ArticlePDF Available

Dark-TRACER: Early Detection Framework for Malware Activity Based on Anomalous Spatiotemporal Patterns



As cyberattacks become increasingly prevalent globally, there is a need to identify trends in these cyberattacks and take suitable countermeasures quickly. The darknet, an unused IP address space, is relatively conducive to observing and analyzing indiscriminate cyberattacks because of the absence of legitimate communication. Indiscriminate scanning activities by malware to spread their infections often show similar spatiotemporal patterns, and such trends are also observed on the darknet. To address the problem of early detection of malware activities, we focus on anomalous synchronization of spatiotemporal patterns observed in darknet traffic data. Our previous studies proposed algorithms that automatically estimate and detect anomalous spatiotemporal patterns of darknet traffic in real time by employing three independent machine learning methods. In this study, we integrated the previously proposed methods into a single framework, which we refer to as Dark-TRACER , and conducted quantitative experiments to evaluate its ability to detect these malware activities. We used darknet traffic data from October 2018 to October 2020 observed in our large-scale darknet sensors (up to /17 subnet scales). The results demonstrate that the weaknesses of the methods complement each other, and the proposed framework achieves an overall 100% recall rate. In addition, Dark-TRACER detects the average of malware activities 153.6 days earlier than when those malware activities are revealed to the public by reputable third-party security research organizations. Finally, we evaluated the cost of human analysis to implement the proposed system and demonstrated that two analysts can perform the daily operations necessary to operate the framework in approximately 7.3 h.
Received December 23, 2021, accepted January 18, 2022, date of publication January 25, 2022, date of current version February 4, 2022.
Digital Object Identifier 10.1109/ACCESS.2022.3145966
Dark-TRACER: Early Detection Framework for
Malware Activity Based on Anomalous
Spatiotemporal Patterns
1National Institute of Information and Communications Technology, Koganei 184-8795, Japan
2Graduate School and Faculty of Information Science and Electrical Engineering, Kyushu University, Fukuoka 819-0395, Japan
Corresponding author: Chansu Han (
This work was supported by the Ministry of Internal Affairs and Communications, Japan, conducted under a Contract of MITIGATE
among Research and Development for Expansion of Radio Wave Resources (JPJ000254).
ABSTRACT As cyberattacks become increasingly prevalent globally, there is a need to identify trends in
these cyberattacks and take suitable countermeasures quickly. The darknet, an unused IP address space,
is relatively conducive to observing and analyzing indiscriminate cyberattacks because of the absence of
legitimate communication. Indiscriminate scanning activities by malware to spread their infections often
show similar spatiotemporal patterns, and such trends are also observed on the darknet. To address the
problem of early detection of malware activities, we focus on anomalous synchronization of spatiotemporal
patterns observed in darknet traffic data. Our previous studies proposed algorithms that automatically
estimate and detect anomalous spatiotemporal patterns of darknet traffic in real time by employing three
independent machine learning methods. In this study, we integrated the previously proposed methods into a
single framework, which we refer to as Dark-TRACER, and conducted quantitative experiments to evaluate
its ability to detect these malware activities. We used darknet traffic data from October 2018 to October
2020 observed in our large-scale darknet sensors (up to /17 subnet scales). The results demonstrate that the
weaknesses of the methods complement each other, and the proposed framework achieves an overall 100%
recall rate. In addition, Dark-TRACER detects the average of malware activities 153.6 days earlier than when
those malware activities are revealed to the public by reputable third-party security research organizations.
Finally, we evaluated the cost of human analysis to implement the proposed system and demonstrated that
two analysts can perform the daily operations necessary to operate the framework in approximately 7.3 h.
INDEX TERMS Anomalous synchronization estimation, darknet, malware activity, spatiotemporal pattern.
In recent years, an increasingly large number of indiscrim-
inate cyberattacks have been observed on the Internet, and
it is therefore becoming increasingly costly to analyze these
attacks. To maintain security of the Internet, it is necessary
to quickly recognize global cyberattack trends, specify their
causes, devise countermeasures, and alert the world of the
details of the threat. For this purpose, it is important to
detect the indiscriminate scanning attack activities caused by
The associate editor coordinating the review of this manuscript and
approving it for publication was Kashif Saleem .
malware at an early stage before a particular attack becomes
a pandemic.
However, it is challenging to identify malware scanning
attacks among the massive amount of benign traffic in regular
networks. Therefore, we adopted unused IP address spaces
(darknets). The term ‘‘darknet’’ refers to observation net-
works, also known as ‘‘network telescopes,’’ and should not
be confused with anonymous communication networks such
as Tor. In the darknet, legitimate communication (noise)
does not occur; therefore, indiscriminate scanning commu-
nication (signal) is observed more noticeably. Thus, the
signal-to-noise ratio is high. This makes it an effective way
to identify trends and tendencies in global cyberattacks.
13038 This work is licensed under a Creative Commons Attribution 4.0 License. For more information, see VOLUME 10, 2022
C. Han et al.:Dark-TRACER: Early Detection Framework for Malware Activity
However, the volume of traffic observed in the darknet is
increasing each year exponentially. Moreover, there are many
communications whose intentions are unknown, as only
the initial communications are observed. For example, in a
darknet, we observe numerous independent cyberattacks
occurring simultaneously, as well as many communications
that are unrelated to attacks, such as scanning activities that
are conducted for benign investigation purposes, communica-
tions with unknown causes, and misconfigured communica-
tions. As a research target, we should distinguish such noisy
communications from malicious attack communications in
Devices infected with similar malware, that is, ones which
share scanning modules, tend to scan in a similar spa-
tiotemporal pattern to compromise new infection targets [1].
Such a tendency is also observed on the darknet [2]. Here,
the distributions of source hosts and destination ports for
packets observed in a certain period are referred to as spatial
features. The features observed in the temporal variation of
these spatial features are thus referred to as spatiotemporal
patterns. The hosts and destination ports that send packets
with similar spatiotemporal patterns are then referred to as
being synchronized. Even in case of small-scale infection
activity of malware, a high degree of synchronicity is expected
to occur in the associated spatiotemporal patterns, and early
detection of malware activity can be realized by estimating
the synchronicity and detecting anomalies.
In our previous studies, we focused on such syn-
chronization and attempted to detect potential malware
activities by estimating the group of source hosts with
high synchronization in their spatiotemporal patterns on
a large-scale darknet. We adopted the following three
different machine learning methods in this study: Graphical
Lasso [3], nonnegative matrix factorization (NMF) [4],
and nonnegative Tucker decomposition (NTD) [5] to esti-
mate the synchronization of spatiotemporal patterns from
packet counts by spatial feature per unit time in darknet
traffic data. The Graphical Lasso algorithm can sparsely
estimate conditionally independent variable pairs that are
not synchronous from a covariance matrix. The NMF and
NTD algorithms can decompose synchronous latent frequent
patterns from data matrices or tensors into superpositions
of multiple groups. We previously proposed the following
different methods to estimate the synchronization in real
time to automatically use the aforementioned algorithms and
detect the source host space groups that show abnormal
synchronization: Dark-GLASSO [6], [7], Dark-NMF [8], and
Dark-NTD [9].
In our previous studies, we confirmed that each method is
capable of detecting malware activities well. However, we did
not comparatively evaluate the methods and examine their
early malware activity detection performance. In this study,
we first modularized the three previously proposed methods
and integrated common components such as feature extrac-
tion and alert issuing into a single framework. We refer to this
integrated framework Dark-TRACER. As the main challenge,
we conducted two experiments on Dark-TRACER—one is
to evaluate the quantitative detection performance, and the
other is to evaluate the feasibility of early detection. In the
first experiment, to quantitatively evaluate the detection
performance of malware activity, we used the ground truth of
reliable malware activity in October 2018, which was man-
ually created, and performed parameter tuning to minimize
false negatives and false positives in each module. Although
we have previously presented the evaluation results of a
conventional method ChangeFinder [10] and the proposed
modules Dark-GLASSO and Dark-NMF, we evaluate Dark-
NTD for the first time using the same criteria. In the second
experiment, we manually generated a new ground truth of
events (from June 2019 to October 2020) that clearly shows
the time of infection spread of malware activities and used it
to evaluate the feasibility of the proposed framework for early
As a result, Dark-GLASSO,Dark-NMF, and Dark-NTD
achieved 97.1%, 100%, and 97.1% recall, respectively.
We also identified the pros and cons of each module and
found that the integration of all the proposed modules into
a single framework, Dark-TRACER, complemented each
individual module’s weaknesses. In addition, the results of
the early detection feasibility evaluation show that Dark-
TRACER can detect threats 153.6 days earlier than when the
threats were revealed to the public by reputable third-party
security research organizations. We also assessed the human
analysis cost and found that daily operation with two analysts
could be performed in an average of 7.3 h, assuming that one
analyst requires 15 min of analysis time per port.
In summary, this study afforded the following contributions:
We integrated our three prior methods (modules) into
a single framework, Dark-TRACER. To the best of our
knowledge, our approach is the first method that focuses
on the synchronization of spatiotemporal patterns of
the darknet traffic. Dark-TRACER can detect malware
activities that show anomalous synchronization.
This work is also the most advanced practical study that
quantitatively evaluated the detection performance of
malware activities and the feasibility of early detection.
We found that Dark-TRACER complements the weak-
nesses of each module, and achieves a 100% recall
rate. In addition, the results demonstrate that Dark-
TRACER detects threats on average 153.6 days earlier
than when the threats are revealed to the public.
We also demonstrated that two analysts can conduct
the necessary daily operations of the framework in
approximately 7.3 h.
Currently, Dark-TRACER is being implemented in real-
world contexts for actual operation. It is expected to provide
information on detected global malware activities to orga-
nizations such as the Computer Security Incident Response
Team (CSIRT) and the Security Operation Center (SOC), and
to assist in their ability to implement prompt countermeasures
such as investigating the causes and conducting detailed
VOLUME 10, 2022 13039
C. Han et al.:Dark-TRACER: Early Detection Framework for Malware Activity
FIGURE 1. Illustration of the framework of Dark-TRACER.
The remainder of the paper is organized as follows. The
proposed framework, Dark-TRACER, and its three modules
are presented in Section II. In Section III, we present the
methodology and results from the parameter tuning and quan-
titative evaluation experiments on the detection performance
of malware activities for each module. Section IV describes
the feasibility evaluation of the proposed method for the early
detection of malware activities. In Section V, we discuss
the advantages of Dark-TRACER through a comprehensive
comparison of each proposed module, consideration of the
likely adversarial attacks, ways to reduce false-positive alerts,
and practical operation methods. Section VI provides a
summary of related work on darknet measurement analysis,
malware activity detection, and investigative scanners, and
we conclude the paper in Section VII.
The overall framework of Dark-TRACER is shown in Fig. 1.
Three algorithms, Graphical Lasso [3], NMF [4], and
NTD [5], are used to estimate the synchronicity of spatiotem-
poral features, and the modules which incorporate these
algorithms are referred to as Dark-GLASSO [6], [7], Dark-
NMF [8], and Dark-NTD [9], respectively, to distinguish
The following advantages over existing malware activity
detection methods can be achieved by focusing on syn-
chronicity: 1) We can reduce the effect of benign noise
communication in the darknet traffic and highlight the
malicious communication. 2) In addition, malware activities
that are difficult to trace by conventional manual operations,
such as threats that are small-scale, orchestrated, or have no
visible explicit spikes, can be captured before the malware
infection becomes widespread by detecting anomalously
synchronized spatial features. 3) Finally, if a malware activity
is found to be synchronized with other malware activities at a
time when the scale of infection is small (i.e., before it spreads
in earnest), it can be detected at that early stage.
The pseudocode of Dark-TRACER framework is presented
in Algorithm 1. The parameters are described in this
section and Section III-C. For more specific details of
the three algorithms employed in this study, the original
paper reporting on each algorithm [3]–[5] or our previous
works [6]–[9] may be referred to. Based on Fig. 1 and
Algorithm 1, the modules are described in greater detail in
Algorithm 1.
Dark-TRACER targets darknet traffic data for analysis.
As mentioned previously, the darknet has the advantage of
a high signal-to-noise ratio, because regular communica-
tion (noise) is not typically observed there, and indiscriminate
scanning communication (signal) is monitored in abundance.
However, not all communications that are observed in the
darknet are malicious communications caused by malware.
Among the totality of communications observed in the
darknet, some communications are not related to attacks,
such as scanning activities for investigation purposes, such as
Shodan and Censys [11],1unexplained communications,
and misconfigured communications. Dark-TRACER is a
framework that detects intrinsic attacks and malware activi-
ties by ignoring and eliminating such noisy communications.
We have implemented a large-scale darknet observation
system, the NICTER project,2which aims to understand
global trends in indiscriminate cyberattacks. Darknet obser-
vation systems (sensors) have been installed in several
countries and organizations, and approximately 300,000
IP addresses are currently being monitored. The observed
data of these darknet sensors differ slightly depending
1 and
13040 VOLUME 10, 2022
C. Han et al.:Dark-TRACER: Early Detection Framework for Malware Activity
Algorithm 1 The Framework of Dark-TRACER
Require: Common: T,M,t,sensor // Dark-GLASSO:γ, λ , K, θ //
Dark-NMF:r, α, β, f// Dark-NTD:˜
Ensure: Alerts
1: while Every tseconds in darknet sensor do
/* Data Observation (Section II-A) */
2: Darknet traffic data for Ts is newly updated, then preprocess it.
