Resilience Analysis of Software-Defined Networks Using Queueing Networks

Abstract

Software-Defined Networks (SDN) are being adopted widely and are also likely to be deployed as the infrastructure of systems with critical real-time properties such as Industrial Control Systems (ICS). This raises the question of what security and performance guarantees can be given for the data plane of such critical systems and whether any control plane actions will adversely affect these guarantees, particularly for quality of service in real-time systems. In this paper we study the existing literature on the analysis of SDN using queueing networks and show ways in which models need to be extended to study attacks that are based on arrival rates and service time distributions of flows in SDN.

A Holistic Approach for Enhancing Critical Infrastructure Protection: Research Agenda

“…The main objective of the review is to serve as a foundation for determining the appropriate asset-centric threat modelling approaches that could be employed for a given scenario. Moreover, our research outcomes which can be categorized into the second part of the holistic approach are published in [18,22,23]. We remarked in [23] that software-defined networks (SDN) are widely being adopted and are likely to be deployed in critical infrastructure.…”

“…Moreover, our research outcomes which can be categorized into the second part of the holistic approach are published in [18,22,23]. We remarked in [23] that software-defined networks (SDN) are widely being adopted and are likely to be deployed in critical infrastructure. However, there is a need to evaluate the security and performance guarantees that can be given for the data plane of such critical systems.…”

Applying software-defined networking to support telemedicine health consultation during and post Covid-19 era

“…SDN is a current network paradigm in which the control plane is separated from the data plane [21]. In contrast to the existing network architecture where the control and data plane are part of the network devices, roles are separated in SDN such that the network devices can become purely forwarding devices with the forwarding instructions issued to them by the control plane and allowing the use of vendorneutral components [22]. In SDN, the main objectives are to simplify the deployment of control plane functions and to support the applications to deal with a single abstracted network environment without regards for the implementation details, in addition to lowering the dependency on proprietary components [10,22].…”

“…In contrast to the existing network architecture where the control and data plane are part of the network devices, roles are separated in SDN such that the network devices can become purely forwarding devices with the forwarding instructions issued to them by the control plane and allowing the use of vendorneutral components [22]. In SDN, the main objectives are to simplify the deployment of control plane functions and to support the applications to deal with a single abstracted network environment without regards for the implementation details, in addition to lowering the dependency on proprietary components [10,22]. Therefore, SDN enables network functions to be directly programmable, which facilitates the automation of network functions and the building of highly scalable and flexible networks that can quickly adapt to changing environments [14].…”

Modelling Adversarial Flow in Software-Defined Industrial Control Networks Using a Queueing Network Model

“…The arrival rate and service time distribution of the adversarial flow can be manipulated by the attacker to achieve desired ends. We had described the idea of adversarial flow in our previous work [2] and opined that modelling based on aggregating distribution or modelling based on analysing the effect of two queues separately may be deployed to study the effect of adversarial flow in the analysis of SDN using queueing network models.…”