=Paper=
{{Paper
|id=Vol-3329/paper-04
|storemode=property
|title=Federated Learning as Enabler for Collaborative Security between not Fully-Trusting Distributed Parties
|pdfUrl=https://ceur-ws.org/Vol-3329/paper-04.pdf
|volume=Vol-3329
|authors=Leo Lavaur,Benjamin Coste,Marc-Oliver Pahl,Yann Busnel,Fabien Autrel
}}
==Federated Learning as Enabler for Collaborative Security between not Fully-Trusting Distributed Parties==
https://ceur-ws.org/Vol-3329/paper-04.pdf
Federated Learning as enabler for Collaborative
Security between not Fully-Trusting Distributed
Parties
Léo Lavaur1,2 , Benjamin Costé3 , Marc-Oliver Pahl1,2 , Yann Busnel1 and
Fabien Autrel1
1
IMT-Atlantique, IRISA
2
Chaire Cybersécurité des Infrastructures Critiques (Cyber CNI)
3
Airbus CyberSecurity
Abstract
Literature shows that trust typically relies on knowledge about the communication partner. Federated
learning is an approach for collaboratively improving machine learning models. It allows collaborators
to share Machine Learning models without revealing secrets, as only the abstract models and not the
data used for their creation is shared. Federated learning thereby provides a mechanism to create trust
without revealing secrets, such as specificities of local industrial systems.
A fundamental challenge, however, is determining how much trust is justified for each contributor
to collaboratively optimize the joint models. By assigning equal trust to each contribution, divergence
of a model from its optimum can easily happen—caused by errors, bad observations, or cyberattacks.
Trust also depends on how much an aggregated model contributes to the objectives of a party. For
example, a model trained for an OT system is typically useless for monitoring IT systems.
This paper shows first directions how heterogeneous distributed data sources could be integrated
using federated learning methods. With an extended abstract, it shows current research directions and
open issues from a cyber-analyst’s perspective.
Keywords
Federated learning, cybersecurity, intrusion detection, distributed trust
1. Introduction
A common cybersecurity goal is thwarting attackers through detection of their actions, un-
derstanding of their methodologies, and increasing the resilience before their next attempts.
However, the considerable variety, thus complexity, of the tools and techniques used by at-
tackers drown their traces in the high volume of legitimate network traffic and endpoints logs.
One way used by defenders (often called “Blue teams”) to act against threat actors is sharing
knowledge about their abuses. However, knowledge gathered during security monitoring or
C&ESAR’22: Computer & Electronics Security Application Rendezvous, Nov. 15-16, 2022, Rennes, France
£ leo.lavaur@imt-atlantique.fr (L. Lavaur); benjamin.b.coste@airbus.com (B. Costé);
marc-oliver.pahl@imt-atlantique.fr (M. Pahl); yann.busnel@imt-atlantique.fr (Y. Busnel);
fabien.autrel@imt-atlantique.fr (F. Autrel)
DZ 0000-0002-0379-7946 (L. Lavaur); 0000-0002-5631-9120 (B. Costé); 0000-0001-5241-3809 (M. Pahl);
0000-0001-6908-719X (Y. Busnel); 0000-0002-8403-515X (F. Autrel)
© 2022 Copyright for this paper by its authors. Use permitted under Creative Commons License Attribution 4.0 International (CC BY 4.0).
CEUR
Workshop
Proceedings
http://ceur-ws.org
ISSN 1613-0073
CEUR Workshop Proceedings (CEUR-WS.org)
Proceedings of the 29th C&ESAR (2022) 65
Federated Learning as Enabler for Collaborative Security between not Full...
incident response is often coupled with private data which cannot be shared for confidentiality
reasons (GDPR, NDA, etc.).
Machine Learning (ML) approaches can help here, as they result in abstract models that are
typically not reversible to their input data [1]. A fundamental problem for observation-based
security is the amount of data needed for having a trustable and reliable impression. Typically,
more data allows for better behavior characterization, thus improving either anomaly detection,
or event classification. Collecting data requires either a long observation time or many data
sources, as comprehensiveness is difficult to reach.
Federated Learning (FL) has been introduced to enable the sharing of local models and to
federate them towards better joint models. Each participant computes on its own an ML-model
using its own data. The resulting model is aggregated with the ones of other participants,
typically by a trusted party, then the new model is shared between each of them.
Consequently, all parties benefit from each other while no one have access to their private
data and algorithms. Sharing models however still faces issues as models can be modified with
backdoors [2], poisoned with adversarial approaches [3] or simply suffers from bad quality
training dataset.
This paper discusses FL sharing approaches through aggregation issues. Section 2 introduces
the concept of Machine Learning (ML) based intrusion detection, and Federated Learning (FL)
as a collaboration enabler. In Section 3, we summarize the literature around federated intrusion
detection, with an extended abstract of a survey. Section 4 outlines an experimental use-case
for Federated Intrusion Detection System (FIDS) application. Section 5 discusses open issues
and the envisioned solutions. Section 6 concludes our proposition.
2. Background: ML for collaboratively defend cyberattacks
One prominent application of FL in the cybersecurity field is for security monitoring and col-
laborative intrusion detection. This section defines the basis for the remainder of the paper.
Security monitoring systems frequently use signature-based Intrusion Detection Systems
(IDSs) to detect known attacks to safeguard companies [4]. However, this strategy suffers
from serious limitations against zero-day and one-day attacks, as well as Advanced Persistent
Threats (APTs), such as Stuxnet [5]. Furthermore, the IoT’s heterogeneity and irregular traffic
cause IDSs to be less effective or inadequate [6]. As a result, researchers began to look into
anomaly detection as a way to improve detection systems.
Here, multiple approaches coexist, depending on objectives and available data. On the one
hand, anomaly-based detection systems compare monitored events to a baseline profile trained
on nominal traffic to assess whether they are malicious or not [7]. On the other hand, pattern-
based classification, aims at extracting patterns from known attacks that have been previ-
ously labelled as such, and then characterize input data according to the extracted patterns.
Therefore, anomaly-based approaches are particularly relevant to detect unknown behaviors,
whereas supervised ones are more equipped for threat characterization.
The source of data will also influence the available features for detection. For example,
endpoint-oriented sensors such as Sysmon monitor processes, use of windows’ registry key
or system calls while network-oriented sensors provides a lot of information about protocols,
66 Proceedings of the 29th C&ESAR (2022)
L. Lavaur, B. Coste, M.-O. Pahl et al
packet length, or IP addresses. Preprocessing can be used to extract additional features from the
raw data, to provide more information. The literature distinguishes three key non-exclusive
approaches: feature extraction, feature embedding, and feature selection [8].
