Mitigation of LLDP Topological Poisoning Attack in SDN Environments
Using Mininet Emulator
Mattia Giovanni Spina
a
, Mauro Tropea
b
and Floriano De Rango
c
Department of Computer Engineering, Modeling, Electronics and Systems (DIMES), University of Calabria,
via P. Bucci 39c, 87036 Rende, Italy
Keywords:
LLDP Attack, SDN, Mininet, Elliptic Curve Cryptography, RSA, HMAC, ECDSA, POX.
Abstract:
Software-Defined Networking (SDN) paradigm permits to have scalability and flexibility in the network man-
agement throughout a centralized control that has the global view of the network topology, but it introduces
new security issues. In this paper, the Link Layer Discovery Protocol (LLDP) topological poisoning attack
has been studied and analysed in order to provide possible mitigation solutions through the use of Mininet
emulator and the POX controller. In particular, it is added to the LLDP protocol the integrity check using three
different types of cryptographic algorithms such as Hash-based message authentication code (HMAC), Digital
Signature Algorithm (DSA) using RSA and Elliptic Curve DSA (ECDSA). The performance evaluation of the
proposal is provided considering a network topology where an attacker hijacks/impersonates an host already
connected to the network.
1 INTRODUCTION
Software-Defined Networking (SDN) was introduced
to provide greater flexibility and agility to computer
networks thanks to the decoupling between control
plane and data plane. Before the introduction of
SDN, networks were managed by specialized network
devices (switches, routers, firewalls) that were pre-
configured using specific and proprietary vendor soft-
ware (protocols, firmware, operative systems) (Liu
et al., 2019). With SDN, network management is
centralized by means of a central entity, called con-
troller, which represents the brain of the whole net-
work and which leads all network devices dynami-
cally programming their behavior in order to adapt it
to specific needs of the network (Rahouti et al., 2022).
This new networking paradigm defines a three layers
architecture:
Application Plane (AP): it comprises services and
applications that rule over the controller and de-
fine the behavior to apply to the network;
Control Plane (CP): it represents the layer in
which resides the controller and in which all the
a
https://orcid.org/0009-0000-8407-2551
b
https://orcid.org/0000-0003-0593-5254
c
https://orcid.org/0000-0003-4901-6233
managing, monitoring and configuration opera-
tions related to data plane devices take place;
Data Plane (DP): it is the layer that comprises for-
warding devices.
Two interfaces are defined to establish the com-
munication between AP-CP and CP-DP that are
Northbound and Southbound interface, respectively.
For the first interface there is no standard communica-
tion protocol to be used and usually RESTful API are
employed. For the latter, one of the standard protocols
widely adopted is OpenFlow (Li et al., 2016). Despite
all the benefits that SDN paradigm can provide, each
of these layers are affected by security vulnerabilities
due to its software based nature.
All applications at the AP layer, since they need
to communicate with the controller, introduce possi-
ble flaws that require authentication mechanisms that
most of the time are neglected. Due to this prob-
lem, attackers can introduce malicious applications
in the network or exploit their vulnerabilities in or-
der to hijack the legacy forwarding rules installed by
the controller on the network devices. Security is-
sues concerning CP are inherently related to the cen-
tralized nature of the controller. If it is hijacked by
malicious users the whole network will be compro-
mised (Melkov and Paulikas, 2021; Fioravanti et al.,
2023). For what concern DP more complex attacks
can be considered such as malicious flow rules injec-
318
Spina, M., Tropea, M. and De Rango, F.
Mitigation of LLDP Topological Poisoning Attack in SDN Environments Using Mininet Emulator.
DOI: 10.5220/0012086200003546
In Proceedings of the 13th International Conference on Simulation and Modeling Methodologies, Technologies and Applications (SIMULTECH 2023), pages 318-325
ISBN: 978-989-758-668-2; ISSN: 2184-2841
Copyright
c
2023 by SCITEPRESS Science and Technology Publications, Lda. Under CC license (CC BY-NC-ND 4.0)
tion and/or tampering aiming at disrupting the actual
topology and the legacy behavior of the overall net-
work devices.
