Game-based Defending against Attacks in Software
Defined Networks with Routing Control
Jie Duan, Ruilin Tian, Yuan Xing, Guofeng Zhao, Shuai Zeng, Yuanni Liu
School of Communication and Information Engineering
Chongqing University of Posts and Telecommunications, Chongqing, 400065, China
Email: {duanjie, zhaogf, zengshuai, liuyn}@cqupt.edu.cn, {trlwork, xingystudy}@foxmail.com
Abstract—This paper aims at defending the data traffic against
an attack in Software Defined Networks (SDNs) by leveraging the
flexibility to collect traffic information and control traffic routing
in SDN. We first model the interaction between an attacker and a
defender as a two-player variable sum game model, and analyze
the Nash equilibrium of the proposed game model. By further
analyzing the Nash equilibrium, we propose some guidelines to
control the traffic routing which can help enhance the traffic
security under attacking. Based on these guidelines, we design an
online algorithm and an offline algorithm to control the traffic
routing. By applying our algorithms to the scenario that the
attacker intends to destroy a node to break the traffic, our
algorithms can greatly reduce the traffic loss when the network
is under attacking.
I. INTRODUCTION
The last few years have witnessed significant increase of
attacks, such as TCP SYN flooding, SSPing, and DDoS attack,
etc. Accordingly, the intrusion detection systems (IDS) are
widely deployed in the networks in order to ensure the network
security. However, most of these IDSes are designed for a
special type of attack, such as deploying high redundant
network and choosing protection node to share the malicious
traffic [1], or designing a cryptograph system to protect data
integrity [2]. These works are not general enough to be
applicable in practice when the attacker is adaptive. Consider
that the main purpose of communication networks is to deliver
the traffic data, in this paper, we are to propose a framework
to control the demand routing and allocate the defending
resources, in order to defend the legitimate traffic against
malicious attacks.
Recent years, Software Defined Networking (SDN) imple-
mented as OpenFlow [3, 4] is proposed to provide a flexible
routing control and traffic information collection framework.
Thanks to SDN, we can deploy a centralized traffic routing
and defending resource allocation framework, such that we
can greatly defend the traffic against the malicious attack
and enhance the traffic security. Accordingly, we focus on
designing the defending framework in a centralized manner.
To propose such defending framework, game theory has
inherent advantages as it can formulate the interaction between
multiple players (e.g. the attacker and the defender in our
work). By leveraging the game theory, we can abstract the
interaction between the attacker and the defender as follows:
the attacker tries to attack some nodes in the networks to
disrupt the traffic (or eavesdrop the information, etc.), and the
defender deploys some schemes to defend its traffic against the
attack. Based on the analysis of this game model, the defender
can optimize its defending resource allocation scheme to
enhance the traffic security.
It should be noted that which node is worth attacking
(and also worth defending) is determined by the node value.
However, the node value is not always fixed and is associated
with the traffic (such as the traffic amount, importance) it
carries. Accordingly, we can control the traffic routing to
optimize the value of each node, and further enhance the traffic
security.
Many works leverage game theory to enhance the network
security. A game model was proposed to study how to allocate
defending resources in a network to counter malicious attacks
in [5]. Similar work was done and the interaction between
the defender and attacker was modeled as a 2-player game
model in [6]. However, The method in [6] assumed players
would attack/defend links to minimize/maximize the maximal
network flow, but only identified which nodes/links are the
most vulnerable ones. Mobility-enabled routing is the most
relevant to our work [7]. Though it applied the game model to
formulate the problem, and further analyzed the game model
solution to guide the traffic steering, it mainly focused on
how to control the topology in the MANET. In our work, the
topology is fixed and is not an ingredient that can be controlled
for optimization.
The technical contribution of this paper can be summarized
as follows:
∙ Formulate the interaction between the attacker and the
defender as a two-player variable sum game model.
∙ Solve the game model to indicate how to allocate the
defending resources in order to protect the traffic.
∙ Deeply analyze the Nash equilibrium and propose guide-
lines to control the routing to enhance the traffic security.
∙ Design routing and defending resource allocation scheme
to enhance the traffic security for both offline and online
cases.
∙ Extensive simulations to show the efficiency of our
framework.
The rest of the paper is organized as follows. Section
II formulates the interaction between the attacker and the
defender, and analyzes the game model. Algorithms to control
the demand routing are illustrated in section III. Section IV
978-1-5386-3180-5/18/$31.00 ©2018 IEEE