVE). CVEs are especially
important for smart contracts with regard to likeli-
hood assessment based on game theory ((Samarji and
et al, 2015), (Kanoun and et al, 2009)), which is
implementing risk scoring (Reguly, 2013). Employ-
ing CAESAIR with CVE scoring (Maghrabi et al.,
2016) and extending it by automated tagging can pro-
vide valuable inputs for information classification and
management life cycle. Such system can be imple-
mented using smart contracts composed for particular
organization. The research presented here should sup-
port cyber analyst by fast and effective establishing of
a cyber situational awareness. Each institution may
have multiple classification profile definitions depen-
dent on network, CI, and the role of the cyber analyst.
The research presented evaluates a system based
on blockchain and smart contract technology that
will automatically classify and manage cyber inci-
dents that could impact cyber situational awareness
reported and analysed by one of the trusted stake-
holders. Within our prototype system, smart con-
tracts trigger incident classification for large amounts
of data by means of knowledge base employing one of
the incident analysis tools. Smart contracts also man-
age incident life cycle if rules coded in related smart
contract are met.
This paper is structured as follows: Section 2
gives an overview of related work and concepts. Sec-
tion 3 explains the cyber incident classification work-
flow and also covers report life cycle issues. Section
4 presents the experimental setup and applied meth-
ods. Evaluation results are presented in subsection
4.2. Section 5 concludes the paper.
2 RELATED WORK
Multiple researchers are developing an automated
technology that will support an information classifi-
1
http://caesair.ait.ac.at
2
https://github.com/certtools/intelmq
3
https://github.com/MISP/MISP
cation system. An attempt to classify the relation-
ships between documents and concepts (Weng et al.,
2006) employs principles of ontology. To improve
information organization and access in construction
management was developed a methodology (Caldas
and Soibelman, 2003) based on automatic hierarchi-
cal classification of construction project documents
according to project components. A survey on various
cyber attacks and their classification (M. and Padma-
vath, 2013) attempts to develop an ontology for cy-
ber security incidents. They classify by characteris-
tics, such as “organized”, “enormous”, “scrupulously
designed”, “demanding time and resource”, and by
purpose and motivations, such as “obstruction of in-
formation”, “retardation of decision making process”,
“denial in providing public services”, “reputation of
the country will be denigrated”, “smashing up legal
interest”. Additionally cyber attacks can be classified
based on severity of involvement, scope or network
types with multiple sub classification terms. Contrary
to this approach, we classify only by threat level that
can differ from organisation to organsation. Our goal
is to focus human expert resources on the most urgent
incidents important for particular organisation.
An information life-cycle model described in
(Harris and Maymi, 2016) is applicable also to the
cyber security domain. Cyber incident reports are
aquired, analysed and become outdated. Effective au-
tomatic classification, retention and disposal policies
can mitigate risks to data and make information man-
agement more effective. Classification of data enables
a company or security operational center to focus their
resources towards most valuable or urgent incidents
and to handle less valuable incidents automatically
saving time and costs.
An overview of the blockchain technologies and
its potential to facilitate money transactions, smart
contracts design, automated banking ledgers and dig-
ital assets is provided in (Peters and Panayi, 2016).
We suggest that the core technology of this approach
can be reused in the cyber security domain by means
of suitable smart contracts.
A distributed peer-to-peer network based on
blockchain technology where non-trusting members
can interact with each other without a trusted in-
termediary, in a verifiable manner was examined in
(Christidis and Devetsikiotis, 2016) for the Internet
of Things (IoT) sector. This mechanism should work
also for the automation of multi-step processes for cy-
ber incident classification and management.
The performance of Blockchain, which is a prob-
abilistic proof-of-work (PoW) based consensus fab-
ric (Vukoli
´
c, 2016) has became an important issue
for the modern cryptocurrency platforms. PoW-based
Cyber Threat Information Classification and Life Cycle Management using Smart Contracts
305
Blockchains can be replaced by BFT state machine
replication, to improve Blockchain scalability limits.