/* Spatiotemporal Feature Extraction (Section II-B) */
3: Generate VhNM×Nh
0,Vhp NM×Nh×Np
/* Algorithms and Anomaly Detection (Section II-C, II-D) */
/** Dark-GLASSO **/
4: if Nh> γ then Vhrandom_sampling(Vh), Nhγend if
5: Precision matrix 61
λRNh×Nhgraphical_lasso(Vh, λ)
6: dλRgraph_density(61
7: outliers1anomaly_detection(dλ,K, θ) in Dark-GLASSO
/** Dark-NMF **/
8: WRM×r,HRr×NNMF(Vhor Vp,r)
9: outliers2anomaly_detection(W, α, β, f) in Dark-NMF
/** Dark-NTD **/
10: outliers3NULL
11: for epochs do
12: G,A(1),A(2) ,A(3) LRA-NTD(Vhp,˜
13: outliers3outliers3|
anomaly_detection(G,A(2),A(3) ,th) in Dark-NTD
14: end for
/* Issuing Alerts (Section II-E) */
15: outliers outliers1|outliers2|outliers3
16: Alerts issuing_alerts(outliers)
17: end while
on their geographical location and the scale of observa-
tion. For this reason, Dark-TRACER analyzes each sensor
Next, as a data preprocessing step, Dark-TRACER analyzes
only TCP-SYN packets because TCP packets other than TCP-
SYN that reach the darknet are not considered to be attack
scans. In addition, the upper 16 bits of the IP address are
adopted as the unit of the source host. This means that
hosts are aggregated on a regional or organizational level.
Finally, to highlight the observation of unknown malware
activities, we excluded well-known and frequently observed
threat ports.
First, we prepared darknet traffic data for a certain period
(Tseconds). We assumed that Nhunique numbers of source
hosts and Npunique numbers of destination ports were
observed in the darknet traffic data. Then, at a sampling
interval of T/Mseconds, the number of packets was counted
for each source host or destination port, and these are
referred to as the spatial feature variables. Here, Mis a
hyperparameter. From the above, three types of tensors
representing spatiotemporal features were generated from
the observed data: VhNM×Nh
0, and
Vhp NM×Nh×Np
0(N0= {0,1,2,· · · }). This feature
extraction was processed in real time and sequentially every
This section briefly introduces the main characteristics of
the sparse structure learning algorithm Graphical Lasso [3]
and the two tensor decomposition algorithms, Nonnegative
Matrix Factorization (NMF) [4] and Nonnegative Tucker
Decomposition (NTD) [5]. In addition to the above algo-
rithms, Dark-TRACER can be applied with other methods
to estimate the synchronization of spatiotemporal features,
but an anomaly detection method that is appropriate for the
method must be considered.
The Graphical Lasso (package name: glasso3) algorithm is
a sparse structure learning method that can calculate the
‘‘intrinsic relationships’’, rather than spurious correlations,
between variables. Here, ‘‘no intrinsic relationship’’ between
two variables is equivalent to the conditional independence
of the two variables given the other variables. In a Gaussian
graphical model, which is a structural learning model that
assumes a multivariate Gaussian distribution, the above
problem can be considered as the problem of estimating
a precision matrix (i.e., an inverse covariance matrix).
Graphical Lasso uses maximum likelihood estimation with
a`1regularization term to obtain a sparse precision matrix,
thereby introducing sparsity into the relationship between
The obtained precision matrix can be represented as an
undirected graph, as shown in the Graphical Lasso section
of Fig. 1. The node set represents the set of variables,
and the edge set represents the ‘‘presence or absence of a
relationship’’ between the variables. In other words, when
there is no relationship between variables, no edges are
drawn between the nodes corresponding to those variables.
Alternatively, if there is a relationship, an edge is drawn.
Graphical Lasso has often been applied to the field of
anomaly detection. Graphical Lasso has been applied to
a wide range of real-world problems, such as outlier
detection [12], [13] based on the relationship of the candidate
outlier with the other variables, such as in Dark-GLASSO, and
for detecting changes in a graph structure [14], [15].
Dark-GLASSO uses Graphical Lasso to estimate and graph
the intrinsic relationship between spatial feature variables
from a spatiotemporal feature matrix (Vhor Vp). This can
be interpreted as a representation of the synchronization
between the variables.
Tensor decomposition is a method of decomposing latent
frequent patterns from a matrix or tensor into a super-
position of multiple groups. Several models have been
proposed depending on the rank of the tensor and the
decomposition method. Tensor decomposition has been
VOLUME 10, 2022 13041
C. Han et al.:Dark-TRACER: Early Detection Framework for Malware Activity
applied in a variety of fields, such as recommendation
systems in the marketing domain [16], feature extraction in
electroencephalograms [17], image classification [18], and
foreground filtering and activity detection in videos [19].
The tensor data generated in Section II-B does not accept
negative values. To make the decomposition results realistic
and interpretable, we employed the tensor decomposition
methods NMF and NTD with non-negative constraints. NMF
is a decomposition method for rank-2 tensors (matrices),
whereas NTD is a decomposition method for rank-Dtensors
(in this case, D=3). NTD can be regarded as an extension
of NMF to higher dimensions. Next, the application process
of the method is briefly explained for each module.
As shown in the NMF part of Fig. 1, NMF is a method
of approximate decomposition of a non-negative matrix
0into a product of two non-negative factor
matrices WRM×r,HRr×N(VWH). Here,
ris the number of basis vectors, and refers to the number
of patterns to be decomposed (r<N,M). The NMF
minimizes the error function ||VWH||2
0,H0) (Frobenius norm). Although several approximate
decomposition algorithms have been proposed for NMF,
we employed the most well-known multiplicative update
algorithm proposed by Lee et al. [4]. In this algorithm, the
initial values of W,Hare given, and the optimization is
performed by alternately updating W,Huntil the error
function is minimized. In Dark-NMF, the values of the
singular value decomposition were taken as the initial values.
In summary, Dark-NMF approximately decomposes a
spatiotemporal feature matrix (Vhor Vp) into two factor
matrices W,Husing NMF. The decomposed matrices
contain the same number of potentially synchronous groups
of spatiotemporal feature variables as the number of bases.
Each component of the two decomposed matrices W,Hcan
be interpreted as follows:
W:Temporal features. Each basis vector represents a
temporal traffic pattern of a different type.
H:Spatial features of the source host or destination
port. The source hosts or the destination port
numbers corresponding to the indices of each
basis vector are presumed to have synchronous
As shown in the NTD part of Fig. 1, NTD decomposes a rank-
Dtensor into one small tensor and several matrices. Dark-
NTD works with rank-3 tensors and decomposes them each
into one small tensor Gand three matrices A(1),A(2) ,A(3).
The tensor VRI1×I2×I3can be decomposed using the
decomposition equation as VG×1A(1) ×2A(2) ×3A(3).
Here, A(n)RIn×Rn(n∈ {1,2,3}),GRR1×···×R3, the
scalars I1,I2,I3denote the length of each axis (mode), and
the symbol ×idenotes the product in the mode-idirection,
which multiplies the tensor by the matrix in mode i. The factor
matrix A(n)can be regarded as an extracted feature for mode
n. The tensor Gis referred to as a core tensor and represents
the weights of the basis vectors of each mode and the strength
of the relationship. In addition, R1,R2,R3are the ranks,
which determine how many basis vectors are extracted for
each mode and can be interpreted as the number of frequent
The NTD algorithm minimizes the error function ||V
G×1A(1) ×2A(2) ×3A(3)||2
F(Frobenius norm). The
optimization is performed by alternately updating Gand A(n).
However, when the tensor to be decomposed, V, becomes
large, the decomposition becomes practically impossible
because a massive amount of memory and computation
is required to perform the exact calculation. To address
this problem, Dark-NTD utilizes the fiber sampling ten-
sor decomposition (FSTD) [20] to perform a low-rank
approximation of the tensor Vin advance. Based on the
results of this low-rank approximation, LRA-NTD [21], which
efficiently approximates NTD, is applied to save memory
and accelerate the decomposition process without reducing
its precision. For details of this acceleration, please refer to
Kanehara et al.’s previous paper on Dark-NTD [9].
In summary, Dark-NTD utilizes the accelerated NTD algo-
rithm to decompose the rank-3 spatiotemporal feature tensor
Vhp into a core tensor Gand three matrices A(1),A(2) ,A(3).
The results of the approximate decomposition contain
the same number of potentially synchronous groups of
spatiotemporal feature variables as the number of bases in
each matrix. Each component of this decomposition result
can be interpreted as follows:
A(1):Temporal features. Each basis vector represents a
temporal traffic pattern of a different type.
A(2):Source host spatial features. The source hosts
corresponding to the indices of each basis vector are
assumed to be synchronized and in the same group.
A(3):Destination port spatial features. The port numbers
corresponding to the indices of each basis vector are
assumed to be received from the same group.
In this section, we present a method for detecting anomalies
in spatial feature variables based on the application results of
each algorithm, module by module.
From a graph of the calculated precision matrix, the degree
of synchronization between the variables is quantified by
the graph density |E|/N(N1). Here, |E|is the number
of elements in the edge set, and Nis the number of spatial
feature variables. The closer the graph density is to 1, the
more strongly all variables are related to each other. Graph
density is also referred to as its sparsity. The graph density
value is calculated from observation data every Tseconds
in a continuous period and is recorded sequentially. Outlier
13042 VOLUME 10, 2022
C. Han et al.:Dark-TRACER: Early Detection Framework for Malware Activity
detection is performed when time-series data of the graph
density value are collected for a period of fixed window size
(K). First, two variances are calculated: one when the largest
element in the time-series data is excluded and the other
when it is not excluded. If the ratio of the two variances
exceeds a threshold θ, it is considered an outlier and is
deleted from the time-series data. The outliers are identified
successively by the next largest element until they no longer
exceed the threshold θ. If there are no outliers and the data
size exceeds the fixed window size (K), the oldest data are
deleted in chronological order. From the above, it is possible
to determine the periods which have anomalous graph density
values as compared to the other periods.
Because the scales of the matrices W,Hare not unique,
we first normalize them. To ensure that the sum of each
column of Wis aligned to 1, we normalize W,Has W=
W3, H=31Husing a diagonal matrix 3Rr×r,
whose diagonal component is an inverse of the sum of each
column. The elements of the normalized Hare scaled to
actual observed packet counts. Spatial features with values
of Hless than 1 are considered inactive features in the
corresponding basis vector. For such active spatial features,
if there are more than βfeatures that exceed α(%) of the
maximum value of elements (maximum number of packets),
active spatial features are judged as anomalies. In addition,
when judging anomalous spatial features, a parameter f
{0,1}is utilized to determine whether to treat all active spatial
features or only those that are more anomalous. From the
above, we can determine the anomalous spatial features of
a specific period.
If there are two or more host spatial features that exceed the
threshold value for A(2), the group of hosts is considered
to have synchronized activities and its IP addresses are
recorded. In addition, Gand A(3) are utilized to identify the
destination port features through which this group of hosts
communicated. From G, we identify a port group of A(3) that
is linked to a group of hosts that have been determined to
have synchronous activity from A(2). In the identified port
group, the destination port features that exceed the threshold
are determined to be the targeted ports in the synchronized
host group of A(2). We can determine the anomalous host
groups and their targeted ports in a specific period.
The final process collects information that has been deter-
mined to be anomalous from each module and outputs an
alert in a uniform format. For Dark-GLASSO, we used
the entire darknet traffic data for a period that has been
identified as anomalous. For Dark-NMF, we used the data
for the spatial features identified as anomalous. If numerous
source hosts sent many packets to a specific destination port,
we aggregated the information regarding the time, destination
port numbers, and source hosts and issued an alert. Finally,
Drak-NTD issued an alert directly using the anomalous host
group, targeted port, and time information found in the
anomaly detection step.
We evaluated the performance of each proposed module and
describe the results of two different experiments to demon-
strate the relationships between modules and their practical-
ity. In the first experiment in this section, we quantitatively
evaluated the ability of each module to accurately detect
malware activities. In Section IV, the second experiment
evaluated the feasibility of the early detection of malware
activities. Darknet traffic was preprocessed using tcpdump
and passed to Dark-TRACER, implemented in the Rlanguage.
All experiments were conducted in a unified manner in Japan
Standard Time, with CPUs running on AMD RYZEN TR
2990WX and 128GB memory.
In this experiment, we manually gathered TCP ports for
which malware activities were clearly observed in October
2018 and generated the ground truth for a total of 35 TCP
ports. This ground truth evaluation aimed to determine a
hyperparameter set that minimized the number of false
negatives, even if there were some false positives in each
module, and evaluated the detection accuracy at that time.