In the context of knowledge sharing, the choice of the input feature is critical to transfer
knowledge between participants. For instance, IP addresses are specific to the local layout of
the network, and might have another meaning in another environment, if they even exist. On
the other hand, given one class of devices, inter-arrival time (IAT) should not vary much be-
tween networks, thus making potentially making it a good choice for transferring knowledge.
With a designated set of features and an input dataset, one can establish a model of the data.
A model is an abstraction that can be used afterward for other tasks, such as characterization.
Algorithmically speaking, a model is a set of mathematical parameters that are inferred from
input data by an algorithm. It can be the statistical parameters of a distribution function, the
condition nodes in a decision tree, or the weights and biases of a neural network. The ability
of a model to be shared partly depends on how easily these parameters can be aggregated.
Over the years, ML have been applied to intrusion detection with substantial results. Three
approaches coexist:
(1) Anomaly detection becomes a binary classification problem with supervised learning.
For effective training, a balanced labeled dataset is required. However, because local
training data is infrequently labeled and models can be affected by unbalanced class
distribution, supervised learning is more difficult to apply in real-world scenarios [9].
(2) For unlabeled data, unsupervised learning is more appropriate. In the case of IDS, we
assume that (i) benign traffic is substantially more common in the testing set than anoma-
lies [10]; and (ii) abnormal packets are statistically different from normal packets.
(3) Semi-supervised learning is a hybrid method that labels only a portion of the training
data. It can be used to bootstrap a detection model with a publicly labeled dataset before
training it on locally collected data afterward.
However, ML algorithms also require a lot of training data to avoid learning biased model,
from the lack of exhaustively or from an overrepresentation of a class in the training set.
To cope with the limitations of ML, especially when training data is locally collected, col-
laborative IDSs have emerged in the literature. However, they are almost always built in a
centralized manner, which induces its own set of issues: (i) centralizing a system typically in-
troduce a single point-of-failure (SPoF) [11]; (ii) centralized IDSs imply sending local data for
training or detection, increasing the risk of information disclosure [12]; (iii) communicating
data over networks also increase bandwidth consumption and latency, which are critical for
intrusion detection [13].
Introduced in 2016 by Google [14], FL promises to cope with these issues. In FL, model
learning is distributed among the participants of the federation. Therefore, local data stay in
the participant’s system, and collaboration is achieved by sharing and aggregating the gener-
ated models. Aggregation can be done by a server [6], [15]–[17], but it might lead to concerns
with trust and privacy. Due to challenges in terms of traceability, integrity, privacy, and trust,
recent research has favored the usage of trusted distributed ledgers [18], [19], multi-party com-
putation (MPC) [20], and privacy-preserving mechanisms like Differential Privacy (DP) [21].
Proceedings of the 29th C&ESAR (2022) 67
Federated Learning as Enabler for Collaborative Security between not Full...
3. The example of collaborative intrusion detection
Since its introduction, FL has been applied to multiple domains, such as intrusion detection,
whose presence is increasing in the literature. In this context, FL allows local detection of
attacks—thus offering low latency and bandwidth—while jointly enriching participant’s models
and preserving data privacy.
This section is an extended abstract of the study published in IEEE Transactions on Network
and Service Management (TNSM) [8]. Contributions of the study are as follows: (1) it examines
the application of FL to the detection and mitigation of attacks; (2) it proposes a reference
architecture that generalizes the selected work and can serve as a starting point for the design of
future FIDSs; (3) it establishes a taxonomy of FIDSs, which provides a framework for comparing
the selected works in this study; and (4) it highlights open questions regarding FIDSs and
identifies associated research directions. The findings presented in this study can easily be
extrapolated to other cybersecurity applications of FL, such as automated forensic analysis, or
malware detection and classification.
Section 3.1 presents the results of the study and the established taxonomy, whereas Sec-
tion 3.2 reviews identified research directions.
3.1. Literature review
The study performs two analyses of the literature. The quantitative analysis is based on ob-
jective metrics that can be extracted from the literature, such as publication date and venue.
Figures 1a and 1b show the evolution of the literature on FL and IDS over time, both being
building blocks of FIDSs.
3000 800
600
Publications
Publications
2000
400
1000 200
0
2000 2005 2010 2015 2020 2010 2012 2014 2016 2018 2020 2022
Year Year
(a) Keywords: intrusion detection system (b) Keywords: federated learning
Figure 1: Evolution of related domains until 08/10/2021, data from Microsoft Academic [22]
—Figures from Lavaur et al. [8] © 2022 IEEE
The qualitative analysis is based on the comparison of existing approaches, using the pro-
posed taxonomy. They can be grouped into four categories: data, local operation, federation,
and aggregation. A fifth meta-category is dedicated to the implementation and evaluation of
the approach.
68 Proceedings of the 29th C&ESAR (2022)
L. Lavaur, B. Coste, M.-O. Pahl et al
The taxonomy (Figure 2) provides twelve characteristics on which all approaches can be
compared:
1. Data source and type: what data is collected and how; heavily depends on the use case.
2. Preprocessing: strategies for data curation, such as normalization and feature selection.
3. ML location: where the ML model is trained and executed.
4. Local algorithms: how the ML model is trained, and its impact on performance.
5. Defense capabilities: the ability of the approach to mitigate attacks.
6. Federation strategy: how the federation is organized; e.g. client selection, architecture.
7. Communication: how data (i.e. models) is exchanged between participants, including
protection mechanisms.
8. FL type: type of FL strategy, depending on the objectives and available data.
9. Aggregation strategy: how models are aggregated, especially with heterogeneous clients.
10. Model target: i.e. the balance between specialization and generalization.
11. Analyzed dataset: the dataset used to evaluate the approach; often Information Technol-
ogy (IT)-focused, and not always available.
12. Costs and metrics: how the approach is evaluated, depending on the use case and objec-
tives.
The structure provided by the taxonomy also allows comparing the selected works. Table 1
summaries the results of the comparison. It shows that most approaches focus on IT network
traffic, using a gateway to collect data and host both learning and detection processes. Most
also use Deep Learnings (DLs) to perform supervised learning and classify traffic. A majority
use unmodified FedAvg for the aggregation, which is the initial FL algorithm that was proposed
by Google [14].