A Blockchain platform comparison (Macdonald
et al., 2017) discusses five general-use Blockchain
platforms and looks at how Blockchain technol-
ogy can be used in applications outside of Bitcoin
(Nakamoto, 2009) to build custom applications on top
of it. This comparison suggests that Ethereum is cur-
rently the most suitable platform and well established
platform. Therefore, for cyber incident analysis we
employ Ethereum Blockchain in its Pyethereum fash-
ion, which supports focused smart contracts testing
environment without the need of mining.
A basis for smart contracts development in cyber
security realm is a solid threat intelligence that is pro-
vided by a number of cyber incident analysis tools.
The CAESAIR tool (Settanni et al., 2016) introduces
the concept of a cyber intelligence analysis system,
called Collaborative Analysis Engine for Situational
Awareness and Incident Response. CAESAIR pro-
vides analytical support for security experts carrying
out cyber incident handling tasks on a national and
international level, and facilitates the identification of
implicit relations between available pieces of infor-
mation. It provides powerful correlation capabilities,
which support the tasks carried out by the analysts of a
Security Operation Center (SOC) during the incident
handling process. CAESAIR evaluates how the col-
lected documents are connected to one another, and
allows the analyst to select the most appropriate cor-
relation method and to flexibly adjust relevance met-
rics.
The research on risk management in SA in-
creasingly gains in importance. The SA framework
(Morita et al., 2011) describes how a person per-
ceives elements of the environment, comprehends and
projects its actions into the future. This framework
employs the situation awareness model that can be
used in the assessment of risk awareness focusing
on the adverse event notification system. Our ex-
pert system takes a similar approach, but focuses
on classification of essential information, rather than
events. The review of existing situation awareness
measurement techniques for their suitability for use
in the assessment of SA in different environments
(Salmon et al., 2006) demonstrates that current SA
measurement techniques are inadequate by them-
selves for use in the assessment of SA, and a multiple-
measure approach utilising different approaches is
recommended. To address this gap, we employ spe-
cific metrics employing outputs from the threat analy-
sis tools. In security planning, it is necessary to anal-
yse data that are often vague and imprecise. In (Bar-
ford et al., 2010) authors survey existing technologies
in handling uncertainty and risk management in cyber
situational awareness, but the focus is on looking for
vulnerabilities in a system, whereby our approach is
focused on secure classification of the raw data at a
lower level that creates a basis for further SA aspects,
such as situation recognition, situation comprehen-
sion and situation projection (Barford et al., 2010).
In the proposed system we intend to apply smart
contracts for cyber incident classification and life cy-
cle management, which is unique for the given do-
main.
3 CYBER INCIDENT
CLASSIFICATION AND
MANAGEMENT USING SMART
CONTRACTS
We evaluate the application of smart contracts to clas-
sify and manage incident reports sended between CIs
experts in order to improve SA. For instance, smart
contracts can be used to estimate that reported cyber
incident is of high relevance, to remove it after some
predefined time, to tag it by acquisition, to search by
tag, to assign access rights (confidential, private, sen-
sitive, public), to periodic check data integrity (pre-
venting manual or hardware corruption) or to deter-
mine data provenance. Using GUI this approach can
also provide visualisation support - show storage con-
tent of smart contract. One smart contract can call
other smart contract to read from his storage. Our
goal is to save storage place, improve performance
and to keep information up-to-date in a trusted way
leveraging distributed nodes nature of the blockchain
technology. Another goal is to minimize calculations
number and to make secure smart contracts.
Once a smart contract is triggered, the analysis
result is automatically propagated among all partici-
pants through inherent blockchain mechanisms. One
of the advantages of this approach is that smart con-
tracts cannot be changed or compromised without be-
ing detected (through hashed transactions) and that
the messages can be verified to originate from a
trusted source (through public key encryption).
The state of smart contracts is stored on the
blockchain and is transparent and accessible to all
registered community members (see Figure 1). The
smart contract code is executed in parallel by a net-
work of miners under consensus regarding outcome
of the execution. The execution of the smart con-
tract results in an update of the contract’s state on the
blockchain that is synchronized with every participat-
ing node through standard peer-to-peer mechanisms.