The OpenFlow is the commonly used and stan-
dardized protocol adopted for the southbound inter-
face. It can use a security layer, namely Transport
Layer Security (TLS), but its use it is not manda-
tory. Therefore, OpenFlow has a lot of vulnerabili-
ties. Among all of them crucial are the ones related
to the OpenFlow discovery procedure mainly due to
the absence of mechanisms of authentication between
controller and DP devices. So, this makes it easy to
perform different typologies of security attacks (Iqbal
et al., 2019). The main idea of this work consists
in the study of one of the typical attacks against an
SDN network related to the topology discovery mech-
anism, the Link Layer Discovery Protocol (LLDP)
fake link fabrication attack. We propose to add in-
tegrity checks to LLDP protocol in order to mitigate
the aforementioned attack against the LDS evaluat-
ing three possible types of security algorithms: Hash-
based message authentication code (HMAC), Digital
Signature Algorithm (DSA) using RSA and, lastly,
Elliptic Curve DSA (ECDSA) (De Rango et al., 2020;
Tropea et al., 2022).
The rest of the paper is organized as follows: Sec-
tion 2 presents a brief overview of the research state-
of-the-art about threats on topology discovery attacks;
Section 3 introduces the LLDP protocol and Section
4 shows the attack due to the crafted LLDP injection.
The proposed mitigations are described in Section
5, based on HMAC, DSA and ECDSA approaches.
Lastly, performance evaluation and conclusions are
presented in Section 6 and Section 7, respectively.
2 RELATED WORKS
The topic of security is very important for the scien-
tific community as proven by the huge literature in
different field of research starting from VPN (Gentile
et al., 2021) and involving various mitigation tech-
niques such as packet marking (Fazio et al., 2020).
Recently, many works have been proposed on security
mechanisms and adequate countermeasures for net-
work architectures based on the new paradigm known
as SDN. In this section, some of these papers are re-
viewed in order to show how the literature has dealt
with the security topic concerning the link discovery
service and in particular the LLDP protocol.
Authors in (Marin et al., 2019) found some vul-
nerabilities in LDS provided by Floodlight controller.
They proposed some guidelines to mitigate vulnera-
bilities in LDS. One of these vulnerabilities is, indeed,
Link Fabrication attack. To mitigate it they imple-
mented a MAC tag based mitigation. We followed
this principle and we considered modern and more re-
sistant techniques such as ECDSA to implement at the
same time a stronger and lightweight integrity check.
In (Chou et al., 2020) a topology anomaly detec-
tion module is proposed which is integrated in the
SDN controller which constantly searches for abnor-
mal LLDP packets that are flowing through the net-
work. The anomaly detection is based on a correlation
analysis on network traffic trying to avoid and miti-
gate topology discovery injection and flood attacks.
(Gao and Xu, 2022) proposed methods to detect
host migration that are caused by host hijacking at-
tacks and to thwart Link Fabrication attacks. Authors
implemented integrity check on LLDP by means of
a tag. However, they did not provide details on the
cryptographic primitive used to compute this tag and,
therefore, it is not possible to determine whether the
integrity check is strong enough in terms of security.
Authors in (Baidya and Hewett, 2020) proposed a
simple defence mechanism based on active port in or-
der to detect host-based and switch-based discovery
attacks. They based the defence mechanism on the
principle that no active port can be used to simultane-
ously connect more than one link otherwise they are
considered fake links. However, this does not prevent
an attacker to forge malicious LLDP packets poison-
ing the topology seen by the controller.
(Al Salti and Zhang, 2022) proposed LINK-
GUARD which is a framework used to make more se-
cure the link discovery process in SDN environments.
It is designed with the aim of detecting Link Fabri-
cation Attack in order to reduce the damage that are
caused by the poisoning of controller’s topology view.
The framework is not based on specific cryptographic
primitives and the only mitigation action is blocking
switch ports on which a LLDP attack is detected.