3.2. Open issues and research directions
As the FL topic gets more mature, research tends to focus on secondary aspects, such as security
and privacy, or on the application of FL to other use cases. This subsection summaries the open
issues and research directions; more details are available in [8].
(i) Performance — Like any detection system, FIDSs are looking for an absolute perfor-
mance: a system with a perfect classification score, producing no false positives or nega-
tives. To this end, several avenues have been identified in the literature, such as the use
of Generative Adversarial Networkss (GANs) or the improvement of feature selection as
input to the model. Moreover, the link between the performance of the system and its
hyper- and meta-parameters is not yet established.
Proceedings of the 29th C&ESAR (2022) 69
Federated Learning as Enabler for Collaborative Security between not Full...
Non-IID
Distribution
IID
1. Data source and type
Usage
Source
Data Communication
Dimensionality reduction
2. Preprocessing Feature extraction
Feature selection
On server
Dedicated device
3. ML location
On gateway
On device Outlier detection
Unsupervised ML Dimensionality reduction
Local operation Clustering
4. Local algorithms
Regression
Supervised ML
Classification
On-device counter-measures
5. Defense capabilities
Network reconfiguration (e.g. SDN)
Architecture (e.g. P2P)
6. Federation strategy Reputation
Taxonomy of FIDS Federation Client selection
Overhead reduction
7. Communication
Encryption
Federated Transfer Learning
8. FL type Vertical Federated Learning
Horizontal Federated Learning
MPC-based aggregation
Aggregation 9. Aggregation strategy Weighted algorithms
Parameter aggregation (e.g. FedAvg)
Generic model
By performance
10. Model target Personalized model
By-class
By-device
Unpublished (as of now)
11. Analyzed dataset Published but custom
Public / Known
Time (e.g. convergence)
Experimentation Training
Resources (e.g. hardware)
Aggregation
12. Costs and metrics Federation
Communication
Latency
Execution
Performance (e.g. accuracy)
Figure 2: Proposed taxonomy for FIDS—Figure from Lavaur et al. [8] © IEEE
(ii) Adaptability and scalability — Distributed systems such as FL are often used to cope
with resource limitations, especially in terms of computation and bandwidth. However,
as pointed out by several selected works, FL faces limitations when dealing with numer-
ous clients. Therefore, further research is needed on FIDSs client selection: dynamic
fusion based on score or time, reputation, number of detected attacks, etc.
(iii) Knowledge transfer — Current solutions focus on federating training and detection for
devices and resources that belong in the same domain. Therefore, open issues include the
70 Proceedings of the 29th C&ESAR (2022)
L. Lavaur, B. Coste, M.-O. Pahl et al
Table 1
Comparative table of selected works—Data from Lavaur et al. [8] © 2022 IEEE
Sa
Fe
Fe
te Aut hys chn ing
de
In
llit on ic ol s
de
Cy ati rnet
ra
fo Int
e- om al ogi
ra F r L l
te
rm e
be on of
Pe U up ise
te
te ed ea FL
d
r P Te T
rso ns erv d
rre ou yst es
d er r
Tr V onta
Se Su arn
M at ni
an er l F
On
Ne
str s V em
na up ise
m
Ho
im ed ng
sfe tic L
liz erv d
tw
Us -bas
ial eh s
lin
i-s erv ng
ic M
riz
e d is
or
ag
el
ne icle
Le T
p
S
k
e- d
e
m ed
tw s
ar L
ba
h
od
a
or
ni
se
i
Ref Training location Data type Local Algorithm Federation Algorithm Dataset Strengths
els
e
ng
ks
d
Abstracted network BIRCH relatively lightweight,
2018 Pahl et al. [23] # # # # # # # # # # # Device Parameter addition Generated
traffic (middleware) K-means online, no labels
Edge-controller offers mitigation,
2019 Rathore et al. [11] # # # # # # # # # # # # Network traffic (SDN) ANN Vector concatenation NSL-KDD
(SDN) decentralized
IoT network online, offers per-class
2019 Nguyen, Marchal, et al. [6] # # # # # # # # # # # Gateway MLP Weight and biases average MIMIC
traffic (TCPdump) models, no labels
Encrypted network
2019 Zhao et al. [24] # # # # # # # # # # # # Gateway GRU FedAvg Generated versatile (multi-task)
traffic (CICFlowMeter)
CICIDS2017
2019 Schneble et al. [25] # # # # # # # # # # # # Gateway Healthcare sensor values FC (shared layers) → FC Weight and biases average ISCXVPN2016 high adaptability, no labels
ISCXTor2016
2019 Cetin et al. [26] # # # # # # # # # # # # # Gateway Network traffic (WIFI) SAE FedAvg AWID –
Air conditioner
2020 Zhang et al. [18] # # # # # # # # # # # # # Gateway CNN-GRU → MLP Homomorphic parameter addition CPS dataset offers traceability (blockchain)
sensor values
2020 Li, Wu, et al. [15] # # # # # # # # # # # # # Gateway MODBUS traffic DAGMM Parameter addition KDD 99 confidentiality (encryption)
IoT network
2020 Rahman et al. [27] # # # # # # # # # # # # # Device ANN CDW_FedAvg Generated –
traffic (TCPdump)
CICIDS2017
IoT network
2020 Chen, Zhang, et al. [28] # # # # # # # # # # # # # Gateway CNN Parameter aggregation NSL-KDD no labels
traffic (TCPdump)
Generated
Network traffic segmented (performance-
2020 Sun, Ochiai, et al. [29] # # # # # # # # # # # # Gateway ANN FedAvg NSL-KDD
(PCAP) based models)
Gateway IoT network traffic LAN-Security knowledge transfer between
2020 Fan et al. [30] # # # # # # # # # # # # CNN Parameter aggregation
(MEC) (TCPdump, CICFlowMeter) Monitoring Project public and private datasets
Network traffic enhanced privacy
2020 Al-Marri et al. [31] # # # # # # # # # # # # # Gateway ANN FedAvg NSL-KDD
(TCPdump) (mimic learning)
Network traffic
2020 Kim, Cai, et al. [32] # # # # # # # # # # # # # Gateway MLP FedAvg NSL-KDD –
(TCPdump)
very lightweight,
Gateway CICIDS2017
2020 Qin, Poularakis, et al. [33] # # # # # # # # # # # # Network traffic (SDN) BNN SignSGD line-speed classification,
(SDN) ISCX Botnet 2014
P4 language compatible
CICIDS2017
Network traffic robust to poisoning,
2020 Chen, Lv, et al. [34] # # # # # # # # # # # # # Gateway GRU-SVM FedAGRU KDD 99
(CICFlowMeter) scalable
WSN-DS
Network traffic online, offers traceability
2020 Hei et al. [35] # # # # # # # # # # # # Device MLP FedAvg DARPA 1999
(TCPdump) (blockchain)
Network traffic (PCAP, relatively lightweight,
2020 Li, Zhou, et al. [36] # # # # # # # # # # # # Gateway CNN Homomorphic parameter addition Generated
CICFlowMeter, Argus) confidentiality (encryption)
IoT network
2021 Popoola et al. [37] # # # # # # # # # # # # # Gateway MLP Parameter aggregation KDD 99 zero-days detection
traffic (TCPdump, Argus)
Network traffic Bot-IoT
2021 Qin and Kondo [38] # # # # # # # # # # # # Device ANN FedAvg relatively lightweight
(TCPdump) N-BaIoT
Network traffic
2021 Liu et al. [39] # # # # # # # # # # # # # Device ELM + AE FedAvg NSL-KDD decentralized
(TCPdump)
Network traffic LAN-Security segmented (performance-
2021 Sun, Esaki, et al. [40] # # # # # # # # # # # # Gateway CNN Parameter aggregation
(PCAP) Monitoring Project based models)
Use case FL type Training Approach
ability to federate clients across different domains. In addition, current methods often
consider that all local models share the same architecture and hyper-parameters. This
limitation makes current FIDSs less versatile and transferable.