ICISSP 2018 - 4th International Conference on Information Systems Security and Privacy
306
USER
REPORTS
INCIDENT
ACQUISITION
STORAGE
BLOCK n BLOCK n+1 BLOCK n+2
USE
STORAGE
ARCHIVAL
STORAGE
BLOCKCHAIN
SMART CONTRACTS
CI1 CI2 CIn
DISPOSAL
STORAGE
Figure 1: The overview of establishing the Cyber Situ-
ational Awareness using Smart Contracts for information
classification and life-cycle management.
By incident acquisition an acquisition smart con-
tract performes classification of a report by threat
level, stores obtained threat level on a blockchain and
initiates life-cycle management process for given in-
cident. In the next steps this report will be used,
archived and disposed.
1. ACQUISITION
2. USE
END
START
THREAT
INTELLIGENCE
TOOLS
4. DISPOSAL
has to be
archived
3. ARCHIVE
no
yes
threat report
1.1 CLASSIFICATION
1.2 TAGGING
1.3 GDPR
2.1 SIMILARITY SEARCH
2.2 STATUS RETRIEVAL
2.3 PROVENANCE RETRIEVAL
2.4 ENRICH
2.5 CHECK DATA INTEGRITY
4.1 REMOVE BY TAG
4.2 REMOVE BY DATE
Figure 2: The workflow for classification and life-cycle
management of cyber incident using smart contracts.
For cyber incident processing we employ four
smart contracts as depicted in Figure 2. The work-
flow execution begins with the reading of an inci-
dent report and parsing of the report content. Input
data, along with expert profile settings that are spe-
cific for an organisation, are passed to the first smart
contract “acquisition”. For the acquisition computa-
tion we employ one of the threat intelligence tools.
By means of threat intelligence tool we obtain cyber
incident content. In the next step we merge the inci-
dent with institutional settings and using smart con-
tract logic to automatically decide which threat level
to assign a given incident to classify it. Classification
occures employing incident text, splitted by words or
phrases, specific terms separated by low, middle and
high threat relevance. The significant terms data pro-
vided by domain experts is stored in a text files. In
the next step we compute risk points counting how
many of threat level terms are included in incident re-
port for each threat level. Finally, we calculate threat
level using one of two methods. Either we estimate
threat level applying threasholds for each level or we
employ weighted method using Formula 1 where we
additionally multiply calculated points on each threat
level with a constant standing for weight of related
threat level.
Domain experts rated the threat level on a scale
of 1-3, where 1 is “low threat” and 3 is “high threat.
Risk points RP are calculated using Formula 1 and
is a sum of high risk points H
rp
multiplied by high
threat weight HT
w
, middle risk points M
rp
multiplied
by middle threat weight MT
w
and low risk points L
rp
multiplied by low threat weight LT
w
.
RP = H
rp
HT
w
+ M
rp
MT
w
+ L
rp
LT
w
(1)
where HT
w
= 3, MT
w
= 2 and LT
w
= 1.
Threat level T
l
can be inferred using high threat
HT
t
and middle threat HT
t
thresholds and weighted
risk points RP from Formula 1 applying Formula 2
T
l
=
3(high) if RP > HT
t
,
2(middle) if RP > MT
t
,
1(low) else RP MT
t
.
(2)
where HT
t
= 10 and MT
t
= 3.
Acquisition step is splitted in different tasks:
SC1.1: Automatic classification by threat level de-
fines one of three threat levels. “high” (3) level re-
quires fast reaction and mediation steps, triage pro-
cess. “medium” (2) level assumes detection of “In-
dicator of Corruption” (IoC) or metrics that indicate
possible vulnerabilities, requires SW update. “low”
(1) level addresses regular cyber security information,
logs, requires attention but should not necessary be a
threat. SC1.2: Tagging means that spcific tags can be
assigned to a report to easy find, shift or remove it
later. SC1.3: Remove personal information (GDPR).
To protect personal data it may be required to remove
personal information from incident report and store
normalized version of incident.
Cyber Threat Information Classification and Life Cycle Management using Smart Contracts
307
In the second step the workflow supports incident
using. Using employs tasks, such as: SC2.1: Auto-
mated similarity search. SC2.2: Status and prove-
nance retrieval. SC2.3: Enrichment with data and
metadata. SC2.4: Periodic check for data integrity
(using hash of incident report - that saves storage
space on blockchain, because instead of storing and
validating of the file content we validate only file
hashes, which are short strings).