The conventional method, ChangeFinder, and the proposed
modules, Dark-GLASSO and Dark-NMF, have already been
tested and the results of those evaluations have been
published [7], [8], whereas Dark-NTD was now evaluated
for the first time using the same criteria. The following
subsection describes the details of the dataset, the parameter
tuning of Dark-NTD, and the comparison results from each
The dataset and the ground truth for evaluation were the
same as those used in the previous reports for Dark-GLASSO
and Dark-NMF and are publicly available.4Specifically,
we employed data from eight darknet sensors Ato H, which
are located around the world and have different observa-
tion scales. The observation scale of each sensor ranges
from approximately 30,000 IP addresses (/17 subnet) to
approximately 2,000 IP addresses (/21 subnet), for a total of
approximately 80,000 IP addresses in the darknet observation
network. The period of data used in the experiment was in the
month of October in 2018. The average number of packets
per day for sensor A, which has the largest observation scale,
was 81.4 M, and the data size was 5,605 MB. To highlight
the observation of unknown malware activities, the following
11 known and constantly observed TCP ports were excluded
during preprocessing: 22, 23, 80, 81, 445, 1433, 2323, 3389,
5555, 8080, 52869.
VOLUME 10, 2022 13043
C. Han et al.:Dark-TRACER: Early Detection Framework for Malware Activity
Next, regarding details of the ground truth, Table 1 shows
the TCP ports where malware activities were monitored
and the characteristics of the malware activities by threat
type. The threat types were primarily classified into Internet
of things (IoT) malware such as Mirai,Hajime, and
HNS (Hide and Seek), vulnerabilities related to router
manufacturers, and vulnerabilities related to other off-the-
shelf service protocols.
As a fingerprint, or key identifier, of Mirai, it is
commonly known that the sequence number matches the
destination IP address in the SYN packet [22], [23].
A fingerprint of Hajime is that its window size is fixed at
14600, and an upper or lower 1 byte of the sequence number
is 0. A common feature of router vulnerability types is that
there are many cases wherein each router manufacturer’s
login screen was confirmed when an HTTP connection was
made to the source hosts that sent the scan. Cohen et al. [24]
also identified that ports 5379, 6379, and 7379 were observed
during the same period. Please refer to that previous paper for
more details, including time-series graphs of the unique host
counts of malware activities in this ground truth.
In this section, we describe how to tune the following five
hyperparameters in the Dark-NTD.
1) sensor: which darknet sensor is used for the observed
2) ˜
Rn: Number of bases in FSTD, a low-rank approxima-
tion method for acceleration.
3) Rn: Number of bases in NTD.
4) epochs: How many times the calculation for the same
data is repeated
5) th: Threshold for alert determination
The above five hyperparameters are tuned by grid search.
The search range and interval include our long-term empir-
ical rules. For the sensor, we compared the performances
of selecting one of the eight darknet sensors against the use
of all eight sensors. Next, the larger the number of bases ˜
FSTD, the better is the low-rank approximation of the original
tensor. In addition, ˜
Rnshould be set to be larger than Rn,
the number of bases in NTD. In this grid search, we worked
within the range of ˜
Rn∈ {25,49,81,121}and Rn∈ {3,5,8}.
Furthermore, because the initial values of FSTD and NTD
are randomly chosen, the calculation results are not unique.
Therefore, we need to know how many times the same data
can be iterated to obtain a stable and sufficient accuracy.
In this tuning experiment, we iterated epochs up to 15 times.
Finally, for alert determination thresholding th, we worked
within a fixed range of {0.05,0.1,0.2,· · · ,0.9}and an
adaptive method called ‘‘Otsu’s thresholding method [32],’
which is a commonly used image thresholding algorithm.
Here we describe the results of the above five parameter
tuning. It was not practical to tune all five parameters
simultaneously, because the number of combinations would
be immense. As an evaluation strategy, we divided the
parameters into two groups: ˜
Rn,Rn, which is directly
FIGURE 2. Results of true positives (TPs) and false positives (FPs) for
each threshold (th) when all eight sensors were utilized. (horizontal axis:
FIGURE 3. Results of true positives (TPs) and false positives (FPs) for
each sensor when the threshold (th) was Otsu’s thresholding method.
(horizontal axis: epochs).
related to NTD, and sensor,epochs, and th, which
are not. Because sensor,epochs, and th have a more
significant impact on detection performance, we first roughly
tuned the detection accuracy with these three parameters
and then fine-tuned it with ˜
Rn,Rn. In this experiment,
the spatiotemporal feature extraction in Section II-B was
performed by generating and using tensors Vhp for October
2018 with the observation time unit Tset to 1,800 s, the
sampling interval Mset to 30, and the online processing time
unit tset to 600 s.
We evaluated the tuning of sensor,epochs, and th after
fixing the values to ˜
Rn=25,Rn=5, which were empirically
used in an earlier study [9]. The results of the evaluation are
presented in Figs. 2 and 3. The horizontal axis represents
the number of epochs, and the vertical axis represents the
number of true positives (TPs) and false positives (FPs) of
the port numbers. When the number of TPs is close to 35 and
the number of FPs is low, we can observe that the detection
13044 VOLUME 10, 2022
C. Han et al.:Dark-TRACER: Early Detection Framework for Malware Activity
TABLE 1. Characteristics of malware activities observed on our darknets in October 2018. Malware activities with similar characteristics are grouped
according to TCP ports.
FIGURE 4. Dispersion results of the number of TPs (red) and FPs (blue)
for each epoch over 15 iterations. Error bars represent standard
deviations. The upper graph shows the results for each th when sensor
Ais set, and the lower graph shows the results for each sensor when th
is Otsu’s thresholding method.
accuracy is excellent. Fig. 2 shows the results for each th
when all eight sensors are utilized. This indicates that th
results in more TPs with fewer epochs while keeping the
number of FPs relatively low when using Otsu’s thresholding
method (otsu) as compared with any other fixed value.
Next, Fig. 3 shows the results for each sensor when the th
used is Otsu’s thresholding method. The results demonstrate
that when using only the sensor A, the same number of TPs
is achieved with fewer epochs while keeping the number of
FPs lower than when using either all sensors in combination,
or other sensor alone. Based on the above, we conclude that
the best solution is to use only Afor sensor, 4 for epochs,
and Otsu’s thresholding method for th.
Next, we provide a brief discussion regarding each
parameter. Fig. 4 shows the dispersion of TPs and #FPs
TABLE 2. Tuning evaluation results of
Rn,Rnin Dark-NTD.
for each epoch over 15 iterations. The red graph is the
mean number of TPs, the blue graph is the mean number
of FPs, and the error bars represent the standard deviations.
From these results, we can conclude that the randomness
of the initial value selection of FSTD and NTD does not
dramatically affect the detection performance because a
similar number of TPs and #FPs was recorded each time.
In terms of the sensor, Dark-NTD recorded a good number of
TPs for sensors with a large observation scale. Finally, Otsu’s
thresholding method achieves a similar level of accuracy to
a fixed value of 0.3 but has the advantage of adaptively
determining a threshold value from the data.
In this section, we set sensor,epochs, and th to the
values determined above, and then evaluated the tuning
of ˜
Rn,Rn. The results are shown in Table 2. Contrary
to expectations, increasing the value of ˜
Rndecreased the
number of TPs and increased both the number of FPs and
the average processing time. This result supports the fact
that even at ˜
Rn=25, we can sample enough important
information (fiber) for low-rank approximations. Therefore,
we determined that ˜
Rn=25 is appropriate. In the case
of Rn, there were no significant differences in the average
processing time for any value, and the greatest number of TPs
was achieved with Rn=5, so we conclude that Rn=5 is
most appropriate.
VOLUME 10, 2022 13045
C. Han et al.:Dark-TRACER: Early Detection Framework for Malware Activity
TABLE 3. Comparative evaluation results of detection performance of
malware activities. Dark-TRACER is the result of integrating Dark-GLASSO,
Dark-NMF’s SET1, and Dark-NTD’s Tuned.
In this section, we quantitatively compare and evaluate the
detection accuracy of malware activities among the modules.
The results are shown in Table 3. ChangeFinder [10] is
an existing method that was applied in NICTER before
proposing each module of Dark-TRACER, and it is an
algorithm for detecting change points in time-series data with
a low computational cost. The ChangeFinder algorithm is
based on the sequential discounting autoregressive (SDAR)
forgetting learning algorithm, which calculates only new
time series data and reduces the influence of past data by
improving the autoregressive model to learn sequentially.
We implemented ChangeFinder on two types of time-
series data: the number of packets and unique source hosts
in 10 min. The parameters of each module used in this
experiment are described below.
autoregressive order =2, forgetting parameter =
0.005, smoothing range (two steps) = {10,5},
threshold for change detection =3
T=600, M=12, t=600, used matrix Vh,K=
432, θ=0.98, λ= {0.4,0.5,0.6,0.7,0.8,0.9},
T=1800, M=30, t=600, used matrices (Vh,Vp),
α=30, β=2, r= {1,2,· · · ,10},f=0 for
SET1,f=1 for SET2
T=1800, M=30, t=600, used tensor
Rn=25, Rn=5, epochs=4,
th=Otsu’s thresholding
Here, λin Dark-GLASSO is a regularization coefficient for
Graphical Lasso. Due to the high computational complexity
of Dark-GLASSO, random sampling was conducted when
the number of hosts Nhexceeded γto maintain real-time
performance. All other parameters are explained in Section II.
Next, we explain the notation used in Table 3. SETs in
Dark-NMF indicates the difference between 0 and 1 settings
FIGURE 5. The number of unique source hosts per hour on NICTER,
where Moobot-related malware activities are observed. A synchronized
fluctuation in the number of hosts was confirmed on September 19 and
September 21. D: The earliest time detected by Dark-TRACER,N: The time
observed by NICTER operators, and P: The time when was is revealed to
the public.
of f.CONV in Dark-NTD is a parameter setting introduced
in previous research, and Tuned is the parameter setting
determined by tuning in the previous section. The difference
between CONV and Tuned is that Tuned has epochs and
uses only sensor A. Note that only Tuned in Dark-NTD
utilizes sensor A, whereas the other modules utilize all
eight sensors. #FNs is the number of false negatives, which is
#TPs+#FNs =35. Recall is an evaluation metric calculated
as #TPs / (#TPs+#FNs).
Lastly, we describe the symbol [0] attached to SET20and
Tuned0. The results of SET2 and Tuned show that the
number of FPs is very high. The primary cause of false
positives is synchronized scans by investigative scanners,
such as Shodan and Censys [11]. To address this problem,
at least temporarily, we attempted to exclude alerts from
investigative scanners by applying a simple rule to the
alert results of SET2 and Tuned. The simple rule was:
if a large number, or a sequential number, of TCP ports
were seen concurrently from the same source hosts in the
alerts, those alerts are excluded. The application results
of the rule were SET20and Tuned0.Dark-NMF has an
effect of halving the number of FPs while maintaining the
number of TPs, whereas Dark-NTD does not have such an
From the results of a comparative evaluation in Table 3,
Dark-TRACER achieves a recall rate of 100%, although there
are some FPs, by integrating the results of each module.
Next, we examine the characteristics of the ports that are false
negative in each module. ChangeFinder tends to perform
poorly at detecting small host sizes and for short- or long-
term constant malware activities. Dark-GLASSO and Dark-
NMF are perform poorly when detecting malware activities
with small host sizes. Furthermore, SET1 tends to be weak in
detecting long-term persistent activities. Dark-NTD tends to
be poor at detecting short-term malware activities. Overall,
the results show that an integration of the three proposed
modules can mutually complement the weaknesses of each
In this section, we assess the feasibility of the early
detection of malware activities. The details of the dataset, the
13046 VOLUME 10, 2022
C. Han et al.:Dark-TRACER: Early Detection Framework for Malware Activity
TABLE 4. Details of the ground truth to assess the feasibility of early detection of malware activity. It consists of 33 TCP ports and 12 types of threat events
for malware activities observed in 17 months from June 2019 to October 2020, wherein the time of infection spread is clearly discernible. (RCE:Remote
Code Execution,C&C:Command and Control,DDoS:Distributed Denial-of-Service,CVE:Common Vulnerabilities and Exposures,PoC:Proof of Concept).
FIGURE 6. The number of unique source hosts per hour on NICTER,
where Moobot-related malware activities are observed. Here, the events
that do not fluctuate synchronously with the ports in Fig. 5 are shown. D:
The earliest time detected by Dark-TRACER,N: The time observed by
NICTER operators, and P: The time when it was revealed to the public.
experimental setup, and the assessment results are described
below. This experiment included the general method of cross-
FIGURE 7. The number of unique source hosts per hour on NICTER where
the malware activities were observed (December 2019 to March 2020). D:
The earliest time detected by Dark-TRACER,N: The time observed by
NICTER operators, and P: The time when it was revealed to the public.
validation in time series data; after learning the optimal
parameters with past data in section 3, we verified them with
future data in this section.
VOLUME 10, 2022 13047
C. Han et al.:Dark-TRACER: Early Detection Framework for Malware Activity
FIGURE 8. Number of unique source hosts per hour on NICTER, where
malware activities were observed (March 2020 to August 2020). D: The
earliest time detected by Dark-TRACER,N: The time observed by NICTER
operators, and P: The time when it was revealed to the public.