Authors in (Popic et al., 2020), in order to mitigate
LLDP vulnerabilities proposed a new mechanism for
LDS based on NETCONF (Network Configuration
Protocol) protocol used as a Southbound Interface to-
gether with a modelling language called YANG (Yet
Another New Generation). However, NETCONF is
not secured, so it represents a possible point for per-
forming attack to the network.
The work in (Alimohammadifar et al., 2018) pro-
posed a link fabrication attack detection based on a
probing techniques, in which probing packets are sent
across the network in order to find fake links. How-
ever, this defense mechanism comes at the cost of a
very high bandwidth consumption caused by the ad-
ditional probing packets.
Considering the reviewed literature, to the best
Mitigation of LLDP Topological Poisoning Attack in SDN Environments Using Mininet Emulator
319
of our knowledge, this is the first attempt to em-
ploy modern cryptographic techniques, such as el-
liptic curve, to enhance LLDP with integrity checks
without the need of deeply modify the protocol and
its behavior, keeping also low the computational and
protocol overhead imposed on it.
3 LINK LAYER DISCOVERY
PROTOCOL (LLDP)
One of the crucial procedures that allows an SDN
controller to retrieve information about the network
topology and keep it constantly updated is the Link
Discovery Service (LDS) (Hong et al., 2015). Ex-
ploiting the LDS, the controller can dynamically dis-
cover the data plane devices distribution, the exist-
ing links between them and other information that
are useful to achieve a consistent management of the
network (Jimenez et al., 2021). LDS implementa-
tions, which referring to OpenFlow is also denoted
as OpenFlow Discovery Protocol (OFDP), rely on the
Link Layer Discovery Protocol (LLDP) in order to re-
trieve information about switches and their links mak-
ing able the switches themselves to embed this infor-
mation inside specific control packets which encapsu-
late the LLDP frame. The LLDP protocol is formally
defined in the IEEE 802.1AB standard (Committee
et al., 2009) and it works at level 2 of the ISO/OSI
stack, and its frame format is described in Figure 1.
Figure 1: LLDP frame format.
In the SDN environment the link discovery proce-
dure is carried out using Open Flow control packets,
namely Packet-In and Packet-Out. Indeed, the LLDP
frame depicted in Figure 1 is encapsulated inside
those specific packets that are exchanged between
controller and switches (Nguyen and Yoo, 2017).
Topology information are exchanged by filling up the
LLDP Data Unit (LLD-PDU) when an LLDP packet
reaches a switch. LLD-PDU is composed of the TLV
- Type, Length, Value - which is used to denote the
needed information to retrieve the topology structure,
such as the Datapath ID (DPID), a 64-bit switch iden-
tifier, the Port ID used to specify the outgoing port of
the switch that will be used to forward the packet to
the next switch, and other useful information. To bet-
ter understand how the link discovery is carried out
we make an example depicted in Figure 2.
Let the considered topology be composed of a sin-
gle SDN controller and two switches s1 and s2. For
Figure 2: LLDP execution flow.
the sake of making easier to understand the example,
we only consider the discovery of the unidirectional
link from switch s1 to switch s2 since the reverse pro-
cedure is carried out in the same way. Initially, the
SDN controller sends a LLDP packet encapsulated in
a Packet-out message to switch s1 specifying in the
LLDP payload the DPID of the switch s1 and its out-
going Port ID. The controller will send this message
for each outgoing port of the switch s1. When s1 re-
ceives the packet-Out it decapsulates the LLDP frame
and sends it to switch s2. Switch s2, upon the receipt
of the LLDP message from s1, adds to the frame its
DPID and the Port ID that identifies the port on which
s2 has received the LLDP packet from s1. Then,
switch s2 encapsulates the frame inside a Packet-In
sending it to the controller that will get the informa-
tion about the link between switch s1 and switch s2,
i.e. the (s1,s1. p1),(s2,s2. p1) link. It is worth noting
that switches must be set with a pre-installed flow rule
imposing that any LLDP packet received on a switch
on a port which is not the port used to connect it with
the controller must be forwarded, as a Packet-in, to
the controller. This procedure will be repeated for
each switch that is managed by the controller and it
will be also performed periodically during time in or-
der to get always fresh topology information making
it able to get a holistic view of the whole network.
