Enabling fast failure recovery in OpenFlow networks

摘要

OpenFlow is a novel technology designed at Stanford University which aims at decoupling the controller software from the forwarding hardware of a router or switch. The OpenFlow concept is based on the approach that the forwarding information base (FIB) of a switch can be programmed via a controller which resides at a separate hardware. The goal is to provide a standardized open management interface to the forwarding hardware of a router or switch. The aim of a project SPARC “SPlit ARchitecture Carrier grade networks” is to deploy OpenFlow in carrier grade networks. Reliability is a major issue to deploy OpenFlow in this networks. This work proposes the addition of a fast restoration mechanism in OpenFlow and evaluates the performance by comparing the switchover time and packet loss to existing restoration options in a current OpenFlow implementation.