In this experiment, we used data from three sensors A(/17
subnet), B(/18 subnet), and D(/20 subnet), selected by
observation scale, instead of all eight sensors. The data period
used in the experiment was 17 months, from June 2019 to
October 2020. To highlight the observation of unknown
malware activity, we excluded known and routinely observed
TCP ports by calculating them for each month at each sensor
as a preprocessing step.
Next, we describe the details of the ground truth used
in the evaluation. In the experiment described in Section 3,
the ground truth included many stationary threats whose
infection spread period of malware activities was unclear;
thus, it was not easy to assess early detection. In this
experiment, we manually generated a new ground truth for
malware activities observed from June 2019 to October 2020,
which represented a set of threats with a clearly identifiable
infection spread period. The newly prepared ground truth
was based on reports and blog posts published by NICTER’s
expert operators.5Among the malware activities observed by
NICTER, we selected malware activities whose origin and
characteristics were clear and for which there were references
by third parties. As a result, we collected 12 types of threat
events on 33 TCP ports. The breakdown of the ground truth is
shown in Table 4. The following information was accurately
the initial period when NICTER began to observe a rapid
increase in the number of packets and hosts on TCP ports
related to threats
the change in the scale of the number of hosts at that time
the period in which threats were revealed to the public
due to references issued by reputable third-party security
research organizations (i.e., reveal date)
characteristics of the threats
Third-party references included recurring activities, such as
BlueKeep,ShenZhen TVT, and MikroTik, which are
attacks on previously known vulnerabilities.
This ground truth considers not only the type of threat but
also its variations, such as the observed infected host size and
the persistence/stationarity of threats. For clarity, hourly time-
series graphs of the number of unique source hosts observed
by NICTER are shown in Figures 5, 6, 7, and 8 for each TCP
TABLE 5. The number of early detected ports, late detected ports,
overlooked ports (#FNs) and their average numbers of days to detection
based on the initial period of observation by NICTER.
TABLE 6. Average number of unique ports by period.
port. The solid vertical line labeled ‘D’ represents the earliest
period detected by Dark-TRACER, the dashed line labeled
‘‘N’ represents the period observed by NICTER operators,
and the dashed line labeled ‘‘P’ represents the period when
was is revealed to the public by reputable third-party security
research organizations.
Figures 5 and 6 are time-series graphs summarizing
the partial TCP ports where Moobot-related threats were
observed. In Fig. 5, port groups with synchronized fluc-
tuations in the number of unique hosts can be confirmed
as belonging to one group on September 19 and one
group on September 21, indicating that large-scale Moobot
activity was observed during this period. Figure 6 shows
events where Moobot features were observed but did not
show synchronized fluctuations with the ports in Fig. 5.
These ports are related not only to Moobot, but also to
the activities of other Mirai variants such as Fbot and
Estella. As shown above, Moobots malware activity
is an orchestrated threat that combines multiple activities.
Next, Figs. 7 and 8 show time-series graphs of partial TCP
ports among threats other than Moobot. Of these partial
TCP ports, we found several threats that were difficult to
detect by conventional methods or by human efforts, such
as threats with no spikes, constant threats, and small-scale
We now describe the processing results for each parameter
with their best parameters, which were the same as in
the experimental setup described in Section III-C. We also
applied the aforementioned simple rule to exclude alerts
caused by investigative scanners. Here, Dark-NMF was
computed with the parameter SET1. Table 5 shows the
number of ports that were detected early, late, or falsely
negative, and their average number of days, based on the
initial period of NICTER observations. The results show that
although there were a few overlooked ports (#FNs) and late
13048 VOLUME 10, 2022
C. Han et al.:Dark-TRACER: Early Detection Framework for Malware Activity
detected ports when considered by the module, all 33 TCP
ports could be detected at an early stage when integrated
with Dark-TRACER. In addition, Dark-TRACER was able
to detect threats on average 126.4 days earlier than the
initial period when threats were first observed by NICTER,
and 153.6 days earlier than the period when threats were
announced to the public.
We also investigated how many ports were alerted for
each module in this experiment. Table 6 shows the average
number of unique ports per period for each module. For
the entire 17-month period, Dark-GLASSO,Dark-NMF,
and Dark-NTD produced alerts for 66, 2,042, and 3,969
unique ports, respectively. When the proposed modules were
integrated into Dark-TRACER, the number of unique ports
was 5,271. We counted the number of unique ports for
each day, week, and month, and the averages are shown in
Table 6. For example, Dark-TRACER issued alerts for an
average of 58.49 ports per day. Assuming that one analyst
requires 15 min of analysis time per port (refer to Section. V-
E), two analysts could perform these daily operations in
approximately 7.3 h (roughly 14.6 h for a single analyst).
It would require approximately 31.5 h for a week and 89.8 h
for a month with two analysts.
From the above two experiments, we found that Dark-
TRACER could tune the parameters of each module so that
the number of FNs was almost non-existent and could also
detect malware activities at a fairly early stage. As a future
challenge, the cost of analysis would be lower if the number
of FPs could be reduced more precisely. In addition, there is a
possibility that expert analysis would disclose other unknown
activities, in addition to the malware activities that were
selected for the ground truth.
In this section, we provide a comprehensive discussion
and insight into the performance of our framework. First,
we demonstrate the advantages of Dark-TRACER and provide
a comprehensive comparison of each proposed module. Then,
we discuss the potential concerns of our approach, such as
adversarial attacks and the reduction of false-positive alerts.
Finally, we present guidelines for the practical application of
As mentioned in the introduction, by focusing on the
synchronization of spatiotemporal patterns in darknet traffic,
we have the following advantages.
Distinguishing between non-attack-related and attack-related
communications from darknet traffic is a difficult task.
Misconfigured or unexplained communications are nuisances
that interfere with the interpretation of darknet traffic
analysis. In this paper, we focused on the fact that hosts
infected with similar malware tend to compromise and scan
FIGURE 9. A 3D graph visualizing a case of anomalous synchronization of
the spatiotemporal patterns detected from the experimental results in
Section III. A scatter plot of partial Vhp during 18:30–19:00 on October 31,
2018, is visualized. Each of the three axes is a time axis in minutes,
a source-host spatial axis, and a destination-port spatial axis, and the
plots represent the observed packets (element values in Vhp). For the
destination-port spatial axis, there are only three points at which
anomalous synchronicity was detected—5379, 6379, and 7379. Host IPs
are plotted in red if they match on multiple ports within one-minute
increments and in green if they do not. The red points are considered to
be synchronized communications caused by factors such as malware
activities, while the green points are considered to be noise
in a synchronized spatiotemporal pattern. By estimating the
synchronicity of spatiotemporal patterns in the darknet traffic
and eliminating communications that do not show synchro-
nization from the scope of analysis, noisy communications
are expected to be scraped off, and malicious communications
can be highlighted.
For understanding, a visualization of the synchronization
of the spatiotemporal patterns is shown in Fig. 9. This is an
alert of malware activity detected by Dark-NMF at sensor
Aduring 18:30-19:00 on October 31, 2018, visualizing Vhp
at that time. The number of packets is plotted in three
dimensions: time, source host, and destination port in one-
minute increments. Figure 9 shows that the number of red
dots indicates the number of communications from the same
host to the same destination ports (5379, 6379, 7379/TCP)
during that time period. As indicated in Table 1, we observed
a scanning attack on the same service Redis at these ports.
Thus, many red dots appear when the spatiotemporal pattern
has anomalous synchronization. In contrast, the green dots
can be regarded as noisy communication. It is assumed that
Dark-TRACER detects anomalies by highlighting the red
dots (e.g., malware activity) while eliminating the green
dots (noise communication). The red dots (synchronization
between spatiotemporal features) do not appear as abundantly
as they appeared in Fig. 9 between arbitrary destination ports
of ordinary darknet traffic.
VOLUME 10, 2022 13049
C. Han et al.:Dark-TRACER: Early Detection Framework for Malware Activity
TABLE 7. A comprehensive comparison of the proposed modules Dark-GLASSO, Dark-NMF, Dark-NTD.
Traditionally, malware activities have been detected based on
changes in time-series data, such as the number of packets and
the number of hosts, using change-point detection algorithms
such as ChangeFinder, or manually by operators. Until
recently, many malware activities were relatively easy to
detect by operators, as they were threats that had severe and
obvious changes in time-series data, threats with clear spikes,
or threats that were simple and had a large scale of infection.
However, in recent years, the amount of communication
observed in the darknet has increased exponentially and
cyberattacks have become more diverse and sophisticated,
such as the Moobot described in Section IV-A. Such
orchestrated threats, which intertwine multiple activities,
small-scale threats, threats without explicit spikes, and
constant threats, are malware activities that are difficult to
detect manually. However, there is a significant possibility
that Dark-TRACER can detect such traditionally hard-to-
detect malware activities. Its feasibility is well demonstrated
by the evaluation results of early detection in Section IV,
wherein various types of malware activities were detected.
Dark-TRACER is not only capable of detecting traditionally
hard-to-detect threats, but also of detecting them early and in
real-time. Even when the scale of malware activity is small,
if there is overlap in the spatial features (e.g., the distribution
of hosts and ports) with another malware activity that has
already been detected, and if there is synchronicity in the
pattern of the number of packets, there is a high probability
that they will be detected together. This implies that Dark-
TRACER can capture the signs of infection before it spreads
in earnest. In this way, by checking the overlapping degree of
host spatial feature variables between alerts from the same
period, it is possible to identify threats that at first glance
appear to be different events, but are actually caused by the
same malware. In fact, as shown in Figs. 5 and 6, Dark-
TRACER can detect orchestrated threats such as Moobot,
in which multiple activities are intertwined, at an early stage
by detecting signs of infection as they spread, even at a small
In this section, we comprehensively compare the proposed
modules Dark-GLASSO, Dark-NMF, Dark-NTD in terms
of accuracy, cost, anomaly detection method, and spatial
features. An overview is given in Table 7, and detailed
explanations are provided in order from the top of the list.
First, we discuss the accuracy aspect. In general, there
was a trade-off between the number of FPs and FNs.
Dark-GLASSO had almost no FPs and Dark-NMF, Dark-
NTD had almost no FNs. As for the performance of
early detection, Table 5 shows that Dark-GLASSO tended
to make detections slightly later, but the other modules
almost always detected threats early. Next, as mentioned
in Section III-C, we examined the characteristics of each
module’s number of FNs ports and considered the types of
threats that each module overlooked. Dark-GLASSO tended
to miss small-scale threats, Dark-NMF tended to miss small-
scale and constant threats, whereas Dark-NTD tended to be
weak at detecting short-term threats. The same tendency
was confirmed by the experimental results described in
Section IV. Because the modules are complementary to
each other, it is recommended to use them in an integrated
manner, rather than using only one in isolation. Table 5 shows
that by integrating the modules into Dark-TRACER, we can
avoid problems of missed or delayed detection. Finally,
the accuracies of Dark-GLASSO and Dark-NTD were low,
unless the observation scale of the darknet sensor was large.
In contrast, Dark-NMF recorded the same level of accuracy
for all eight sensors of different sizes used in Section III.
13050 VOLUME 10, 2022
C. Han et al.:Dark-TRACER: Early Detection Framework for Malware Activity
Dark-NMF is very computationally inexpensive and does
not require any particular preprocessing. In contrast, Dark-
GLASSO and Dark-NTD are computationally expensive and
require specific preprocessing. Given the spatial feature sizes
Nhand Np,graphical lasso has a cubic time complexity of
h), NTD has a quadratic time complexity of O(Nh·Np),
and NMF has a linear time complexity of O(Nh) or O(Np).
In this study, Dark-TRACER was processed online sequen-
tially at 10-min intervals. Therefore, we adjusted the param-
eters and preprocessed each module to finish the processing
within 10 min for one data slot. Dark-NMF did not require
any preprocessing. Next, as described in Section III-C, Dark-
GLASSO performs random sampling preprocessing when the
number of hosts Nhexceeds γ. In a previous paper [7],
we reported that the output alerts were quite unstable when γ
was lower than the expected average value of Nh.Dark-NTD,
as described in Section II-C, applies FSTD [20] to preprocess
the tensor Vto a low-rank approximation and preliminarily
samples only the essential parts. The larger the number of
bases ˜
Rnin FSTD, the better the low-rank approximation of
the original tensor. However, as shown in Table 2, the results
of tuning ˜
Rndemonstrated that increasing the value of ˜
worsened the accuracy. In all experiments, the processing
time of each module was no longer than 10 min (CPU: AMD
RYZEN TR 2990WX). For one data slot, Dark-NMF took
approximately 1 min, Dark-NTD approximately 5 min, and
Dark-GLASSO approximately 7 min.
An important factor in the cost of alert analysis is the
number of ports that must be analyzed per unit period.
As shown in Table 6, Dark-GLASSO has the lowest cost.
For Dark-GLASSO, only 66 ports needed to be analyzed
during the entire experiment in Section IV, whereas the other
modules required 30 to 60 times more cost.
Next, we discuss methodological differences in anomaly
detection. Dark-NMF and Dark-NTD decompose spatiotem-
poral features into latent frequent patterns and then perform
anomaly detection for each group of decomposed spatial
features. This decomposition can be regarded as a favorable
condition for detecting local events, but it is also one of the
reasons why the number of detected anomalous events (alerts)
becomes very large, although it should be noted that the
number of alerts can be adjusted by tuning the parameters.