4 LINK FABRICATION ATTACKS:
FORGERY AND RELAY
Due to the lack of adequate security measures in the
LDS, an attacker can exploit this procedure to carry
out several types of attack that could bring serious
damage to the network: Fake Link Fabrication, De-
nial of Service, Man-In-The-Middle. We evaluate a
scenario in which an attacker, which is inside of the
SDN network or is able to hijack and control a host in-
SIMULTECH 2023 - 13th International Conference on Simulation and Modeling Methodologies, Technologies and Applications
320
side of it, injects fake topology information, exploit-
ing the lack of integrity checks, leading the controller
lose the right view of the network. The considered
attack is the Fake Link Fabrication. It can be catego-
rized in two types: Relay and Forgery (Gao and Xu,
2022). In the first case, a malicious user has to obtain
a genuine LLDP packet that is flowing inside the net-
work. This can be achieved by monitoring the Open-
Flow traffic and sniffing one of the exchanged LLDP
packet. Once the attacker obtains the genuine/legacy
LLDP packet, he modifies its information by replac-
ing the DPID of the legacy destination switch (L
s
)
with a DPID of a switch A
s
that is not actually con-
nected to L
s
. So, this tampered packet is forwarded
through the network waiting until the discovery pro-
cedure is executed and the controller save the fake link
between L
s
and A
s
in its links database. The latter type
of Link Fabrication is easier to achieve because an
attacker can simply forge from scratch a new LLDP
packet which will determine a fake link between two
arbitrary switches in the network.
5 PROPOSED MITIGATION
In order to mitigate the LLDP fake link fabrication
attack, the openflow.discovery module was modified
enhancing it with a set of algorithms to provide the
integrity feature to the LLDP protocol: HMAC, DSA-
RSA, ECDSA. In each of these methods the integrity
check is guaranteed used as data the DPID, Port ID
and TTL of the LLDP packet generated by the con-
troller. Before sending the LLDP packets to the
switches in the network the controller computes a tag
based on this data and attach it to the packet. When
the LLDP packet comes back from the network to the
controller filled with the data gathered by switches,
the controller will compute again the tag using the
same data and will check if the computed tag and the
tag into the packet are the same. If it is true the con-
troller will be sure that the packet is not been modified
during the transmission otherwise it drops the packet.
This procedure is depicted in Figure 3 in which there
is an attempt of LLDP packet tampering in order to
create a fake link.
5.1 Hash-Based Message
Authentication Code (HMAC)
LLDP injection attack mitigation can be achieved
through the use of HMAC, a hash-based authentica-
tion technique that allows to verify the integrity and
authenticity of the exchanged messages. In this work,
a HMAC with a SHA-256 hash function has been
Figure 3: LLDP attack example.
used. The following formula describes how HMAC
generates a tag S(k, m) starting from a message m and
a key k:
S(k, m) = H(k opad || H(k ipad || m)) (1)
The 128-bit key k, generated in a pseudo-random
way, is updated every time the controller starts a new
discovery cycle on the network. Message m is formed
by the concatenation of DPID, PortID and T T L.
The use of HMAC helps mitigate the threat of
LLDP injection, as forged LLDP messages will not
be able to pass HMAC-based authentication and will
be rejected by the network if the attacker tries to per-
form the attack shown first would have no effect, as
the integrity of the LLDP frame would not be veri-
fied. The complexity derived by the use of HMAC is
not very high as there is no key management between
the different hosts, but only the controller takes care
of it.