Finally, depending on threat level after some pe-
riod of time, incident can be archived (step 3) or re-
moved e.g. by date or by tag (step 4).
We believe that this automatic smart-contracts-
based approach would significantly facilitate incident
classification and management and could be used
by analysts for the defence of critical infrastructure.
The suggested method would make SA analysis less
costintensive and would perform with higher through-
put. However, as is typical in this area, a human-based
approach performs with higher accuracy.
4 EVALUATION
In this section we report on measurements of the auto-
mated cyber incident classification, how long it takes
for smart contracts to be executed and validated. We
carried out measurements for varying incident report
categories and different volumes of data in the knowl-
edge base. The goal of this evaluation was to lever-
age the domain expert knowledge base for cyber in-
cident classification and management as described in
the workflow (see Fig. 2), pointing out threat level
relevant for Situation Awareness.
4.1 Evaluation Data Set
One responsibility of the cyber analyst is to priori-
tise a received cyber incident and to mitigate it or to
carry out a selected cyber incident response. For this
evaluation, we differentiate between high, middle and
low priority. High priority means that the incident has
high severity and mitigation steps should be carried
out. These types of incidents are tagged with a num-
ber 3. Low priority incidents are tagged with a num-
ber 1.
This evaluation took place on an Intel Core i7-
3520M 2.66GHz computer using Python on Ubuntu
OS. We evaluate 5850 cyber incident reports from the
“seclists” feed
4
from the last three years address-
ing four report categories: “fulldisclosure”, “bug-
traq”, “pen-test” and “nmap-dev”. The “fulldis-
closure” category contains messages from public,
4
http://seclists.org/
vendor-neutral forum for detailed discussion of vul-
nerabilities and exploitation techniques, as well as
tools, papers, news, and events of interest to the com-
munity. The “bugtraq” category is a general security
mailing list. The “pen-test” category discloses tech-
niques and strategies that would be useful to anyone
with a practical interest in security and network au-
diting. The “nmap-dev” category comprises an un-
moderated technical development forum for debating
ideas, patches, and suggestions regarding proposed
changes to Nmap
5
and related projects.
The specific cyber security terms were obtained
from the cyber security glossary
6
.
We assume that employing of described smart
contracts approach should classify cyber incidents
among a very large number of incident reports fa-
cilitating further cyber analysis and incident manage-
ment in the future. We also expect that smart contracts
written in Serpent language
7
, supported by Python
2.7 workflow and subsequent analysis will demon-
strate both good performance and sufficient accuracy.
4.2 Experimental Results and
Interpretation
0
50
100
150
200
250
300
350
HIGH THREAT REPORTS
fulldisclosure bugtraq pen-test nmap-dev
Figure 3: Plot for distribution of high threat incident reports
over last three years shared quarterly.
In the test scenario we investigate incident reports
from “seclists” feed to classify them by threat level
and to automatically manage them from acquisition to
disposal without involvement of human analyst (see
Table 1). Due to huge number of results in this table
we describe only selected classification results, which
demonstrate typical cases. The first column contains
categories of the incidents. “FD” is “fulldisclosure”,
“BT” is “bugtraq”, “ND” is “nmap-dev” and “PT”
means “pen-test”. Next three columns show incident
timestamp splittet in year, month and quarter. Graph-
ical results for the whole dataset in Figures 3, 4 and 5
5
https://nmap.org/
6
https://scottschober.com/glossary-of-cybersecurity-
terms/
7
https://github.com/ethereum/wiki/wiki/Serpent
ICISSP 2018 - 4th International Conference on Information Systems Security and Privacy
308
Table 1: Excerpt of classification results for cyber incident reports by their acquisition using smart contracts.