In contrast, Dark-GLASSO detects anomalies from all spatial
features without decomposing the spatiotemporal features.
This is a favorable condition for detecting global events
and is one of the reasons that the number of anomalous
events (alerts) detected is small.
In addition, Dark-GLASSO requires Kof the past data to
perform anomaly detection. Each time we change the value
of the parameters or introduce a new sensor, Dark-GLASSO
has to wait for Kperiods to obtain the detection results.
Other modules, however, do not require waiting in such cases
because they do not require past data.
Finally, we discuss differences in the application of spatial
features. In Dark-TRACER, two types of spatial features can
be handled: host space and port space. For darknet traffic in
a short unit time, the size of the port space Nptends to be
larger than the size of the host space Nh. In Dark-GLASSO,
the port spatiotemporal feature matrix Vpis not employed
because it becomes computationally intractable when the
size of the spatial features becomes large. Dark-NMF can
handle not only the host spatiotemporal feature matrix Vh
but also the port spatiotemporal feature matrix Vpdue to its
low computational complexity. Finally, because Dark-NTD
is designed to handle a three-dimensional spatiotemporal
feature tensor Vhp from the beginning of the proposal, it can
calculate the host/port space simultaneously.
This section discusses adversarial attacks that an attacker
might implement to evade detection by Dark-TRACER.
Possible attempts include multiplying noise, distributing
spatial features, and reducing the frequency of temporal
features, which might prevent the malware from capturing
spatiotemporal features when the framework performs scan-
ning compromises.
The case wherein dummy scans, which are unrelated
to an attack, are attached to a true attack to confuse the
detection framework.
In this case, more data will be observed, and syn-
chronization of the spatiotemporal features will be
captured more strongly, resulting in better detection
of true attacks. However, some of the detected events
may contain dummy scan information, which may be
troublesome for analysis.
The case wherein multiple true attacks from many
attack groups are distributed and executed simultane-
The advantage of Dark-NMF and Dark-NTD is that they
can be decomposed into several patterns with similar
spatiotemporal features. Therefore, Dark-TRACER can
detect anomalies by dividing potential attack groups
into several groups, even when orchestrated attacks are
conducted simultaneously.
The case of a slow stealth scan attack.
Depending on the degree of stealthiness, if a stealth scan
attack is too slow, the synchronization of the observed
spatiotemporal features becomes weak, making the
attack difficult to detect. However, slow stealth scans
are not efficient for an attacker who wants to spread
the damage of their malware as quickly as possible,
because the speed of spreading the malware infection
is significantly slower. Slow stealth scans are generally
considered to have purposes other than the spread of the
VOLUME 10, 2022 13051
C. Han et al.:Dark-TRACER: Early Detection Framework for Malware Activity
malware infection directly, and thus are not the target of
detection in this study.
As mentioned in Section III-C, the primary cause of
false-positive alerts (#FPs) issued by Dark-TRACER is
synchronized scanning by organizations for investigative
purposes. Organizations such as Shodan and Censys,
which deploy various cyberspace intelligence information as
search engines, regularly scan the entire Internet space at a
high frequency. Because such Internet-wide scans are fast and
large-scale, they are observed in our darknet sensor networks
and are represented as synchronized spatiotemporal patterns.
Consequently, Dark-TRACER achieved a low number of
FNs, whereas the number of FPs from investigative scanners
is very high.
We believe that if Dark-TRACER can distinguish between
alerts caused by investigative scanners and alerts caused by
malware activities in a secondary manner, after detecting
anomalous synchronous spatiotemporal features without
missing them, the inefficient situation wherein there are
many FPs can be significantly improved. In this study,
to temporarily solve this challenge, we applied a simple rule
that excluded alerts when a large number, or a sequential
number, of TCP ports were seen simultaneously from the
same source hosts in the alerts. In the first experiment
described in Section III, we found that Dark-NMF halved
the number of FPs while maintaining a high number of
TPs, whereas Dark-NTD did not. In the second experiment
in Section IV, by applying our simple rule, we were able
to reduce the number of unique ports from 64,103 to 5,271
for the entire period alerted by Dark-TRACER.In this
way, we have demonstrated the feasibility of significantly
improving malware detection by secondarily examining
alerts. In future work, we would like to develop a model for
classifying or clustering scanners for investigative purposes
and automatically create a blacklist so that we can eliminate
alerts caused by investigative scanners with better accuracy
than the simple rule used in the present study.
Each of the three independent proposed modules has its own
strengths and weaknesses, and they complement each other
through their collaboration into a single framework Dark-
TRACER. From the two experiments presented in this paper,
it was found that Dark-TRACER can achieve a 100% recall
rate in the detection accuracy of malware activities and can
also accomplish early detection. In this section, we discuss
how Dark-TRACER can be operated in a practical manner.
First, we consider each module separately. Dark-GLASSO
has a small number of FNs, but because there are few FPs,
the precision rate #TPs / (#TPs+#FPs) is high. When it
is not possible to spend much time on the analysis of the
detection results, or when the analysis of global malware
activities is sufficient, it is practical to employ only the
detection results of Dark-GLASSO. Next, Dark-NMF and
Dark-NTD, which use nonnegative tensor decomposition
methods, are beneficial for detecting local malware activities
because they can detect many FPs while incurring almost no
FNs at an early stage. In particular, Dark-NMF is effective
in detecting anomalous synchronization because it does not
require preprocessing, has a very low computational cost,
and shows good detection accuracy, even for darknet sensors
with small observation scales. In contrast, Dark-NTD has a
very high potential for detecting small-scale threats that are
typically considered difficult to detect with other modules and
is useful for capturing fairly localized events. As described
above, each module has different characteristics and can
be utilized according to nature of the precise situation,
or all modules can be fully leveraged into an integrated
framework as in Dark-TRACER, taking advantage of their
complementary relationship.
Finally, we discuss a secondary analysis method for
the detection results of Dark-TRACER. The alerts issued
by Dark-TRACER contain information on IP addresses,
targeted ports, and the timestamps of the hosts that are
identified as abnormal. However, this information alone is
often not enough to accurately determine malware activity.
As mentioned in Section III-A, some malware activities, such
as Mirai and Hajime, are known to have fingerprints in
their initial scan packets. In large-scale scans such as malware
and scanners which operate for investigative purposes, packet
headers are often designed to have fingerprints in order to
scan faster [44], [45]. Previous research has also reported that
scanners use fingerprints to distinguish their scan results from
backscatters [46].
The question arises as to what specific information should
be checked. The following steps are considered useful for
secondary analysis of Dark-TRACER alerts:
1) Computing the statistics of packet headers of detected
alerts and find characteristic header information
(including known fingerprints such as Mirai and
2) Checking whether honeypots in an interactive observa-
tion network have observed any communication related
to the detected alerts, and if so, analyzing what type of
communication occurred interactively.
3) Collating and analyzing the presence of information
related to the detected alert in third-party threat
intelligence information (e.g., CVEs, vulnerabilities,
and reports).
This is the actual workflow of the security operations center
at NICTER.NICTER operations experts are expected to ana-
lyze the aforedescribed collation process in approximately
15 min per port of an alert. However, this does not necessarily
imply that the causes and details of all events can be clarified.
In order to increase the number of events that can be clarified
as much as possible, it is necessary to collate more abundant
information. In the future, we intend to extend Dark-TRACER
by considering a wide range of applications, such as a
mechanism to reduce false positives, improve both recall
13052 VOLUME 10, 2022
C. Han et al.:Dark-TRACER: Early Detection Framework for Malware Activity
FIGURE 10. Total annual number of observed packets per IP address on
and precision, and automatically associate threat intelligence
from third parties [47].
We describe related work on how darknets are leveraged
in measurement analysis and malware activity detection.
We also provide related studies and directions for identi-
fying investigative scanners on darknet traffic, which will
inevitably become necessary in the future.
The darknet has attracted extensive attention in the field of
network security, and many researchers are actively engaged
in research on its development, analysis, and visualiza-
tion [55]. Previous research [2], [56]–[58] has discussed the
fundamentals of various darknet configurations, deployment
techniques, and sensor placement techniques, and clarified
the effectiveness of darknets. In addition, profiling, filtering,
and classification have been intensively studied for the
measurement of darknets. In the rest of this subsection,
we present related work on IoT malware analysis and general
darknet measurement analysis. A summary of the related
studies is shown in Table 8.
Fig. 10 shows that the amount of observed traffic in
NICTER’s darknet observation network with a total of
300,000 IP addresses has increased rapidly over the past
few years. The main reason for this is the IoT malware
‘‘Mirai’, which appeared in 2016 [22]. In contrast to
conventional botnets, IoT malware scans multiple ports in
parallel to form a large-scale botnet that can spread the
infection faster [48]. Moreover, IoT malware variants behave
competitively with each other and are repeatedly destroyed
and reinfected over a short period [23]. The emergence of such
diverse and sophisticated IoT malware further complicates
cyber threats and makes it difficult to examine the actual
current state of malware strategies. Therefore, it is essential
to have a mechanism to investigate IoT botnets while they are
still persistent and to rapidly and precisely detect potential
Apart from malware activity detection, which is discussed
in the next subsection, the following studies were conducted
in other areas of darknet measurement analysis. Dainotti et al.
contributed to a census-like analysis of how the IP address
space is used by developing malware and evaluating methods
to remove spoofed traffic from darknets and live net-
works [49]. Durumeric et al. analyzed a large-scale darknet
to investigate Internet-wide scanning activities and identify
patterns of extensive horizontal scanning operations [50].
Fachkha et al. devised an inference and characterization
module to identify and analyze the probing activities of cyber-
physical systems (CPS) by extracting various features from
large amounts of darknet data and performing correlational
analyses [51]. Jonker et al. introduced a framework to protect
against DoS attacks based on various data sources, including
darknet traffic data [52]. They found that one-third of all /24
networks on the Internet had suffered at least one DoS attack
in the past two years. Shaikh et al. identified unsolicited IoT
devices by collecting IP header information from darknet
traffic data and classifying them using several machine
learning algorithms [53]. Akiyoshi et al. proposed a method
to detect emerging scanning activities and their scale by
analyzing the correlation between traffic in honeypots and
darknets [54]. Most of the measurement analysis studies using
darknets have been applied to understand the general trend
of malicious communications observed in darknets. Thus, for
detailed analysis, many studies use not only darknet data but
also trap-based monitoring systems such as honeypots.
A summary of the related works referred to in this subsection
is provided in Table 9. According to a survey paper on
darknets [55], the technique of filtering misconfigured traffic
has not yet been fully explored and is an ongoing challenge
that deserves more attention from the research community.
We consider that our method can filter out misconfigured
traffic by detecting anomalies in the synchronization of
spatiotemporal patterns. Furthermore, our method is unique
in that it can detect global cyber threats/malware activities
in real time in a uniform format by focusing on the syn-
chronization of anomalous spatiotemporal patterns among
many indiscriminate suspicious scans that reach large-scale
Here, we present some prior research that had a similar
scope to our problem and used darknet traffic but did
not focus on synchronization. There are several methods
to detect anomalies by detecting change points in darknet
traffic, such as ChangeFinder that was introduced as a
comparison method in a previous study [10], [59]–[61].
Ahmedet al. proposed a sliding window-based adaptive
cumulative sum (CUSUM) algorithm, which is a sequential
analysis method for detecting drastic changes in darknet
traffic [59]. Inoue et al. [60] employed the ChangeFinder
algorithm [10] to detect sudden change points in darknet
traffic with a low computational cost. Ban et al. proposed
an abrupt-change detection algorithm that can detect botnet
VOLUME 10, 2022 13053
C. Han et al.:Dark-TRACER: Early Detection Framework for Malware Activity
TABLE 8. Summary of related works on darknet measurement analysis including IoT malware (Mirai).
TABLE 9. Summary of related works on darknet detecting malware activity from darknet traffic.
probe campaigns with a high detection rate by searching
for temporal coincidences in botnet activities observed on
the darknet [61]. The aforementioned change detection
methods all share the same drawback—they cannot achieve
high accuracy without focusing on specific protocol ports
because they detect change points without distinguishing
between many sources of noisy communications, such as
misconfigured traffic. As shown in the experimental results
of ChangeFinder herein, the accuracy was low when the
algorithm was applied to the entire traffic without focusing
on a specific port. In addition, applying the change detection
method to a specific port would result in many alerts, which
would require considerable computational and analysis costs.
Next, we review recent related works on the analysis
of malware activity using darknet data (mainly clustering).