(iv) Security and trust — Using ML or FL to detect intrusions can introduce new threats
to the system, such as poisoning. Several works have examined the vulnerabilities of
FL systems and proposed countermeasures. With FL, poisoning becomes easier, as any
participant can theoretically impact everyone else’s model. Structure of models depends
on the architecture of the underlying algorithm, models trained cannot be aggregated
easily [41]. Furthermore, as ML, and especially DL, lacks explainability, the content of a
model is difficult to infer. Its aggregation with others is therefore made more risky. Fu-
ture works are required in this direction to properly assess the content of a model before
aggregation with others. Current solutions require an increase in the trust attributable
to clients for model aggregation, inspired by the state of the art in collaboration systems
and information sharing platforms.
(v) Self-defense and self-healing — Current research on FIDSs focuses on intrusion detec-
tion and attack classification. Defense is barely represented in the literature. However,
Proceedings of the 29th C&ESAR (2022) 71
Federated Learning as Enabler for Collaborative Security between not Full...
technologies such as Software-defined networking (SDN) offer rapid resiliency capabili-
ties; and recent work studies the effectiveness of such defense mechanisms. New emerg-
ing applications such as self-defense and self-healing systems could benefit from FIDSs
and other FL-based technologies.
(vi) Model convergence — Models can differ from one client to another, especially in het-
erogeneous contexts like intrusion detection. Consequently, model convergence is made
more difficult. Current research focus on optimizing parameters by considering aggre-
gation as an optimization problem. For instance, Charles et al. [42] use meta-learning to
infer the right parameters, thus optimizing the aggregation afterward. Weighting mech-
anisms are also present in the literature to improve the convergence [43].
(vii) Dataset representativity — Existing public datasets are not representative of FIDSs
potential deployment environments. Indeed, they are often datasets produced for tradi-
tional machine learning algorithms but split for federated purposes, like NSL-KDD [44],
UNSW-NB15 [45], or CIC-IDS2017 [46]. However, this approach introduces biases, as
features or times series are all related to the same original event. Similar approaches us-
ing adversarial examples for malware analysis which modify features instead of original
binaries faces the inverse feature-mapping problem [47], [48].
Some of these issues depend on works from other related fields, such as ML for performance
or FL for scalability. However, specificities of the FIDSs use case require more concrete research
questions. Especially, the topics of security, trust, and resilience, are critical for a collaborative
security use case.
4. Ensuring trust and personalization in FL-based intrusion
detection
As introduced in Section 2, FL can be used in multiple cybersecurity settings and use cases,
from distributed statistic inference [49] to authentication [50]. Even in the more specific con-
text of intrusion detection (i.e. FIDS), numerous use cases are studied, e.g. IT networks [29],
Industrial Internet of Things (IIoT) [15], or smart healthcare [51]. This variety in use cases
comes with various data types, architectures, and constraints. Making knowledge transferable
among heterogeneous use cases and data is a long-term goal for FIDSs (see Section 3.2).
However, heterogeneity is currently a major challenge for FL [52]. Therefore, we focus on a
specific use case, namely FIDS in IT networks. We study the impact of heterogeneity on FL in
this setting, and research solutions to mitigate it.
4.1. Use case definition
We consider a typical use case inspired from the industry, where actors invested in collabora-
tion are organizations aiming at improving their local detection. We assume that each organi-
zation has interests in sharing information, but has highly sensitive data that cannot be shared.
For example, Security Operation Centers (SOCs) perform security monitoring through the pro-
cessing of customer data (which can contain personal identifiable information) that cannot be
72 Proceedings of the 29th C&ESAR (2022)
L. Lavaur, B. Coste, M.-O. Pahl et al
shared. On the opposite, with the rise of cyber-criminal services [53], attackers tend to lead
similar attacks against different information systems. Two SOCs in such situation would share
their Indicators of Compromise (IoCs) or corresponding ML-models to detect all attacks after
the first that succeed. In this context, FL can be used to train a FIDS model on a distributed
dataset, while preserving the privacy of each organization. For instance, existing structures
such as Information Sharing and Analysis Centers (ISACs) or inter-SOCs could benefit from
such a system, which enables collaboration while protecting company secrets.
This setting is called cross-silo [52], as opposed to cross-device. It is worth noting that FIDSs
do not exclude cross-device settings, for instance in endpoint detection [5]. However, cross-
silo is more relevant for our use case, as it is more relatable for the industry, and it is easier to
implement in a testing infrastructure. In cross-silo, fewer clients (10–1000) operate with more
data each, as well as more powerful computation capabilities. Participants are also deemed
more reliable in terms of availability, as the local learning process is performed on a dedicated
device. Hence, it is not dependent on whether the hosting device (e.g. an employee computer) is
turned on or off. Finally, organizations often operate with higher-stake privacy requirements,
as they process data from their customers, their employees, and themselves.
We firstly also focus on horizontal federated learning (HFL), where participants have the
same features, but different samples. In HFL, participants have similar objectives (i.e. ML tasks)
and want to improve their models, but cannot build a centralized dataset due to privacy or legal
concerns. This is the most common setting in FIDSs, as it serves the goal of improving behavior
characterization, and having access to knowledge that cannot be inferred with only local data.