Cat Year M Q Source ID Time L
LT
Terms
M
MT
Terms
H
HT
Terms
SUM TL WTL
SC
ID
FD 2017 Jan q1
Wolfgang
feedyourhead at
75 0.371 1 subject 1 bug 3
attacker
firewall
password
5 3 3 68
FD 2015 Feb q1
Scott
Arciszewski
53 0.370 1 key 2
bug
spam
0 3 2 2 1304
FD 2015 Feb q1
Praveen D
90 0.451 1 response 0 0 1 1 1 1314
BT 2017 Jan q1
Vulnerability
Lab
18 0.677 2
capability
investigation
4
access
authen-
tication
author-
ization
encode
7
alert
attack
attacker
hack
phishing
risk
vulne-
rability
13 3 3 3419
BT 2017 Jun q2
SEC
Consult
Vulnerability
Lab
56 0.532 3
investigation
penetration
work
5
cipher
crypt-
analysis
decrypt
decryption
encryption
5
attack
attacker
risk
signature
vulne-
rability
13 3 3 3829
BT 2016 Jan q1
Slackware
Security
Team
75 0.332 1 key 1
authen-
tication
0 2 1 1 4009
BT 2016 Apr q2
Salvatore
Bonaccorso
158 0.432 0 1 bug 0 1 1 1 4215
ND 2017 Mar q1
Henri
Doreau
226 0.533 2
event
work
1 bug 0 3 2 2 4831
ND 2015 Nov q4
Peter
Houppermans
107 0.600 2
interoperability
penetration
3
authen-
tication
decryption
interope-
rability
3
attack
password
signature
8 3 3 6849
ND 2015 Oct q4
Mark
Scrano
63 0.496 1 bitcoin 0 0 1 1 1 6853
PT 2017 Jul q3
Hafez
Kamal
1 0.252 2
NFC
penetration
1 access 0 3 2 2 7994
PT 2016 Feb q1
Francisco
Amato
2 0.357 1 penetration 1 bug 0 2 1 1 8071
PT 2016 Dec q4
ERPScan
inc
0 0.521 2
penetration
work
1
cyber
security
6
attack
attacker
exploit
impact
threat
vulne-
rability
9 3 3 8072
also splitted in quaters on X axis. Column “Source”
depicts an incident source that can be a person or an
organisation. Incident ID in “seclists” terms is pre-
sented in column “ID”. The columns “L”, “M” and
“H” to the right of the “ID” column present how many
risk points were detected for given incident respec-
tivly. And columns “LT Terms”, “MT Terms” and
“HT Terms” show detected significant threat terms for
low, middle and high threat level. The total number
0
10
20
30
40
50
60
70
80
90
100
MIDDLE THREAT REPORTS
fulldisclosure bugtraq pen-test nmap-dev
Figure 4: Plot for distribution of middle threat incident re-
ports over last three years shared quarterly.
of risk points for each incident can be found in the
column “SUM”. Resulting threat level and weighted
threat level are shown in columns “TL” and “WTL”
respectively. Finally, column “SC ID” contains smart
contract id of incident on the blockchain.
As a use case scenario assume that Security Op-
erational Team has received an incident report from
Vulnerability Lab in January 2017. On receiving
of this report our smart contract triggers automati-
0
50
100
150
200
250
LOW THREAT REPORTS
fulldisclosure bugtraq pen-test nmap-dev
Figure 5: Plot for distribution of low threat incident reports
over last three years shared quarterly.
Cyber Threat Information Classification and Life Cycle Management using Smart Contracts
309
cal analysis and classification of this incident report.
According to Table 1 we see that this incident ob-
tains smart contract identifier SCID 3419 and contract
identifies 2 “LT Terms” (capability, investigation), 4
“MT Terms” (access, authentication, authorization,
encode) and 7 “HT Terms” (alert, attack, attacker,
hack, phishing, risk, vulnerability). Going through
the contract logic we estimate both the regular threat
level and the weighted threat level as a “high threat”
(3). Therefore, our smart contract has automatically
analyzed and classified this incident report as a “high
threat”. That means it should be handled in the first
place with high priority. Incident is automatically
tagged and enriched with additional data from cyber
security feeds and tools like “whois”
8
and “nmap”
9
. Links to similar incidents are established. All this
facilitates Triage process for a cyber analyst and per-
forms analysis steps that usually are done manually.