5.2 Digital Signature Algorithm (DSA)
A second way to mitigate LLDP injection can be
achieved through the use of a digital signature, an au-
thentication technique based on public-key cryptog-
raphy. The sender of an LLDP message uses their
private key to create a digital signature of the mes-
sage, which is then sent along with the message. The
receiver gets the message and verifies the digital sig-
nature using the sender’s public key. If the digital
signature verification is successful, the receiver can
be confident that the message comes from the legiti-
mate sender and has not been modified during trans-
mission. Using a digital signature therefore helps
mitigate the threat of LLDP packets tampering, as
forged or replayed LLDP messages will not be able
to generate a valid digital signature using the legiti-
mate sender’s private key, and will be rejected by the
network. Digital signature generation and verification
Mitigation of LLDP Topological Poisoning Attack in SDN Environments Using Mininet Emulator
321
require cryptographic computation which can lead to
increased processing load on network devices.
In our case, the digital signature based on the RSA
algorithm with a 2048-bit key was used.
5.3 Elliptic Curve Digital Signature
Algorithm (ECDSA)
LLDP injection mitigation can also be achieved
through the use of a digital signature based on Ellip-
tic Curves Cryptography (ECC), a public key cryp-
tography method that uses elliptic curves (Saho and
Ezin, 2020). In practice, the ECC allows the use
of shorter keys than other encryption techniques, but
with the same cryptographic strength. The use of a
digital signature based on ECC therefore helps mit-
igate the threat of LLDP injection, as forged LLDP
messages will not be able to generate a valid digi-
tal signature using the legitimate sender’s private key,
and will be rejected by the network. Furthermore, us-
ing ECC to generate digital signatures is also more se-
cure than other public key cryptography techniques,
as it is more resistant to some cryptographic attack
techniques such as brute force attack and key lookup
attack. Two types of curves have been used: ran-
dom curves (r) and Koblitz curves (k) (Brown, 2010).
Koblitz curves are characterized by their particular
structure, while random curves are generated ran-
domly. Koblitz curves type is characterized by some
inherently mathematical properties that allow to make
an efficient implementation of the group of opera-
tions that are used during the keys generation over the
curve. However, it is common to believe that more
randomly selected parameters (i.e. random curves)
can provide a more entropic force and, thus more se-
curity strength. But there is the suspect that random
coefficients can be chosen in order to provide a back-
door (Su
´
arez-Albela et al., 2018). The choice among
one curves type or the other is strictly related to the
specific needs of the application in which they are
used. If we need a computational effective imple-
mentation we can use Koblitz, otherwise if we want
to give a more entropic feature to the key generation
procedure we can choice random curves.
6 PERFORMANCE EVALUATION
In this section we provide the performance evaluation
of the proposed mitigation varying the integrity algo-
rithm. Our experimental setup is composed of a com-
puter running a virtual machine with Ubuntu 18.04
operative system running the latest version of Mininet
(Tropea and Palmieri, 2022) and the POX SDN con-
Figure 4: Considered network scenario.
troller. Different simulation scenarios have been con-
sidered during performance evaluation phase and they
are described in Table 1.
Table 1: Considered scenarios for performance evaluation.
Scenario Description
S1 Normal behavior
S2 HMAC mitigation
S3 REC-256 mitigation
S4 REC-384 mitigation
S5 KEC-283 mitigation
S6 KEC-409 mitigation
S7 DSA-RSA 2048 mitigation
S8 System under LLDP attack
6.1 Attack Scenario
To replicate the crafted fake link injection attack, we
considered the topology depicted in Figure 4.
In our scenario, the host h1 represents the ma-
licious user which is sniffing on S1 to get a legacy
LLDP packet with the DPID of S1. Once it gets this
packet, it will replace the DPID of S1 with the DPID
of S3. Then, forwarding this tampered packet it will
make the controller to believe that there exists a direct
link between S1 and S3, which actually does not ex-
ist in the legacy topology. This kind of attack can be
the starting point to launch more harmful attacks such
as Man-In-The-Middle or Denial of Service (DoS)
(Potrino et al., 2019). In the first one, the attacker by
letting the controller see a fake link between two arbi-
trary switches can sniff all the data exchanged on this
fake link that is accurately installed and controlled by
the attacker itself. In the case of the DoS, instead, the
attacker can create fake links among switches lead-
ing to complex-loop topology that can cause a huge
amount of extra data exchanged on the network lead-
ing to a high utilization of the bandwidth and, as a
consequence, to an extra-overhead that can saturate
the network itself. It has also to be mentioned, in the
SIMULTECH 2023 - 13th International Conference on Simulation and Modeling Methodologies, Technologies and Applications
322
context of DoS attack, the Spanning Tree Service that
is implemented by Open Flow controllers. It is a ser-
vice that is triggered each time that a topology update
occurs in order to block switch ports that are redun-
dant and that can lead to useless loop in the network.
