A cognitive model‐based approach for autonomic fault management in OpenFlow networks. (23rd August 2013)

• Record Type: Journal Article
• Authors: Kim, Sungsu; Kang, Joon‐Myung; Seo, Sin‐seok; Hong, James Won‐Ki
• Is Part Of: International journal of network management. Volume 23:Number 6(2013:Nov./Dec.)
• Page Start: 383
• Page End: 401
• Publication Date: 2013-08-23
• Subjects: Computer networks -- Management -- Periodicals; 004.6
• DOI: 10.1002/nem.1839
• Languages: English
• ISSNs: 1055-7148

Abstract

SUMMARY: Autonomic network management is an approach to the management of complex networks and services that incorporates the detection, diagnosis and reconfiguration, as well as optimization, of their performance. A control loop is fundamental as it facilitates the capture of the current state of the networks and the reconfiguration of network elements without human intervention. For new networking architectures such as software‐defined networking and OpenFlow networks, in which the control plane is moved onto a centralized controller, an efficient control loop and decision making are more crucial. In this paper, we propose a cognitive control loop based on a cognitive model for efficient problem resolving and accurate decision making. In contrast to existing control loops, the proposed control loop provides reactive, deliberative and reflective loops for managing systems based on analysis of current status. In order to validate the proposed control loop, we applied it to fault management in OpenFlow networks and found that the protection mechanism provides fast recovery from single failures in OpenFlow networks, but it cannot cover multiple‐failure cases. We therefore also propose a fast flow setup (FFS) algorithm for our control loop to manage multiple‐failure scenarios. The proposed control loop adaptively uses protection and FFS based on analysis of failure situations. We evaluate the proposed control loop and the FFS algorithm by conducting failure recovery experiments and comparing its recovery time to those of existing methods. Copyright © 2013 John Wiley & Sons, Ltd. Abstract : In this paper a cognitive control loop based on a cognitive model for efficient problem resolving and accurate decision making is proposed. The proposed control loop provides reactive, deliberative, and reflective loops for managing systems based on analysis of current status. We propose a fast flow setup (FFS) algorithm for our control loop to manage multiple‐failure cases. We evaluated the proposed control loop and the FFS algorithm by conducting failure recovery experiments in OpenFlow networks and showed that our proposed method outperforms those of existing methods.