Due to employment of blockchain technology aggre-
gated information can not be lost or biased. An expert
is able to retrieve incident status or provenance infor-
mation from blockchain at any time. According to
evaluated classification level, smart contract defines
timestamps for automated archival and disposal of in-
cident data. Therefore, a cyber analyst does not need
to care about incident life cicle and can focus her re-
sources on Triage for urgent cases.
The smallest duration for one smart contract op-
eration 0.252 seconds shows report with SCID 7994
and the longest operation time 0.677 report with SCID
3419. This difference can be explained by different
report size (we calculate hash for report content) and
different risk points numbers (3 for SCID 7994 vs. 13
for SCID 3419). We can see that reports are coming
from persons, such as “Scott Arciszewski” or “Henri
Doreau” but also from companies, such as “SEC Con-
sult Vulnerability Lab” or “Slakware Security Team”.
This table also gives easy overview over detected sig-
nificant terms, such as “attack”, “hack”, “phishing”
for high threat incidents, “access”, “authentication”,
“encode” for middle threat incidents and “key”, “ca-
pability”, “investigation” for low level threats. Also
cyber analyst can immediately see automatic calcu-
lated threat level of incident for different calculation
methods. Having smart contract ID analyst is able to
retrieve status data of particular incident report from
blochchain using smart contract (e.g. hash, prove-
nance, time, tags, owner).
The experimental results regarding category
overview are presented in Table 2 that shows distri-
bution of high, middle and low threat level incidents
for different incident categories. This table demon-
8
https://www.whois.com/whois/
9
https://nmap.org/
strates that most incident reports (2429) are coming
from “nmap-dev” category following by “fulldisclo-
sure” (1872) and “bugtraq” (1447) categories. Most
of incident reports belong to the low thrat level (2461)
but report number classified as high threat is also high
(1967). Most high threat level reports are coming
from “fulldisclosure” (724) and “bugtraq” (758) cat-
egories. That means that these categories should be
addressed first by incident management.
The experimental results are visualized in Figures
3, 4 and 5 and show distribution of threat incident re-
ports over last three years respective for high, mid-
dle and low threat levels. Each incident category is
flagged by specific color. X axis is range of the num-
ber of incidents and Y axis is a time scale splitted in
quarters. These figures demonstrate that most produc-
tive category for high (up to 325) and low (up to 215)
threats is a “bugtraq” category, whereas “nmap-dev”
(93) and “fulldisclosure” (97) are dominating for mid-
dle thrat reports. For given period of time most active
phase for all levels is from “Q4-2015” to “Q3-2016”.
Visualization of incident reports provides an analyst
with fast and descriptive situational awareness pic-
ture. To focus on particular part analyst can perform
fine tuning and adjust time scale or select particular
category or source.
Table 2: Overview about aggregated threat reports for dif-
ferent threat categories.
Threat category High Threat Middle Threat Low Threat Total
fulldisclosure 724 558 590 1872
bugtraq 758 147 542 1447
pen-test 55 43 4 102
nmap-dev 430 674 1325 2429
Sum 1967 1422 2461 5850
These results demonstrate that a semi-automatic
approach for incident classification and visualisation
is very effective and it is a signicant improvement
compared with manual analysis for monitoring and
validation of design for critical infrastructure. Result-
ing actions of the presented analysis tool may be mit-
igation steps, validation, checking or updating of the
software.
5 CONCLUSIONS
In this work we have presented an automated ap-
proach to classify and manage incident reports for
establishing cyber situational awareness using smart
contracts. We have combined expertise gathered dur-
ing the development of a cyber intelligence tool with
the power of the smart contracts approach. The main
contribution of this work is a real-time solution that
could replace human input for a large number of cyber
ICISSP 2018 - 4th International Conference on Information Systems Security and Privacy
310
incident analysis tasks in order to eliminate irrelevant
information and to focus on important information to
promptly perform mitigation steps. Another contribu-
tion is the employment of smart contract techniques
to provide an automated trusted system for incident
management life-cycle that allows automatic acquisi-
tion, classification, use, archiving and disposal. The
presented method employs a domain expert knowl-
edge base collected through a cyber intelligence tools
to detect Situational Awareness risks. An additional
advantage of this approach is a reduction of human
analysis costs. Ultimately, our research will lead to
the creation of automated security assessment tools
with more effective handling of cyber incidents.