If the attacker can leverage the LDS, it can create and
add a fake link between two legacy switches creating
a loop that will be detected by the Spanning Tree Ser-
vice that, as a consequence, will block the ports that
connect those two switches making them not able to
communicate anymore (Hong et al., 2015).
6.2 Network Performance Analysis
Using Wireshark, we have collected data relating to
the bandwidth used for each type of mitigation, con-
sidering a fixed simulation time. As can be seen in
Figure 5, the mitigation that uses the greatest band-
width is the digital signature, since the key used by the
algorithm is 2048 bits. For digital signatures based
on elliptic curves, bandwidth consumption is lower
because the keys used have a smaller length. In par-
ticular, on the digital signature based on ECC, it can
be noted that an increase in the length of the key leads
to an increase in bandwidth consumption. The miti-
gation with the use of HMAC turns out to be the least
expensive compared to all the others. However, ellip-
tic curve methods provide a higher degree of security,
so there is a trade-off between security and network
bandwidth consumption to be considered.
Regarding the generation and signing time of an
LLDP frame, a series of timing data was collected and
averaged. It can be seen from the Figure 6 that, due
to the generation of a message for verifying the in-
tegrity of the data, the time increases with respect to
the normal behavior where there is no verification of
data integrity. In particular, as we can imagine, it can
be observed that, using HMAC we got the less gen-
eration and verification time due to the use of high
efficient hash function SHA-256 and the effective im-
plementation of the HMAC algorithm. So we regis-
ter a slight increase in time while providing integrity
checks. When we deal with more complex methods
the increasing in time is strongly related to the length
of the key which is involved. Indeed, when elliptic
curve is used we registered an increasing of time in
generation and signature which is proportional to the
size of the key. The signature verification procedure
algorithm for elliptic curve is more computational ex-
pensive with respect to the one used in RSA and this
turns out to be verified by our experiments as can be
seen in Figure 6. The same consideration can be done
for the LLDP frame processing phase which can be
appreciated in Figure 7. Also in this case we have to
face a trade-off between security degree (higher when
using ECC and lower when using RSA) and process-
ing time (lower when using RSA).
6.3 CPU and RAM Performance
Analysis
CPU and RAM data usage was collected using a
Python script which is able to interact with the un-
derlying operative system. Following the aforemen-
tioned results about LLDP frame processing it can be
noticed in Figure 8 that this trend is confirmed for
the CPU usage. Indeed, when using the HMAC the
amount of consumed CPU is negligible with respect
to the normal behavior with the advantage of provid-
ing the needed integrity check. Instead, when using
the ECC the consumed CPU increases following the
trend of the LLDP frame processing time.
Since the signature generation and verification al-
gorithm does not change when changing the curve
type but it is related to the size of the employed key,
the amount of CPU does not show an evident trend
variation, see Figure 8. The high value registered for
the CPU usage when using RSA is due to the most
expensive computational overhead caused by the keys
generation procedure. These results highlight the ben-
efit of using elliptic curves since they provide a very
efficient and lightweight keys generation procedure,
which does not effect the overall overhead, and at the
same time a more security strength degree. Therefore,
in order to provide an integrity check for the LLDP
protocol, it can be evaluated the use of HMAC or
ECC considering a trade-off between the needed se-
curity degree and the computational overhead. Figure
9 compares the data collected on the use of memory
(RAM) by the controller. In this case there is no dif-
ference between the digital signature and the elliptic
curves, which use a slightly higher memory load than
the base case and HMAC. This is because a 128-bit
key is used in HMAC, while in other cases the key is
256 or 384 for digital signature based on random el-
liptic curves, 283 or 409 for digital signature based on
Koblitz elliptic curves, and 2048 for RSA-based digi-
tal signature. Using Wireshark it was possible to ana-
lyze the network in the already considered scenarios.