As mentioned in the previous subsection on darknet mea-
13054 VOLUME 10, 2022
C. Han et al.:Dark-TRACER: Early Detection Framework for Malware Activity
surement analysis, most of the communications reaching
the darknet since 2016 have seen a considerable increase
in traffic volume due to IoT malware. Therefore, many
clustering methods targeting the analysis of IoT malware
activities have been studied in recent years. Pour et al.
learned to classify compromised IoT devices and non-IoT
devices using convolutional neural networks [62]. Using
the classification training results, they extracted features
embedded in unsolicited scan flows and deployed hierar-
chical agglomerative clustering. As a result, the authors
uncovered 440,000 compromised IoT devices and 350 IoT
botnets. Torabi et al. leveraged the Shodan IoT search
engine and darknet traffic data to infer compromised IoT
devices and to characterize and cluster the generated scanning
campaigns [63]. The authors discovered newly targeted
ports and emerging IoT malware/botnets, highlighting their
persistence and evolutionary process. Cohen et al. proposed
a method for sequential clustering of aggressive hosts having
similar intentions from scalable darknet traffic by embedding
a port sequence for each source host [24]. By tracking the
clusters, they detected recurrent or new attacks and found
several new campaigns of malicious port sequences.
Such clustering methods are considered to be able to
classify malware activities, investigative scans, and noisy
communications (e.g., misconfigured traffic) to some extent.
However, although the clustering methods can distinguish
distinctive hosts or packets as a group, they cannot detect
anomalies and thus cannot identify which clusters are
anomalous or distinctive. Therefore, it is necessary to
prioritize the clusters to be analyzed based on some criteria.
In contrast, as Dark-TRACER issues alerts in real time upon
anomaly detection, it is easy to know what to analyze first,
thereby making the framework highly practical. In addition,
clustering methods are not suitable for the early detection of
unknown or small-scale malware activities because clusters
are not formed unless the amount of observed features of
malware activities is large. As shown in the experimental
results herein, Dark-TRACER has good potential to detect
small-scale malware activities in an early stage.
BotSniffer [1] and BotMiner [64] proposed a framework
for detecting C&C traffic and malicious activities based on
the spatiotemporal correlation method. However, the scope of
their problem setting was different from that of ours, because
BotSniffer and BotMiner only target specific protocols of
actual network traffic and therefore cannot detect global
cyber threats and malware activities in the entire Internet
space. To the best of our knowledge, there is no related
work that focuses on synchronization in the same scope as
the present study. As described in Table 9, prior papers [6]
and [7] were published as prototypes of Dark-GLASSO,
prior paper [8] was published as a prototype of Dark-NMF,
and, finally, prior paper [9] was published as a prototype
of Dark-NTD. However, as the previous methods have been
considered independently, their relationship has remained
unclear. In this study, we integrated and evaluated the three
previous methods as Dark-TRACER and clarified that they
complement each other. In addition, although the previous
methods can detect malware in real time, the feasibility
of early detection of malware activity was not evaluated;
therefore, in this study, we evaluated the early detection
performance and the analysis cost.
We also present several other related works that utilized
darknet traffic to identify and detect malicious communica-
tions, based on the reports of a recent survey paper [65].
Kumar et al. proposed a model that learns from darknet data
and benign traffic data to test whether it can classify malicious
communications [66]. However, the model only classifies
whether the traffic flow is malicious or benign, and because it
learns all at once, it does not specifically identify what kind of
maliciousness it has classified. Bou-Harb et al. investigated
orchestrated probing campaigns by considering a clustering
method for time-series traffic data [67]. However, this
method does not detect anomalies and does not distinguish
noisy communications. Ali et al. classified DDoS attacks
using Resource Allocating Network with Locality Sensitive
Hashing) (RAN-LSH), which employs LSH to select data
for training and achieves fast online learning by training
only selected data [68]. However, because this method only
analyzes backscatter traffic and targets to classify DDoS
attacks, it is not suitable for detecting and classifying malware
activities, which do not involve backscatter.
We conclude this section by sharing related works that
have distinguished between investigative scanners, which
is an issue that must be considered in future studies.
A summary of the related works mentioned in this sub-
section is presented in Table 10. Many massive Internet-
wide scanners are observed on the darknet, including both
public scanning activities and malware activities. Recently,
high-performance scanning tools such as ZMap [44] and
Masscan [69] have been deployed, and Mazel et al. profiled
the utilization of such tools [45]. The results revealed that
many entities openly engage in scanning activities on a
large scale and on a constant basis. Because such harmless
and large-scale investigative scanners perform activities with
relatively synchronized spatiotemporal patterns, many of
these scanners were incorrectly detected in the results of this
study. Therefore, we must consider how to distinguish such
harmless investigative scanners from malware activities.
As mentioned in the previous subsection, DANTE [24]
embeds port sequences of darknet traffic by source host in
a given time frame and performs clustering. By comparing
with the previous time frame and labeling the clusters,
it is possible to track campaigns and detect recurrent or
new attacks. In large-scale scans using scanning tools,
such as ZMap or Masscan, or large-scale scans using
malware, fingerprints are often attached to packet headers
to perform faster scans [46]. It has also been reported
in [46] that fingerprints are provided to distinguish scan
results from backscatters. In contrast, Tanaka et al. proposed
VOLUME 10, 2022 13055
C. Han et al.:Dark-TRACER: Early Detection Framework for Malware Activity
TABLE 10. Summary of related works on identifying investigative scanners.
a method based on a genetic algorithm to automatically
identify fingerprints embedded in TCP/IP headers from
darknet traffic [70]. They succeeded in identifying unknown
fingerprints from data corresponding to a short period.
Identifying the fingerprints of investigative scanners and
tracing the scanners using the DANTE mechanism may
enable us to distinguish scanners from malware activities.
Additionally, Wan et al. clarified that the coverage rate
differs depending on the scan source and that the observed
hosts differ depending on the region of the observation
network [71]. Analyzing the darknet observation networks
in various regions, such as the autonomous system (AS) and
in various countries, is a way to obtain a more detailed and
precise understanding of the actual scanning situation.
In this study, we introduced three independent machine
learning methods to automatically estimate the synchro-
nization of the spatiotemporal patterns of darknet traffic
in real time and to detect anomalies. Those three methods
are: Dark-GLASSO,Dark-NMF, and Dark-NTD. We also
proposed Dark-TRACER, which integrates all three methods
into a single framework. We found that Dark-TRACER
was able to complement the weaknesses of each module,
achieving a 100% recall rate and detecting all malware
activities in the experiment. It detected the malware on
average 153.6 days earlier than the time when the threats
were revealed to the public by reputable third-party security
research organizations. In addition, we found that two
analysts could perform the daily operations necessary to
detect these threats in approximately 7.3 h.
Currently, our most serious challenge is the large number
of false positives. In this study, we confirmed that even
a simple rule-based approach can effectively reduce the
number of false-positive alerts. As described in Sections V-
D and VI-C, our future work is to reduce the number of
false positives by identifying the fingerprints of investigative
scanners and building a model to track them. By reducing the
number of false positives, the analysis cost can be lowered.
In addition, we intend to automate the secondary collision
analysis mentioned in Section V-E to elucidate the causes
and details of the alerts detected by Dark-TRACER. Finally,
we plan to deploy Dark-TRACER in the real world and
detect threats and malware activities in real-time to aid rapid
The authors would like to thank an Associate Professor
Katsunari Yoshioka from Yokohama National University
and Prof. Noboru Murata from Waseda University for their
valuable comments.
[1] G. Gu, J. Zhang, and W. Lee, ‘‘BotSniffer: Detecting botnet command and
control channels in network traffic,’’ in Proc. Netw. Distrib. Syst. Secur.
Symp. (NDSS), 2008, pp. 1–19.
[2] M. Bailey, E. Cooke, F. Jahanian, A. Myrick, and S. Sinha, ‘‘Practical
darknet measurement,’’ in Proc. 40th Annu. Conf. Inf. Sci. Syst., Mar. 2006,
pp. 1496–1501.
[3] J. Friedman, T. Hastie, and R. Tibshirani, ‘‘Sparse inverse covariance esti-
mation with the graphical lasso,’Biostatistics, vol. 9, no. 3, pp. 432–441,
Dec. 2007.
[4] D. Lee and H. S. Seung, ‘‘Algorithms for non-negative matrix factoriza-
tion,’’ in Proc. 13th Int. Conf. Neural Inf. Process. Syst. (NIPS), 2000,
pp. 535–541.
[5] Y.-D. Kim and S. Choi, ‘‘Nonnegative tucker decomposition,’’ in Proc.
IEEE Conf. Comput. Vis. Pattern Recognit., Jun. 2007, pp. 1–8.
[6] C. Han, J. Shimamura, T. Takahashi, D. Inoue, M. Kawakita, J. Takeuchi,
and K. Nakao, ‘‘Real-time detection of malware activities by analyzing
darknet traffic using graphical lasso,’’ in Proc. 18th IEEE Int. Conf. Trust,
Secur. Privacy Comput. Commun. (TrustCom), Aug. 2019, pp. 144–151.
[7] C. Han, J. Shimamura, T. Takahashi, D. Inoue, J. Takeuchi, and K. Nakao,
‘‘Real-time detection of global cyberthreat based on darknet by estimating
anomalous synchronization using graphical lasso,’IEICE Trans. Inf. Syst.,
vol. 103, no. 10, pp. 2113–2124, Oct. 2020.
[8] C. Han, J. Takeuchi, T. Takahashi, and D. Inoue, ‘‘Automated detection of
malware activities using nonnegative matrix factorization,’’ in Proc. IEEE
Int. Conf. Trust, Secur. Privacy Comput. Commun. (TrustCom), Oct. 2021.
[9] H. Kanehara, Y. Murakami, J. Shimamura, T. Takahashi, D. Inoue,
and N. Murata, ‘‘Real-time botnet detection using nonnegative tucker
decomposition,’’ in Proc. 34th ACM/SIGAPP Symp. Appl. Comput.,
Apr. 2019, pp. 1337–1344.
[10] J. Takeuchi and K. Yamanishi, ‘‘A unifying framework for detecting
outliers and change points from time series,’IEEE Trans. Knowl. Data
Eng., vol. 18, no. 4, pp. 482–492, Apr. 2006.
[11] Z. Durumeric, D. Adrian, A. Mirian, M. Bailey, and J. A. Halderman,
‘‘A search engine backed by internet-wide scanning,’’ in Proc. 22nd ACM
SIGSAC Conf. Comput. Commun. Secur., 2015, pp. 542–553.
[12] T. Ide, A. Khandelwal, and J. Kalagnanam, ‘‘Sparse Gaussian Markov
random field mixtures for anomaly detection,’’ in Proc. IEEE 16th Int.
Conf. Data Mining (ICDM), Dec. 2016, pp. 955–960.
[13] A. J. Gibberd and J. D. B. Nelson, ‘‘High dimensional changepoint
detection with a dynamic graphical lasso,’’ in Proc. IEEE Int. Conf.
Acoust., Speech Signal Process. (ICASSP), May 2014, pp. 2684–2688.
[14] T. Idé, A. C. Lozano, N. Abe, and Y. Liu, ‘‘Proximity-based anomaly
detection using sparse structure learning,’’ in Proc. SIAM Int. Conf. Data
Mining, Apr. 2009, pp. 97–108.
[15] S. Liu, T. Suzuki, and M. Sugiyama, ‘‘Supportconsistency of direct sparse-
change learning in Markov networks,’’ in Proc. 29th AAAI Conf. Artif.
Intell., 2015, pp. 2785–2791.
[16] Y. Koren, R. Bell, and C. Volinsky, ‘‘Matrix factorization techniques
for recommender systems,’IEEE Comput., vol. 42, no. 8, pp. 30–37,
Aug. 2009.
13056 VOLUME 10, 2022
C. Han et al.:Dark-TRACER: Early Detection Framework for Malware Activity
[17] Q. Zhao, C. F. Caiafa, D. P. Mandic, L. Zhang, T. Ball,
A. Schulze-Bonhage, and A. and Cichocki, ‘‘Multilinear subspace
regression: An orthogonal tensor decomposition approach,’’ in Proc. 25th
Annu. Conf. Neural Inf. Process. Syst., 2011, pp. 1269–1277.
[18] A. H. Phan and A. Cichocki, ‘‘Tensordecompositions for feature extraction
and classification of high dimensional datasets,’IEICE Nonlinear Theory
Appl., vol. 1, no. 1, pp. 37–68, 2010.
[19] A. Anandkumar, P. Jain, Y. Shi, and U. N. Niranjan, ‘‘Tensor vs. matrix
methods: Robust tensor decomposition under block sparse perturbations,’
in Proc. 19th Int. Conf. Artif. Intell. Statist., (AISTATS), vol. 51, 2016,
pp. 268–276.
[20] C. F. Caiafa and A. Cichocki, ‘‘Generalizing the column–row matrix
decomposition to multi-way arrays,’Linear Algebra its Appl., vol. 433,
no. 3, pp. 557–573, Sep. 2010.
[21] G. Zhou, A. Cichocki, Q. Zhao, and S. Xie, ‘‘Efficient nonnegative
tucker decompositions: Algorithms and uniqueness,’IEEE Trans. Image
Process., vol. 24, no. 12, pp. 4990–5003, Dec. 2015.
[22] M. Antonakakis, T. April, M. Bailey, M. Bernhard, E. Bursztein,
J. Cochran, Z. Durumeric, J. A. Halderman, L. Invernizzi,
M. Kallitsis, D. Kumar, C. Lever, Z. Ma, J. Mason, D. Menscher,
C. Seaman, N. Sullivan, K. Thomas, and Y. and Zhou, ‘‘Understanding the
Mirai botnet,’’ in Proc. 26th USENIX Secur. Symp., 2017, pp. 1093–1110.
