Abstract:

Aiming at node failure, we propose a proactive and fast recovery mechanism, called node failure protection (NFP). The mechanism can ensure the resiliency of the network, and packets can be rerouted to alternative paths by the switches themselves without involving the controller. In the meantime, the alternative paths do not partially overlap the working paths in the main flow table. Experimental results indicate that the NFP is more reliable than reactive approaches as the scale gets larger, and it can fulfill the demand of restoration speed in carrier-grade networks.

Key words: SDN, OpenFlow, node failure, proactive protection