Malicious Code and Access Control in Software-Defined-Networks

Christian Röpke

9. GI FG SIDAR Graduierten-Workshop über Reaktive Sicherheit (SPRING), 2014


Abstract

Software-Defined Networking (SDN) is a new paradigm for building networks and it may change the network market drastically. The main idea is to decouple network control features from packet forwarding hardware. This promises many advantages like more reliable and easier to manage networks or better cost-effectiveness compared to traditional networks. Because of that, the last few years people from both academia and industry has taken a great interest in this concept, and also big companies like Google benefit from SDN by increasing backbone performance and fault tolerance while reducing operation costs. In practice, so called SDN controllers manage programmable switches via open protocols like OpenFlow. Network control features are implemented as so called SDN applications which tell SDN controllers how to program the switches.

Beside its advantages, with SDN we also have to face new threats and attack vectors. One threat may arise through malicious SDN applications but little is known about it up to now. On the one hand, such programs could harm on network-level, e.g., by re-programming switches in order to deny certain network connections or reroute traffic to an adversary. On the other hand, they could attack SDN controllers on system-level, e.g., by executing arbitrary code or shutting down the SDN controller program.

[PDF]

Tags: Malicious Code, Software-Defined Networking