[23] H. Griffioen and C. Doerr, ‘‘Examining Mirai’s battle over the Internet
of Things,’’ in Proc. ACM SIGSAC Conf. Comput. Commun. Secur.,
Oct. 2020, pp. 743–756.
[24] D. Cohen, Y. Mirsky, M. Kamp, T. Martin, Y. Elovici, R. Puzis, and
A. Shabtai, ‘‘DANTE: A framework for mining and monitoring darknet
traffic,’’ in Proc. 25th Eur. Symp. Res. Comput. Secur. (ESORICS).
Springer, 2020, pp. 88–109.
[25] FOX-IT. (2016). Recent Vulnerability in Eir D1000 Router Used to
Spread Updated Version of Mirai DDoS Bot. [Online]. Available:
[26] S. Herwig, K. Harvey, G. Hughey, R. Roberts, and D. Levin, ‘‘Measure-
ment and analysis of hajime, a peer-to-peer IoT botnet,’’ in Proc. Netw.
Distrib. Syst. Secur. Symp., 2019, pp. 1–15.
[27] Netlab 360. (2018). HNS Botnet Recent Activities. [Online]. Available:
[28] Netlab 360. (2018). 7,500+MikroTik Routers are Forwarding
Owners’ Traffic to the Attackers, How is Yours?. [Online]. Available:
owners% -traffic-to-the-attackers-how-is-yours-en/
[29] Huawei. (2017). Security Notice—Statement on Remote Code Execution
Vulnerability in Huawei HG532 Product. [Online]. Available:
01-% hg532-en
[30] Tenable Research Blog. (2018). Tenable Research Advisory: Peekaboo
Critical Vulnerability in NUUO Network Video Recorder. [Online]. Avail-
critica% l-vulnerability-in-nuuo-network-video-recorder
[31] Netlab 360. (2018). BCMPUPnP_Hunter: A 100k Botnet
Turns Home Routers to Email Spammers. [Online]. Available:
home-ro% uters-to-email-spammers-en/
[32] N. Otsu, ‘‘A threshold selection method from gray-level histograms,’’IEEE
Trans. Syst., Man, Cybern. Syst., vol. SMC-9, no. 1, pp. 62–66, Jan. 1979.
[33] I. Ilascu. (2019). New Echobot Botnet Variant Uses
Over 50 Exploits to Propagate. [Online]. Available:
varia% nt-uses-over-50-exploits-to-propagate/
[34] Netlab 360. (2019). The Botnet Cluster on the [Online].
[35] Microsoft. (2019). Remote Desktop Services Remote Code Execu-
tion Vulnerability.[Online]. Available:
guide/en-U.S./vulnerability/CVE-2019-07% 08
[36] Bashis. (2018). Shenzhen TVT Digital Technology co.
ltd & oem DVR/NVR/IPC API RCE. [Online]. Available:
[37] V. Yarmak. (2020). Full Disclosure: 0day Vulnerability (Backdoor) in
Firmware for Xiaongmai-Based DVRs, NVRs and IP Cameras. [Online].
[38] K. Hsu, H. Zhang, Z. Zhang, and R. Nigam. (2020). Grandstream
and Draytek Devices Exploited to Power New Hoaxcalls DDoS Botnet.
[Online]. Available:
[39] Pierre. (2020). Multiple Vulnerabilities found in Zyxel CNM Secuman-
ager. [Online]. Available:
zyxel-secumanager-0day-vuln% erabilities.html
[40] F-Secure. (2020). Saltstack Authorization Bypass. [Online]. Available:
[41] REW. Linksys e-Series—Remote Code Execution. [Online]. Available:, 2014.
[42] J. Lands. (2020). Priority Threat Actors Adopt Mirai Source Code.
[Online]. Available:
[43] M. Althibyani. (2020). Oracle Weblogic Server—Unauthenticated
rce Via Get Request. [Online]. Available: https://www.exploit-
[44] Z. Durumeric, E. Wustrow, and J. A. Halderman, ‘‘Zmap: Fast internet-
wide scanning and its security applications,’’ in Proc. 22th USENIX Secur.
Symp., 2013, pp. 605–620.
[45] J. Mazel, R. Fontugne, and K. Fukuda, ‘‘Profiling internet scanners:
Spatiotemporal structures and measurement ethics,’’ in Proc. Netw. Traffic
Meas. Anal. Conf. (TMA), Jun. 2017, pp. 1–9.
[46] H. Griffioen and C. Doerr, ‘‘Discovering collaboration: Unveiling slow,
distributed scanners based on common header field patterns,’’ in Proc.
NOMS IEEE/IFIP Netw. Oper. Manage. Symp., Apr. 2020, pp. 1–9.
[47] T. Takahashi, Y. Umemura, C. Han, T. Ban, K. Furumoto, O. Nakamura,
K. Yoshioka, J. Takeuchi, N. Murata, and Y. Shiraishi, ‘‘Designing
comprehensive cyber threat analysis platform: Can we orchestrate
analysis engines?’’ in Proc. IEEE Int. Conf. Pervasive Comput. Com-
mun. Workshops Affiliated Events (PerCom Workshops), Mar. 2021,
pp. 376–379.
[48] R. Tanabe, T. Tamai, A. Fujita, R. Isawa, K. Yoshioka, T. Matsumoto,
C. Gañán, and M. van Eeten, ‘‘Disposable botnets: Examining the anatomy
of IoT botnet infrastructure,’’ in Proc. 15th Int. Conf. Availability, Rel.
Secur., Aug. 2020, pp. 7:1–7:10.
[49] A. Dainotti, K. Benson, A. King, K. Claffy, M. Kallitsis, E. Glatz, and
X. Dimitropoulos, ‘‘Estimating internet address space usage through
passive measurements,’’ ACM SIGCOMM Comput. Commun. Rev., vol. 44,
no. 1, pp. 42–49, Dec. 2013.
[50] Z. Durumeric, M. Bailey, and J. A. Halderman, ‘‘An internet-wide view
of internet-wide scanning,’’ in Proc. 23rd USENIX Secur. Symp., 2014,
pp. 65–78.
[51] C. Fachkha, E. Bou-Harb, A. Keliris, N. Memon, and M. Ahamad,
‘‘Internet-scale probing of CPS: Inference, characterization and orchestra-
tion analysis,’’ in Proc. Netw. Distrib. Syst. Secur. Symp., 2017, pp. 1–15.
[52] M. Jonker, A. King, J. Krupp, C. Rossow, A. Sperotto, and
A. Dainotti, ‘‘Millions of targets under attack,’’ in Proc. Internet
Meas. Conf., Nov. 2017, pp. 100–113, doi: 10.1145/3131365.3131383.
[53] F. Shaikh, E. Bou-Harb, J. Crichigno, and N. Ghani, ‘‘A machine learning
model for classifying unsolicited IoT devices by observing network
telescopes,’’ in Proc. 14th Int. Wireless Commun. Mobile Comput. Conf.
(IWCMC), Jun. 2018, pp. 938–943.
[54] R. Akiyoshi, D. Kotani, and Y. Okabe, ‘‘Detecting emerging large-scale
vulnerability scanning activities by correlating low-interaction honeypots
with darknet,’’ in Proc. IEEE 42nd Annu. Comput. Softw. Appl. Conf.
(COMPSAC), Jul. 2018, pp. 658–663.
[55] C. Fachkha and M. Debbabi, ‘‘Darknet as a source of cyber intelligence:
Survey, taxonomy, and characterization,’’ IEEE Commun. Surveys Tuts.,
vol. 18, no. 2, pp. 1197–1227, 2nd Quart., 2016.
[56] M. Bailey, E. Cooke, F. Jahanian, and J. Nazario, ‘‘The internet motion
sensor—A distributed blackhole monitoring system,’’ in Proc. Netw.
Distrib. Syst. Secur. Symp. (NDSS), 2005, pp. 167–179.
[57] V. Yegneswaran, P. Barford, and D. Plonka, ‘‘On the design and use of
internet sinks for network abuse monitoring,’’ in Recent Adv. Intrusion
Detection, 7th Int. Symp. (RAID), vol. 3224, 2004, pp. 146–165.
[58] D. Moore, ‘‘Network telescopes: Tracking denial-of-service attacks and
internet worms around the globe,’’ in Proc. 17th Conf. Syst. Admin. (LISA),
2003. [Online]. Available:
[59] E. Ahmed, A. Clark, and G. Mohay, ‘‘A novel sliding window based change
detection algorithm for asymmetric traffic,’’ in Proc. IFIP Int. Conf. Netw.
Parallel Comput., Oct. 2008, pp. 168–175.
VOLUME 10, 2022 13057
C. Han et al.:Dark-TRACER: Early Detection Framework for Malware Activity
[60] D. Inoue, K. Yoshioka, M. Eto, M. Yamagata, E. Nishino, J. Takeuchi,
K. Ohkouchi, and K. Nakao, ‘‘An incident analysis system NICTER and
its analysis engines based on data mining techniques,’’ in Proc. Neural Inf.
Process. 15th Int. Conf. (ICONIP), vol. 5506, 2008, pp. 579–586.
[61] T. Ban, L. Zhu, J. Shimamura, S. Pang, D. Inoue, and K. Nakao, ‘‘Detection
of botnet activities through the lens of a large-scale darknet,’’ in Neural
Information Processing. Cham, Switzerland: Springer, 2017, pp. 442–451,
doi: 10.1007/978-3-319-70139-4_45.
[62] M. S. Pour, A. Mangino, K. Friday, M. Rathbun, E. Bou-Harb, F. Iqbal,
S. Samtani, J. Crichigno, and N. Ghani, ‘‘On data-driven curation, learning,
and analysis for inferring evolving Internet-of-Things (IoT) botnets in the
wild,’Comput. Secur., vol. 91, Apr. 2020, Art. no. 101707.
[63] S. Torabi, E. Bou-Harb, C. Assi, E. B. Karbab, A. Boukhtouta, and
M. Debbabi, ‘‘Inferring and investigating IoT-generated scanning cam-
paigns targeting a large network telescope,’’ IEEE Trans. Dependable
Secure Comput., vol. 19, no. 1, pp. 402–418, Jan. 2020.
[64] G. Gu, R. Perdisci, J. Zhang, and W. Lee, ‘‘Botminer: Clustering analysis of
network traffic for protocol- and structure-independent botnet detection,’’
in Proc. 17th USENIX Secur. Symp., 2008, pp. 139–154.
[65] P. S. Joshi and H. Dinesha, ‘‘Survey on identification of malicious activities
by monitoring darknet access,’’ in Proc. 3rd Int. Conf. Smart Syst. Inventive
Technol. (ICSSIT), Aug. 2020, pp. 346–350.
[66] S. Kumar, H. Vranken, J. van Dijk, and T. Hamalainen, ‘‘Deep in the dark:
A novel threat detection system using darknet traffic,’’ in Proc. IEEE Int.
Conf. on Big Data (Big Data), Dec. 2019, pp. 4273–4279.
[67] E. Bou-Harb, M. Debbabi, and C. Assi, ‘‘A time series approach for
inferring orchestrated probing campaigns by analyzing darknet traffic,’’ in
Proc. 10th Int. Conf. Availability, Rel. Secur., Aug. 2015, pp. 180–185.
[68] S. H. A. Ali, S. Ozawa, T. Ban, J. Nakazato, and J. Shimamura, ‘‘A neural
network model for detecting DDoS attacks using darknet traffic features,’’
in Proc. Int. Joint Conf. Neural Netw. (IJCNN), Jul. 2016, pp. 2979–2985.
[69] R. Graham. (2013). Masscan: The Entire Internet in 3 Minutes,
Errata Security Blog. [Online]. Available: https://blog.erratasec.
[70] A. Tanaka, C. Han, T. Takahashi, and K. Fujisawa, ‘‘Internet-wide scanner
fingerprint identifier based on TCP/IP header,’’ in Proc. 6th Int. Conf. Fog
Mobile Edge Comput. (FMEC), Dec. 2021.
[71] G. Wan, L. Izhikevich, D. Adrian, K. Yoshioka, R. Holz, C. Rossow, and
Z. Durumeric, ‘‘On the origin of scanning: The impact of location on
internet-wide scans,’’ in Proc. ACM Internet Meas. Conf., Oct. 2020,
pp. 662–679.
CHANSU HAN (Member, IEEE) received the
B.E. degree in computer science and the M.S.
and Ph.D. degrees in informatics engineering
from Kyushu University, in 2016, 2018, and
2021, respectively. He is currently a Researcher
with the National Institute of Information and
Communications Technology (NICT), Japan. His
research interests include analyzing and solving
problems in the cybersecurity field (especially
networks and malware) using machine learning.
JUN’ICHI TAKEUCHI (Member, IEEE) received
the B.Sc. degree in physics and the Dr.Eng. degree
in mathematical engineering from the Univer-
sity of Tokyo, in 1989 and 1996, respectively.