We also start with unsupervised learning algorithms, as the presence of labels not guaranteed
in real-world settings.
Figure 3 depicts an example topology, inspired by IT networks from the industry. We assume
that each organization has a dedicated IT network, that might vary in terms of architecture,
probe location, or services (see Section 4.2).
4.2. Experiment presentation
The chosen use case inherently highlights two of the seven open issues identified in the litera-
ture: heterogeneity and trust. Furthermore, as mentioned in Section 3.2, the lack of a dataset
that is representative of this use case undermines existing works on the topic. Therefore, we
first focus on two complementary tasks: (1) the creation of a representative, distributed dataset;
and (2) the development of solutions to mitigate heterogeneity and provide trust between dis-
tributed participants.
To cope with the lack of appropriated dataset in the literature, we propose to build a new
dataset generation platform with federation and distributed systems in mind. The platform re-
lies on virtualization topologies to generate normal traffic [54], and defined attack scenarios to
evaluate the detection performance. These topologies will be notably used to generate benign
traffic and train local unsupervised learning algorithms—literature typically use autoencoders
for this task [28], [38], [55]. The traffic generated at this step must be representative of a real
IT-focused network with users, internal resources (such as file sharing and web applications),
administration and supervision services, and access to the Internet.
We then aim to evaluate existing approaches on this more realistic and demanding use case.
Proceedings of the 29th C&ESAR (2022) 73
Federated Learning as Enabler for Collaborative Security between not Full...
We expect most of them to yield less promising results, even when they claim to be able to
deal with heterogeneity. On the other hand, FL literature abounds of works on heterogeneous
data, and aggregation algorithms, such as FedProx [56] or Fed+ [57], might provide good re-
sults. In this case, proving empirically that such approaches are able to cope with the realistic
heterogeneity of our use case remains a major contribution.
At the same time, we develop a new FL approach that deal with heterogeneity and trust. Sec-
tion 5 presents strategies envisioned to achieve this goal. When considered as a collaboration
system, FL is highly connected to research on trust and reputation systems. Hence, we plan
to leverage existing works on the topic, such as [58], to provide a reputation-aware FL system.
This approach will be evaluated on the new dataset, and compared to existing strategies.
4.2.1. Parameters
To measure the ways in which heterogeneity can manifest itself, we define varying parameters
that are used to generate the topologies. This can be used to generate a worst-case scenario,
where all the parameters are set differently in each topology.
In regard to the use case presented in Section 4.1, we consider the following parameters:
(i) architecture — the network architecture of the topology defines how services are inter-
connected, how the traffic is captured, and where data collection is performed. For exam-
ple, a topology with a single main gateway which captures traffic, and several services
on the same network, will produce a different dataset when compared with a star-shaped
topology with multiple subnets. Appropriated metrics are required to characterize the
impact of these differences, e.g. size (number of hosts, of subnets), mean number of hops
between a service and the last gateway, and so on.
(ii) services — different services can rely on different protocols, and therefore generate dif-
ferent kind of data, with different behaviors. For example, a service using TCP will
induce connection establishment, and therefore a lot of traffic back-and-forth, whereas
something based on UDP will produce a more continuous stream of data. Therefore,
different services (and protocols) might have different normal behaviors, causing hetero-
geneity among participants. The list of considered services must be adapted depending
on the considered attack scenarios.
(iii) maturity — security practices vary between organizations, depending on their threat
model, previous expertise, and budget. For example, a company might have a dedicated
security team, and therefore be able to implement a more mature security policy, whereas
a small company might not have the resources to do so. This parameter is important to
consider, as it can impact the quality of the dataset, e.g. by having unseen attacks in the
training data, supposed to be benign traffic.
(iv) probe location — while we assume that all participants extract the same features (due
to HFL settings, see Section 4.1), the location of the probes can vary. In the same archi-
tecture, one collection point at the gateway, or distributed probes in each subnet, will
produce different datasets.
74 Proceedings of the 29th C&ESAR (2022)
L. Lavaur, B. Coste, M.-O. Pahl et al
LAN Users LAN Admin
WAN / Internet employee-pc1 employee-pc3
.101 .102
centreon-server admin-pc it-pc
insider-pc .101 .102
.10
.105
employee-pc2 employee-pc4
.103 .104
user-pc remote-pc outsider-pc
.2.3 .5.101 .13.37 192.168.1.0/24 192.168.100.0/24
.1 .1 .1
10.0.0.0/8 .2
192.168.201.0/24 192.168.202.0/24
.1 .2 .1
wan-fw lan-fw monitor-fw
.1 .1
.1
LAN Local services LAN Monitor
DMZ
ids syslog-server splunk-server
ad-server odoo-server
shopping-server ftp-server
KEY DATA
.10 .11
0
.10 .12 .12
1
2
3
4
.10 .12
exchange-server dns-server
tomcat-server db-server .11 .13
.11 .13
192.168.10.0/24 192.168.99.0/24
192.168.11.0/24
Figure 3: Example topology of a IT network
Furthermore, some services are required to build working topologies, such as NTP, DHCP, or
DNS. We use a typical SME topology as a starting point for the experiments, which is presented
in Figure 3. This topology is composed of a dozen of machines, divided between employee
workstations, servers, etc. Variations of this topology will be also generated, and will be used
as other participants.
4.2.2. Evaluation
To evaluate approaches on any dateset, we measure their ability to detect relevant attack scenar-
ios. Typical attacks targeting such infrastructures are automated, like ransomware and botnets.
A lone attacker doing recognition and enumeration in the network is also to be considered, but
more advanced attacks such as APTs are out of the scope of this use case.
To make a rigorous evaluation, we rely on the state of the art to select attacks, such as the
MITRE ATT&CK framework [59], or the most used attacks in dataset literature. The list of
attacks that are considered on a given topology is deeply correlated to the services available
in this topology. For instance, cache poisoning requires cache-based service in the topology,
such as a DNS relay-server. On the other hand, organizations establish threat models depend-
ing on the services they host or use. Considered attacks may include: Distributed Denial of
Service (DDoS) on the internet-facing web services, host and port scan,or web vulnerability
exploitation (injections, ...).