We have observed that the pattern of the exchanged
packets in the network is almost identical in every sit-
uation since the attack is based on sending a single
packet, which does not affect the network in any way.
For this reason, identifying such type of attacks it is
pretty complex since it can be considered as a very
silent attack. Thus, the additional integrity check pre-
vents the malicious user to carry out this kind of at-
tacks.
Mitigation of LLDP Topological Poisoning Attack in SDN Environments Using Mininet Emulator
323
Figure 5: Network used bandwidth.
Figure 6: LLDP frame generation time.
Figure 7: LLDP frame processing time.
7 CONCLUSIONS
This paper focuses on SDN network security, which
introduces new risks due to control/data plane separa-
tion. The fake link fabrication attack is analyzed, and
an integrity check is proposed using three cryptog-
raphy primitives (HMAC, DSA-RSA, ECDSA) and
Figure 8: Average Used CPU.
Figure 9: Average Used RAM memory.
Figure 10.
evaluated using Mininet and a POX controller. Re-
sults show performance in terms of network and re-
source consumption. Choosing between cryptogra-
phy primitives depends on network security needs and
operational restrictions. ECDSA is best for security,
while HMAC is a good alternative for a simpler and
more efficient approach.
SIMULTECH 2023 - 13th International Conference on Simulation and Modeling Methodologies, Technologies and Applications
324
REFERENCES
Al Salti, I. and Zhang, N. (2022). LINK-GUARD:
An Effective and Scalable Security Framework for
Link Discovery in SDN Networks. IEEE Access,
10:130233–130252.
Alimohammadifar, A., Majumdar, S., Madi, T., Jarraya, Y.,
Pourzandi, M., Wang, L., and Debbabi, M. (2018).
Stealthy Probing-Based Verification (SPV): An Ac-
tive Approach to Defending Software Defined Net-
works Against Topology Poisoning Attacks. [Online;
accessed 11. Mar. 2023].
Baidya, S. S. and Hewett, R. (2020). Link discovery at-
tacks in software-defined networks: Topology poison-
ing and impact analysis. J. Commun., 15(8):596–606.
Brown, D. R. (2010). Sec 2: Recommended elliptic curve
domain parameters. Standars for Efficient Cryptogra-
phy.
Chou, L.-D., Liu, C.-C., Lai, M.-S., Chiu, K.-C., Tu, H.-H.,
Su, S., Lai, C.-L., Yen, C.-K., and Tsai, W.-H. (2020).
Behavior anomaly detection in sdn control plane: a
case study of topology discovery attacks. Wireless
Communications and Mobile Computing, 2020:1–16.
Committee, L. S. et al. (2009). Ieee standard for local and
metropolitan area networks–station and media access
control connectivity discovery.
De Rango, F., Potrino, G., Tropea, M., and Fazio, P. (2020).
Energy-aware dynamic internet of things security sys-
tem based on elliptic curve cryptography and mes-
sage queue telemetry transport protocol for mitigat-
ing replay attacks. Pervasive and Mobile Computing,
61:101105.
Fazio, P., Tropea, M., Voznak, M., and De Rango, F. (2020).
On packet marking and markov modeling for ip trace-
back: A deep probabilistic and stochastic analysis.
Computer Networks, 182:107464.
Fioravanti, G., Spina, M. G., and De Rango, F. (2023). En-
tropy based ddos detection in software defined net-
works. In 2023 IEEE 20th Consumer Communica-
tions & Networking Conference (CCNC), pages 636–
639. IEEE.
Gao, Y. and Xu, M. (2022). Defense against software-
defined network topology poisoning attacks. Tsinghua
Science and Technology, 28(1):39–46.