REFERENCES
Barford, P., Dacier, M., Dietterich, T., Fredrikson, M., Gif-
fin, J., Jajodia, S., Jha, S., Li, J., Liu, P., Ning, P., Ou,
X., Song, D., Strater, L., Swarup, V., Tadda, G., Wang,
C., and Yen, J. (2010). Cyber sa: Situational aware-
ness for cyber defense. In Jajodia, S., Liu, P., Swarup,
V., and Wang, C., editors, Cyber Situational Aware-
ness, volume 46 of Advances in Information Security,
pages 3–13. Springer US.
Caldas, C. H. and Soibelman, L. (2003). Automating hier-
archical document classification for construction man-
agement information systems. Automation in Con-
struction, 12(4):395 – 406.
Christidis, K. and Devetsikiotis, M. (2016). Blockchains
and smart contracts for the internet of things. IEEE
Access, 4:2292–2303.
Harris, S. and Maymi, F. (2016). Cissp all-in-one exam
guide. CISSP book, seventh edition, chapter 2, pages
189 – 245.
Kanoun, W. and et al (2009). Success likelihood of ongoing
attacks for intrusion detection and response systems.
volume 3, pages 83–91. IEEE.
Kott, A. and Wang, C. (2014). Cyber Defense and Situa-
tional Awareness. Springer International Publishing
Switzerland.
M., U. and Padmavath, G. (2013). A survey on various cy-
ber attacks and their classification. nternational Jour-
nal of Network Security, 15(5):390–396.
Macdonald, M., Liu-Thorrold, L., and Julien, R. (2017).
The blockchain: A comparison of platforms and their
uses beyond bitcoin. The University of Queensland.
Maghrabi, L., Pfluegel, E., and Noorji, S. F. (2016). Design-
ing utility functions for game-theoretic cloud security
assessment: a case for using the common vulnerabil-
ity scoring system. In 2016 International Conference
On Cyber Security And Protection Of Digital Services
(Cyber Security), pages 1–6.
Morita, P., Burns C.M., H., and He, Q. (2011). Situa-
tion awareness and risk management understanding
the notification issues. In Stud Health Technol Inform.
PubMed.
Nakamoto, S. (2009). Bitcoin: A peer-to-peer electronic
cash system.
Peters, G. W. and Panayi, E. (2016). Understanding Mod-
ern Banking Ledgers Through Blockchain Technolo-
gies: Future of Transaction Processing and Smart
Contracts on the Internet of Money, pages 239–278.
Springer International Publishing, Cham.
Reguly, T. (2013). Does anybody really care
about vulnerability scoring? [online],
Available: https://www.tripwire.com/stateof-
security/risk-based-security-for-executives/risk-
management/doesanybody- really-care-about-
vulnerability-scoring/.
Salmon, P., Stanton, N., Walker, G., and Green, D. (2006).
Situation awareness measurement: A review of appli-
cability for {C4i} environments. Applied Ergonomics,
37(2):225 – 238.
Samarji, L. and et al (2015). Coordination and concurrency
aware likelihood assessment of simultaneous attacks.
volume 153, pages 524–529.
Settanni, G., Shovgenya, Y., Skopik, F., Graf, R., Wurzen-
berger, M., and Fiedler, R. (2016). Correlating cyber
incident information to establish situational awareness
in critical infrastructures. In 2016 14th Annual Con-
ference on Privacy, Security and Trust (PST), pages
78–81, Auchland, New Zealand. IEEE.
Vukoli
´
c, M. (2016). The Quest for Scalable Blockchain
Fabric: Proof-of-Work vs. BFT Replication, pages
112–125. Springer International Publishing, Cham.
Weng, S.-S., Tsai, H.-J., Liu, S.-C., and Hsu, C.-H. (2006).
Ontology construction for information classification.
Expert Systems with Applications, 31(1):1 – 12.
Wood, G. (2014). Ethereum: A secure decentralised gener-
alised transaction ledger.
Cyber Threat Information Classification and Life Cycle Management using Smart Contracts
311