Experiments are conducted on a private infrastructure. The test bed consist of three servers,
two for the virtualization, and one for computation that will be dedicated to executing the
required ML algorithms. While in real-world settings, the computation is performed on the
participants’ devices, we offload the computation to a dedicated server. This allows us to fo-
cus on the aggregation and trust aspects of the problem, and to avoid the complexity of the
distributed computation. To enable sound experiments [60], we plan to provide access to all
produced artifacts, including the dataset, the code, and the topology specifications, as well as
to the testbed itself.
Proceedings of the 29th C&ESAR (2022) 75
Federated Learning as Enabler for Collaborative Security between not Full...
5. Discussion
As the literature shows, FIDSs falter at aggregating heterogeneous models. Therefore, the
experiments detailed in section 4 provide means to measure and quantify the issue. While
literature on FL contains works on dealing with heterogeneous clients, this is still an issue in
the context of intrusion detection. Hence, we expect state-of-the-art approaches to perform
rather poorly, when compared to the results obtained on a homogeneous dataset. We consider
several approaches to cope with these issues.
First, metric-based model weighting could be used to give less importance to models that
deviate too much from the others. Zhang et al. [18] use a centroid-distance weighting algorithm
that cope with the heterogeneity in IIoT. More generally, weighting algorithms can help to
merge only relevant models, depending on a set of defined metrics. Other metrics could be used
to further tune the model aggregation, like a numerical estimation of how much information
the model can bring.
Another related topic is the measurement of the training data’s quality. In fact, as the learned
model is an abstraction of the data upon which it was trained, low-quality data would lead to a
low-quality model. Bringing a low-quality model in a federation could undermine every one’s
security. However, we must define what makes training data of quality, and how this quality
can be measured. Furthermore, weighting models according to their quality means the system
needs to be able to compare them, thus having access to the other’s data.
Moreover, existing works on federated learning rely on participant clustering [58], [61] to
improve model specialization. Often, these approaches aim at either reducing heterogeneity
between clients, or detecting and excluding malicious participants. Part of the challenge here
reside in the metrics to chose to build clusters. Aforementioned metrics such as data quality, or
information estimation, could be also used in clustering. In the context of intrusion detection,
the formation of clusters can have an indirect impact on participants’ security.
Finally, the lack of information inherent to the abstract nature of the model is a major hurdle
to estimate its value for aggregation. Therefore, we also consider adding metadata around
models, describing what the model contains without giving out too much information about
local data and configurations Such metadata would allow choosing which models one client
is interested in, depending on its use case. We believe an à-la-carte model aggregation would
help to cope with heterogeneity issues.
6. Conclusion
In this paper, we focused on the application of federated learning to the cybersecurity field.
Federated learning increases trust among partners as they do not need to share data, contrarily
to traditional ML-approaches. However, federated learning faces some limitations when feder-
ating models built on heterogeneous data. We therefore offer to address this issue within exper-
iments conducted on a platform with appropriate generated datasets. This platform displays
several use-cases with different partners in order to study aggregation-models parameters. We
then outline envisioned solutions which will be subjects of future work.
These contributions can have a significant impact on increasing the security of organizations.
76 Proceedings of the 29th C&ESAR (2022)
L. Lavaur, B. Coste, M.-O. Pahl et al
In fact, while existing research has addressed the privacy aspects of collaboration through fed-
erated learning and other privacy-preserving mechanisms, maintaining trust in heterogeneous
collaboration is still a challenge. We showed that federated learning can help with the creation
of a common trusted security model for systems that are inherently distributed. Our approach
enables non-trusting parties to collaborate for the joint goal of increasing their cybersecurity
without revealing critical internals.
References
[1] M. Catillo, A. D. Vecchio, A. Pecchia, and U. Villano, “A critique on the use of machine learning on
public datasets for intrusion detection,” in Communications in Computer and Information Science,
Springer International Publishing, 2021.
[2] S. Goldwasser, M. P. Kim, V. Vaikuntanathan, and O. Zamir, “Planting undetectable backdoors in
machine learning models,” 2022. arXiv: 2204.06974 [cs.LG].
[3] I. J. Goodfellow, J. Shlens, and C. Szegedy, “Explaining and harnessing adversarial examples,” 2014.
arXiv: 1412.6572 [stat.ML].
[4] N. Chaabouni, M. Mosbah, A. Zemmari, C. Sauvignac, and P. Faruki, “Network Intrusion Detec-
tion for IoT Security Based on Learning Techniques,” IEEE Communications Surveys & Tutorials,
2019.
[5] G. Karantzas and C. Patsakis, “An Empirical Assessment of Endpoint Detection and Response
Systems against Advanced Persistent Threats Attack Vectors,” en, Journal of Cybersecurity and
Privacy, 2021. (visited on 07/20/2021).
[6] T. D. Nguyen, S. Marchal, M. Miettinen, H. Fereidooni, N. Asokan, and A.-R. Sadeghi, “DÏoT:
A Federated Self-learning Anomaly Detection System for IoT,” in 2019 IEEE 39th International
Conference on Distributed Computing Systems (ICDCS), IEEE, 2019.
[7] P. García-Teodoro, J. Díaz-Verdejo, G. Maciá-Fernández, and E. Vázquez, “Anomaly-based net-
work intrusion detection: Techniques, systems and challenges,” Computers & Security, 2009.
[8] L. Lavaur, M.-O. Pahl, Y. Busnel, and F. Autrel, “The Evolution of Federated Learning-based Intru-
sion Detection and Mitigation: A Survey,” IEEE Transactions on Network and Service Management,
Special Issue on Network Security Management, 2022.
[9] E. M. Campos, P. F. Saura, A. González-Vidal, J. L. Hernández-Ramos, J. B. Bernabe, G. Baldini,
and A. Skarmeta, “Evaluating Federated Learning for Intrusion Detection in Internet of Things:
Review and Challenges,” en, arXiv:2108.00974 [cs], 2021. (visited on 12/02/2021).
[10] V. Chandola, A. Banerjee, and V. Kumar, “Anomaly detection: A survey,” en, ACM Computing
Surveys, 2009. (visited on 03/20/2022).
[11] S. Rathore, B. Wook Kwon, and J. H. Park, “BlockSecIoTNet: Blockchain-based decentralized se-
curity architecture for IoT network,” Journal of Network and Computer Applications, 2019.
[12] C. V. Zhou, C. Leckie, and S. Karunasekera, “A survey of coordinated attacks and collaborative
intrusion detection,” en, Computers & Security, 2010. (visited on 07/21/2021).
[13] ENISA, “Actionable Information for Security Incident Response,” Tech. Rep., 2014.