Gentile, A. F., Fazio, P., and Miceli, G. (2021). A survey
on the implementation and management of secure vir-
tual private networks (vpns) and virtual lans (vlans)
in static and mobile scenarios. In Telecom, volume 2,
pages 430–445. MDPI.
Hong, S., Xu, L., Wang, H., and Gu, G. (2015). Poisoning
network visibility in software-defined networks: New
attacks and countermeasures. In Ndss, volume 15,
pages 8–11.
Iqbal, M., Iqbal, F., Mohsin, F., Rizwan, M., and Ahmad, F.
(2019). Security issues in software defined network-
ing (sdn): risks, challenges and potential solutions.
International Journal of Advanced Computer Science
and Applications, 10(10).
Jimenez, M. B., Fernandez, D., Rivadeneira, J. E., Bellido,
L., and Cardenas, A. (2021). A survey of the main
security issues and solutions for the sdn architecture.
IEEE Access, 9:122016–122038.
Li, W., Meng, W., and Kwok, L. F. (2016). A survey on
openflow-based software defined networks: Security
challenges and countermeasures. Journal of Network
and Computer Applications, 68:126–139.
Liu, Y., Zhao, B., Zhao, P., Fan, P., and Liu, H. (2019).
A survey: Typical security issues of software-defined
networking. China Communications, 16(7):13–31.
Marin, E., Bucciol, N., and Conti, M. (2019). An in-depth
look into sdn topology discovery mechanisms: Novel
attacks and practical countermeasures. In Proceedings
of the 2019 ACM SIGSAC Conference on Computer
and Communications Security, pages 1101–1114.
Melkov, D. and Paulikas, S. (2021). Security benefits and
drawbacks of software-defined networking. In 2021
IEEE Open Conference of Electrical, Electronic and
Information Sciences (eStream), pages 1–4. IEEE.
Nguyen, T.-H. and Yoo, M. (2017). Analysis of link dis-
covery service attacks in sdn controller. In 2017
International Conference on Information Networking
(ICOIN), pages 259–261. IEEE.
Popic, S., Vuleta, M., Cvjetkovic, P., and Todorovi
´
c, B. M.
(2020). Secure topology detection in software-defined
networking with network configuration protocol and
link layer discovery protocol. In 2020 International
Symposium on Industrial Electronics and Applica-
tions (INDEL), pages 1–5. IEEE.
Potrino, G., De Rango, F., and Santamaria, A. F. (2019).
Modeling and evaluation of a new iot security system
for mitigating dos attacks to the mqtt broker. In 2019
IEEE Wireless Communications and Networking Con-
ference (WCNC), pages 1–6. IEEE.
Rahouti, M., Xiong, K., Xin, Y., Jagatheesaperumal, S. K.,
Ayyash, M., and Shaheed, M. (2022). Sdn security re-
view: Threat taxonomy, implications, and open chal-
lenges. IEEE Access, 10:45820–45854.
Saho, N. J. G. and Ezin, E. C. (2020). Comparative study
on the performance of elliptic curve cryptography al-
gorithms with cryptography through rsa algorithm. In
CARI 2020-Colloque Africain sur la Recherche en In-
formatique et en Math
´
ematiques Apliqu
´
ees.
Su
´
arez-Albela, M., Fraga-Lamas, P., and Fern
´
andez-
Caram
´
es, T. M. (2018). A practical evaluation on
rsa and ecc-based cipher suites for iot high-security
energy-efficient fog and mist computing devices. Sen-
sors, 18(11):3868.
Tropea, M. and Palmieri, N. (2022). Software defined net-
working emulator for network application testing. In
Open Architecture/Open Business Model Net-Centric
Systems and Defense Transformation 2022, volume
12119, pages 50–61. SPIE.
Tropea, M., Spina, M. G., De Rango, F., and Gentile, A. F.
(2022). Security in wireless sensor networks: A cryp-
tography performance analysis at mac layer. Future
Internet, 14(5):145.
Mitigation of LLDP Topological Poisoning Attack in SDN Environments Using Mininet Emulator
325