Open Access
Subscription Access
Control-Plane based Failure Detection for Highly Available Software-Defined Network
Abstract
To ensure high availability, redundancies such as active/standby controllers have been adopted on production SDN networks. However, the complex functionalities embodied by a SDN network make it difficult to ascertain the presence of failures. Not all failures will exhibit as component crashes to be captured by heartbeat-based failure detectors. We propose a novel failure detector for SDN network based on transparent monitoring of control-plane messages and real-time modeling of network state. By detecting deviation of the model, the proposed failure detector is capable of detecting a wide range of failures. The detector does not depend on specific SDN applications or controllers. We implemented a prototype for OpenFlow-based [3] SDN network and provide the preliminary results confirming the feasibility of a control-plane based failure detector.
Keywords
software defined network; failure detector; high availability; control plane
Citation Format:
Patrick Ngai, Tung-Yueh Lin, Yu-Sung Wu, Chien-Hua Lee, "Control-Plane based Failure Detection for Highly Available Software-Defined Network," Communications of the CCISA, vol. 23, no. 1 , pp. 6-16, Jan. 2017.
Patrick Ngai, Tung-Yueh Lin, Yu-Sung Wu, Chien-Hua Lee, "Control-Plane based Failure Detection for Highly Available Software-Defined Network," Communications of the CCISA, vol. 23, no. 1 , pp. 6-16, Jan. 2017.
Full Text:
PDFRefbacks
- There are currently no refbacks.
Published by Chinese Cryptology and Information Security Association (CCISA), Taiwan, R.O.C
CCCISA Editorial Office
E-mail: ccisa.editor@gmail.com