[14] B. McMahan, E. Moore, D. Ramage, S. Hampson, and B. A. y. Arcas, “Communication-efficient
learning of deep networks from decentralized data,” in Proceedings of the 20th international confer-
ence on artificial intelligence and statistics, A. Singh and J. Zhu, Eds., ser. Proceedings of machine
learning research, PMLR, 2017.
Proceedings of the 29th C&ESAR (2022) 77
Federated Learning as Enabler for Collaborative Security between not Full...
[15] B. Li, Y. Wu, J. Song, R. Lu, T. Li, and L. Zhao, “DeepFed: Federated Deep Learning for Intrusion
Detection in Industrial Cyber-Physical Systems,” IEEE Transactions on Industrial Informatics, 2020.
[16] J. Ren, H. Wang, T. Hou, S. Zheng, and C. Tang, “Federated Learning-Based Computation Offload-
ing Optimization in Edge Computing-Supported Internet of Things,” IEEE Access, 2019.
[17] K. Bonawitz, V. Ivanov, B. Kreuter, et al., “Practical Secure Aggregation for Privacy-Preserving
Machine Learning,” in Proceedings of the 2017 ACM SIGSAC Conference on Computer and Commu-
nications Security, New York, NY, USA: ACM, 2017.
[18] W. Zhang, Q. Lu, Q. Yu, et al., “Blockchain-based Federated Learning for Device Failure Detection
in Industrial IoT,” IEEE Internet of Things Journal, 2020.
[19] U. Majeed and C. S. Hong, “FLchain: Federated Learning via MEC-enabled Blockchain Network,”
in 2019 20th Asia-Pacific Network Operations and Management Symposium (APNOMS), IEEE, 2019.
[20] T. D. Nguyen, P. Rieger, H. Yalame, et al., “FLGUARD: Secure and Private Federated Learning,” en,
arXiv:2101.02281 [cs], 2021. (visited on 05/18/2021).
[21] M. Kim, O. Gunlu, and R. F. Schaefer, “Federated Learning with Local Differential Privacy: Trade-
offs between Privacy, Utility, and Communication,” 2021.
[22] A. Sinha, Z. Shen, Y. Song, H. Ma, D. Eide, B.-J. ( Hsu, and K. Wang, “An Overview of Microsoft
Academic Service (MAS) and Applications,” en, in Proceedings of the 24th International Conference
on World Wide Web, Florence Italy: ACM, 2015. (visited on 10/22/2021).
[23] M.-O. Pahl and F. X. Aubet, “All Eyes on You: Distributed Multi-Dimensional IoT Microservice
Anomaly Detection,” 14th International Conference on Network and Service Management, CNSM
2018, 1st Workshop on Segment Routing and Service Function Chaining, 2018.
[24] Y. Zhao, J. Chen, D. Wu, J. Teng, and S. Yu, “Multi-Task Network Anomaly Detection using Feder-
ated Learning,” en, in Proceedings of the Tenth International Symposium on Information and Com-
munication Technology - SoICT 2019, Hanoi, Ha Long Bay, Viet Nam: ACM Press, 2019. (visited
on 06/07/2021).
[25] W. Schneble and G. Thamilarasu, “Attack detection using federated learning in medical cyber-
physical systems,” en, 2019.
[26] B. Cetin, A. Lazar, J. Kim, A. Sim, and K. Wu, “Federated Wireless Network Intrusion Detection,”
en, in 2019 IEEE International Conference on Big Data (Big Data), Los Angeles, CA, USA: IEEE,
2019. (visited on 10/25/2021).
[27] S. A. Rahman, H. Tout, C. Talhi, and A. Mourad, “Internet of Things Intrusion Detection: Central-
ized, On-Device, or Federated Learning?” en, IEEE Network, 2020. (visited on 06/01/2021).
[28] Y. Chen, J. Zhang, and C. K. Yeo, “Network anomaly detection using federated deep autoencoding
gaussian mixture model,” in Machine learning for networking, S. Boumerdassi, É. Renault, and P.
Mühlethaler, Eds., Cham: Springer International Publishing, 2020.
[29] Y. Sun, H. Ochiai, and H. Esaki, “Intrusion Detection with Segmented Federated Learning for
Large-Scale Multiple LANs,” en, in 2020 International Joint Conference on Neural Networks (IJCNN),
Glasgow, United Kingdom: IEEE, 2020. (visited on 10/01/2021).
[30] Y. Fan, Y. Li, M. Zhan, H. Cui, and Y. Zhang, “IoTDefender: A Federated Transfer Learning Intru-
sion Detection Framework for 5G IoT,” en, in 2020 IEEE 14th International Conference on Big Data
Science and Engineering (BigDataSE), Guangzhou, China: IEEE, 2020. (visited on 10/04/2021).
[31] N. A. A.-A. Al-Marri, B. S. Ciftler, and M. Abdallah, “Federated Mimic Learning for Privacy Pre-
serving Intrusion Detection,” en, arXiv:2012.06974 [cs], 2020. (visited on 10/25/2021).
78 Proceedings of the 29th C&ESAR (2022)
L. Lavaur, B. Coste, M.-O. Pahl et al
[32] S. Kim, H. Cai, C. Hua, P. Gu, W. Xu, and J. Park, “Collaborative Anomaly Detection for Internet
of Things based on Federated Learning,” en, in 2020 IEEE/CIC International Conference on Com-
munications in China (ICCC), Chongqing, China: IEEE, 2020. (visited on 10/25/2021).
[33] Q. Qin, K. Poularakis, K. K. Leung, and L. Tassiulas, “Line-Speed and Scalable Intrusion Detection
at the Network Edge via Federated Learning,” en, 2020.
[34] Z. Chen, N. Lv, P. Liu, Y. Fang, K. Chen, and W. Pan, “Intrusion Detection for Wireless Edge
Networks Based on Federated Learning,” en, IEEE Access, 2020. (visited on 10/25/2021).
[35] X. Hei, X. Yin, Y. Wang, J. Ren, and L. Zhu, “A trusted feature aggregator federated learning for
distributed malicious attack detection,” en, Computers & Security, 2020. (visited on 10/25/2021).
[36] K. Li, H. Zhou, Z. Tu, W. Wang, and H. Zhang, “Distributed Network Intrusion Detection Sys-
tem in Satellite-Terrestrial Integrated Networks Using Federated Learning,” en, IEEE Access, 2020.
(visited on 10/25/2021).
[37] S. I. Popoola, R. Ande, B. Adebisi, G. Gui, M. Hammoudeh, and O. Jogunola, “Federated Deep
Learning for Zero-Day Botnet Attack Detection in IoT Edge Devices,” en, IEEE Internet of Things
Journal, 2021. (visited on 10/01/2021).