From 1989 to 2006, he worked with NEC Corpora-
tion, Japan. In 2006, he joined Kyushu University,
Fukuoka, Japan, where he is currently a Professor
in mathematical engineering. From 1996 to 1997,
he was a Visiting Research Scholar with the
Department of Statistics, Yale University, New
Haven, CT, USA. His research interests include mathematical statistics,
information geometry, information theory, data science, and machine
learning. He is a member of IEICE and JSIAM.
the Ph.D. degree in telecommunications from
Waseda University, in 2005. He was with the Tam-
pere University of Technology as a Researcher,
from 2002 to 2004, and Roland Berger Ltd., as a
Business Consultant, from 2005 to 2009. Since
2009, he has been with the National Institute
of Information and Communications Technol-
ogy, where he is currently a Research Manager.
His research interests include cybersecurity and
machine learning.
DAISUKE INOUE (Member, IEEE) received the
B.E. and M.E. degrees in electrical and computer
engineering and the Ph.D. degree in engineering
from Yokohama National University, in 1998,
2000, and 2003, respectively. He is currently the
Director of the Cybersecurity Laboratory, National
Institute of Information and Communications
Technology (NICT). He has received several
awards, including the Asia-Pacific Information
Security Leadership Achievements (ISLA),
in 2014.
13058 VOLUME 10, 2022
... Hosts infected with a similar type of malware that shares a scanning module tend to perform scans in a similar spatiotemporal pattern, indicating synchronization. Focusing on such synchronization by malware, we have proposed an anomaly detection framework called Dark-TRACER for early detection of indiscriminate scanning attack activities originating from malware before an outbreak of a cyberattack [2]. ...
... In our previous study [2], we found that Dark-TRACER detected critical cyber threats on average 126.4 days earlier than when they were first observed in NICTER. However, we have not been able to verify whether there is a correlation between these early detections and the actual threats. ...
... If we can distinguish and filter investigative scan activities, we should be able to focus on critical threats and scan activities exclusively. This paper outlines four ways of discriminating between investigative and offensive scanners: (1) automatic fingerprint identification of scanner-specific characteristics, (2) clustering of source hosts based on destination port sets, (3) tracing of scan campaigns, and (4) prediction of scan campaigns. Our goal is to build a synergistic hybrid darknet analysis system by incorporating the findings of these studies in Dark-TRACER as functions other than anomaly detection. ...
... The following is a brief description of our research activities on this challenge. To solve this issue, we developed an AI-based solution that automatically detects the rise of new malware scanning activities earlier using a large-scale darknet environment (an observation environment using unused IP addresses) [9], [10]. As shown in Fig. 6, this method focuses on the fact that scanning behavior from hosts infected with the same malware exhibits synchronized characteristics in terms of scanning methods, timing, etc., since the same scan transmission module is used to perform the scan. ...
... As shown in Fig. 6, this method focuses on the fact that scanning behavior from hosts infected with the same malware exhibits synchronized characteristics in terms of scanning methods, timing, etc., since the same scan transmission module is used to perform the scan. In other words, when synchronized scan behaviors are detected on the darknet, it can be determined that the group of hosts emitting those scans are infected with the same malware [10]. By using this method, we believe it is possible to catch signs of new malware activity at the earliest possible stage, detect infected host groups, and link this to incident response, including guiding vulnerable equipment users to countermeasures to minimize the damage caused by the malware. ...
In this paper, we developed the latest IoT honeypots to capture IoT malware currently on the loose, analyzed IoT malware with new features such as persistent infection, developed malware removal methods to be provided to IoT device users. Furthermore, as attack behaviors using IoT devices become more diverse and sophisticated every year, we conducted research related to various factors involved in understanding the overall picture of attack behaviors from the perspective of incident responders. As the final stage of countermeasures, we also conducted research and development of IoT malware disabling technology to stop only IoT malware activities in IoT devices and IoT system disabling technology to remotely control (including stopping) IoT devices themselves.
... The DN environment [11] offers a hidden segment of online space for conducting CyCr activities. DN monitoring enhances traditional malware defenses by mining and finding behavioural regularities of Vulnerability Originators (VOs). ...
... In malware communication, ARL [11] may be used in processing (transaction) sets characterized by indications of regularity. We picked three indicators of interest to study among the numerous variables presented in the IP Headers (IP and TCP) of packets: Destination_Port(DP), WS and Sequence_Number(SN). ...
The Mirai botnet source code, which was distributed on the dark web, is still changing as malware designers modify it to produce more sophisticated varieties of the botnet—the Satori—based on the malware source code. An botnet that is tied to the Internet of Things, infested with malware, and administered by offenders creates a backdoor for cybercriminals to carry out destructive operations throughout the system. The work provides association rule learning, which reveals design characteristics and warning indications of hazards using data (a data stream) collected for darknet vulnerability research. The negative effects of threat-originating computers associated with malware scripts are investigated by looking at trends in IoT-linked applications (like destination port, TCP window lengths, and kind of service).This paper provides an study of an attack (both before and after) with the first source scripts development of a conventional IoT malware, the Satori, as well as an uncommon insight into the viral processes of an attack campaign. Trials demonstrate that the suggested system is efficient in anticipating and tracking new virus tendencies in cyberspace, demonstrating that it is a viable way for streamlining and accelerating the discovery of new illicit patterns in online activities. Satori-based mathematical modelling of destination ports, ToS, and TCP window size is provided with NICT/16 DN sensor dataset that is employed in the research.
... The top-ranked aimed ports currently, 6379-TCP 23-TCP 22-TCP 443-TCP 3389-TCP 2375-TCP 80-TCP 123-UDP 5060-UDP 8088-TCP 81-TCP 2376-TCP 389-UDP 53-UDP 8080-TCP 5555-TCP 8443-TCP 10443-TCP 0-ICMP 445-TCP 8545-TCP 5900-TCP 9200-TCP 161- This result is somewhat expected if one considers the rise of IoT applications and the botnet families that target Telnet services on IoT devices (e.g., see [47]). Further, Redis vulnerabilities are recently popularly mined for Cryptojacking [8] and other application-level attacks [25]. Looking at Figure 3 in [17], we also notice that ZMap/Masscan currently play a prominent role in Internetwide scanning whereas in 2014 their presence was minimal (as expected, since they were relatively unknown tools then). ...
Full-text available
Aggressive network scanners, i.e., ones with immoderate and persistent behaviors, ubiquitously search the Internet to identify insecure and publicly accessible hosts. These scanners generally lie within two main categories; i) benign research-oriented probers; ii) nefarious actors that forage for vulnerable victims and host exploitation. However, the origins, characteristics and the impact on real networks of these aggressive scanners are not well understood. In this paper, via the vantage point of a large network telescope, we provide an extensive longitudinal empirical analysis of aggressive IPv4 scanners that spans a period of almost two years. Moreover, we examine their network impact using flow and packet data from two academic ISPs. To our surprise, we discover that a non-negligible fraction of packets processed by ISP routers can be attributed to aggressive scanners. Our work aims to raise the network community's awareness for these "heavy hitters", especially the miscreant ones, whose invasive and rigorous behavior i) makes them more likely to succeed in abusing the hosts they target and ii) imposes a network footprint that can be disruptive to critical network services by incurring consequences akin to denial of service attacks.
... Data coming from darknets have helped profiling attack strategies [30,37,44,53], detecting and characterizing Internet scans [28,29,48] and studying malware spread [54]. For instance, authors of [35,55] rely on genetic algorithms to automatically modify the fingerprints of attacks and automatically identify new variants. Such prior studies rely on ad-hoc algorithms to characterize darknet traffic and have repeatedly proved the value of such networks. ...
Darknets are probes listening to traffic reaching IP addresses that host no services. Traffic reaching a darknet results from the actions of internet scanners, botnets and possibly misconfigured hosts. Such peculiar nature of the darknet traffic makes darknets a valuable instrument to discover malicious online activities, e.g., identifying coordinated actions performed by bots or scanners. However, the massive amount of packets and sources that darknets observe makes it hard to extract meaningful insights, calling for scalable tools to automatically identify and group sources that share similar behaviour. We here present i-DarkVec, a methodology to learn meaningful representations of Darknet traffic. i-DarkVec leverages Natural Language Processing techniques (e.g., Word2Vec) to capture the co-occurrence patterns that emerge when scanners or bots launch coordinated actions. As in NLP problems, the embeddings learned with i-DarkVec enable several new machine learning tasks on the darknet traffic, such as identifying clusters of senders engaged in similar activities. We extensively test i-DarkVec and explore its design space in a case study using real darknets. We show that with a proper definition of services , the learned embeddings can be used to (i) solve the classification problem to associate unknown sources’ IP addresses to the correct classes of coordinated actors, and (ii) automatically identify clusters of previously unknown sources performing similar attacks and scans, easing the security analyst’s job. i-DarkVec leverages a novel incremental embedding learning approach that is scalable and robust to traffic changes, making it applicable to dynamic and large-scale scenarios.
... Their system could detect orchestrated source IP addresses together with the target destination ports. Han et al. [50] finally integrated graphical lasso [25], NMF [27], and NTD [29] into an anomaly detection framework Dark-TRACER and achieved a 100% recall rate on darknet traffic data (up to /17 subnet scales) from October 2018 to 2020. ...
Full-text available
Adversaries perform port scanning to discover accessible and vulnerable hosts as a prelude to cyber havoc. A darknet is a cyberattack observation network to capture these scanning activities through reachable yet unused IP addresses. However, the enormous amount of packets and superposition of diverse scanning strategies prevent extracting significant insights from the aggregate traffic. Some coordinated scanners disperse probe packets whose TCP/IP header follows a unique pattern to determine whether the received packets are valid responses to their probes or are part of other background traffic. We call such a pattern a fingerprint. For example, a probe packet from a Mirai-infected host satisfies a pattern whereby the destination IP address equals the sequence number. A fingerprint indicates that the source host has been involved in a particular scanning campaign. Although some fingerprints have been discovered and known to the public, there are and will be more undiscovered ones. We intend to unveil these fingerprints. Our preliminary work automatically identified flexible fingerprints but overlooked low-rate and coordinated scanners. In this work, we improved the fingerprint identifier, enabling it to detect these stealth scans. Moreover, we revealed the scans’ objectives by investigating destination port sets. We associated fingerprints with threat intelligence and verified their reliability. Our approach identified all well-known and eight unknown fingerprints on one month’s worth of darknet data collected from about three-hundred thousand unused IP addresses. We disclosed the fingerprints of the Mozi botnet and destination port sets that were previously unreported.
... threats [Endo et al., 2022]. We previously developed a synchronization anomaly detection framework called Dark-TRACER with the goal of detecting malware-induced indiscriminate scanning attack activities before an attack goes viral [Han et al., 2022, Han et al., 2020. We found that while Dark-TRACER detected threats an average of 126.4 days earlier than when they were first observed and recognized by NICTER, it also detected many false positive alerts that were unrelated to (or had little to do with) attack activities. ...
Full-text available
The Mirai botnet, composed primarily of embedded and IoT devices, took the Internet by storm in late 2016 when it overwhelmed several high-profile targets with massive distributed denial-of-service (DDoS) attacks. In this paper, we provide a seven-month retrospective analysis of Mirai's growth to a peak of 600k infections and a history of its DDoS victims. By combining a variety of measurement perspectives, we analyze how the bot-net emerged, what classes of devices were affected, and how Mirai variants evolved and competed for vulnerable hosts. Our measurements serve as a lens into the fragile ecosystem of IoT devices. We argue that Mirai may represent a sea change in the evolutionary development of botnets-the simplicity through which devices were infected and its precipitous growth, demonstrate that novice malicious techniques can compromise enough low-end devices to threaten even some of the best-defended targets.
Full-text available
With the rapid evolution and increase of cyberthreats in recent years, it is necessary to detect and understand it promptly and precisely to reduce the impact of cyberthreats. A darknet, which is an unused IP address space, has a high signal-to-noise ratio, so it is easier to understand the global tendency of malicious traffic in cyberspace than other observation networks. In this paper, we aim to capture global cyberthreats in real time. Since multiple hosts infected with similar malware tend to perform similar behavior, we propose a system that estimates a degree of synchronizations from the patterns of packet transmission time among the source hosts observed in unit time of the darknet and detects anomalies in real time. In our evaluation, we perform our proof-of-concept implementation of the proposed engine to demonstrate its feasibility and effectiveness, and we detect cyberthreats with an accuracy of 97.14%. This work is the first practical trial that detects cyberthreats from in-the-wild darknet traffic regardless of new types and variants in real time, and it quantitatively evaluates the result.
Trillions of network packets are sent over the Internet to destinations which do not exist. This ‘darknet’ traffic captures the activity of botnets and other malicious campaigns aiming to discover and compromise devices around the world. In this paper, we present DANTE: a framework and algorithm for mining darknet traffic. DANTE learns the meaning of targeted network ports by applying Word2Vec to observed port sequences. To detect recurring behaviors and new emerging threats, DANTE uses a novel and incremental time-series cluster tracking algorithm on the observed sequences. To evaluate the system, we ran DANTE on a full year of darknet traffic (over three Tera-Bytes) collected by the largest telecommunications provider in Europe, Deutsche Telekom and analyzed the results. DANTE discovered 1,177 new emerging threats and was able to track malicious campaigns over time.