[38] Y. Qin and M. Kondo, “Federated Learning-Based Network Intrusion Detection with a Feature
Selection Approach,” en, in 2021 International Conference on Electrical, Communication, and Com-
puter Engineering (ICECCE), Kuala Lumpur, Malaysia: IEEE, 2021. (visited on 10/04/2021).
[39] H. Liu, S. Zhang, P. Zhang, X. Zhou, X. Shao, G. Pu, and Y. Zhang, “Blockchain and Federated
Learning for Collaborative Intrusion Detection in Vehicular Edge Computing,” en, IEEE Transac-
tions on Vehicular Technology, 2021. (visited on 10/04/2021).
[40] Y. Sun, H. Esaki, and H. Ochiai, “Adaptive Intrusion Detection in the Networking of Large-Scale
LANs With Segmented Federated Learning,” en, IEEE Open Journal of the Communications Society,
2021. (visited on 10/04/2021).
[41] X. Cao, Z. Li, H. Yu, and G. Sun, “CoFED: Cross-silo Heterogeneous Federated Multi-task Learning
via Co-training,” en, arXiv:2202.08603 [cs], 2022. (visited on 02/25/2022).
[42] Z. Charles and J. Konecny, “Convergence and Accuracy Trade-O s in Federated Learning and
Meta-Learning,” en, 2021.
[43] H. T. Nguyen, V. Sehwag, S. Hosseinalipour, C. G. Brinton, M. Chiang, and H. Vincent Poor, “Fast-
Convergent Federated Learning,” en, IEEE Journal on Selected Areas in Communications, 2021.
(visited on 05/25/2022).
[44] M. Tavallaee, E. Bagheri, W. Lu, and A. A. Ghorbani, “A detailed analysis of the KDD CUP 99 data
set,” in 2009 IEEE Symposium on Computational Intelligence for Security and Defense Applications,
IEEE, 2009.
[45] S. A. V. Jatti and V. J. Kishor Sontif, “UNSW-NB15: A Comprehensive Data set for Network Intru-
sion Detection Systems,” International Journal of Recent Technology and Engineering, 2019.
[46] I. Sharafaldin, A. Habibi Lashkari, and A. A. Ghorbani, “Toward Generating a New Intrusion De-
tection Dataset and Intrusion Traffic Characterization,” en, in Proceedings of the 4th International
Conference on Information Systems Security and Privacy, Funchal, Madeira, Portugal: SCITEPRESS
- Science and Technology Publications, 2018. (visited on 10/14/2021).
[47] F. Pierazzi, F. Pendlebury, J. Cortellazzi, and L. Cavallaro, “Intriguing properties of adversarial ml
attacks in the problem space,” 2019. arXiv: 1911.02142 [cs.CR].
[48] D. Park and B. Yener, “A survey on practical adversarial examples for malware classifiers,” in
Reversing and Offensive-oriented Trends Symposium, ACM, 2020. arXiv: 2011.05973 [cs.CR].
Proceedings of the 29th C&ESAR (2022) 79
Federated Learning as Enabler for Collaborative Security between not Full...
[49] J. R. Trocoso-Pastoriza, A. Mermoud, R. Bouyé, F. Marino, J.-P. Bossuat, V. Lenders, and J.-P.
Hubaux, Orchestrating Collaborative Cybersecurity: A Secure Framework for Distributed Privacy-
Preserving Threat Intelligence Sharing, en, 2022. (visited on 09/12/2022).
[50] M. Alazab, S. P. R M, P. M, P. Reddy, T. R. Gadekallu, and Q.-V. Pham, “Federated Learning for
Cybersecurity: Concepts, Challenges and Future Directions,” en, IEEE Transactions on Industrial
Informatics, 2021. (visited on 12/02/2021).
[51] E. Ashraf, N. F. F. Areed, H. Salem, E. H. Abdelhay, and A. Farouk, “FIDChain: Federated Intru-
sion Detection System for Blockchain-Enabled IoT Healthcare Applications,” en, Healthcare, 2022.
(visited on 07/05/2022).
[52] P. Kairouz, H. B. McMahan, B. Avent, et al., “Advances and Open Problems in Federated Learning,”
en, arXiv:1912.04977 [cs, stat], 2021. (visited on 04/01/2022).
[53] E. Brumaghin, A. Khodjibaev, M. Thaxton, and A. Zobec, Attackers leveraging dark utilities ”c2aas”
platform in malware campaigns, 2022.
[54] A. Dey, B. Costé, É. Totel, and A. Bécue, “Realistic simulation of users for it systems in cyber
ranges,” 2021.
[55] G. d. C. Bertoli, L. A. P. Junior, A. L. d. Santos, and O. Saotome, Generalizing intrusion detection
for heterogeneous networks: A stacked-unsupervised federated learning approach, en, 2022. (visited
on 09/12/2022).
[56] T. Li, A. K. Sahu, M. Zaheer, M. Sanjabi, A. Talwalkar, and V. Smith, “Federated Optimization in
Heterogeneous Networks,” en, arXiv:1812.06127 [cs, stat], 2020. (visited on 09/20/2021).
[57] P. Yu, A. Kundu, L. Wynter, and S. H. Lim, “Fed+: A Unified Approach to Robust Personalized
Federated Learning,” en, arXiv:2009.06303 [cs, math, stat], 2021. (visited on 01/31/2022).
[58] Z. Chen, P. Tian, W. Liao, and W. Yu, “Zero Knowledge Clustering Based Adversarial Mitigation in
Heterogeneous Federated Learning,” IEEE Transactions on Network Science and Engineering, 2021.
[59] B. E. Strom, A. Applebaum, D. P. Miller, K. C. Nickels, A. G. Pennington, and C. B. Thomas, “Mitre
att&ck™: Design and philosophy,” MITRE CORP BEDFORD MA, Tech. Rep., 2018.
[60] R. Uetz, C. Hemminghaus, L. Hackländer, P. Schlipper, and M. Henze, “Reproducible and Adapt-
able Log Data Generation for Sound Cybersecurity Experiments,” en, in Annual Computer Security
Applications Conference, Virtual Event USA: ACM, 2021. (visited on 08/08/2022).
[61] Z. Li, T. Wang, and N. Li, Differentially Private Vertical Federated Clustering, en, 2022. (visited on
08/16/2022).
80 Proceedings of the 29th C&ESAR (2022)