🤖 AI Summary
This study addresses the coordination challenges in in-band SDN control plane deployments with multiple controllers, where controller discovery, state synchronization, and failure recovery must be achieved without expanding switch forwarding state. The authors propose a boundary-switch-based local forwarding graph mechanism that confines inter-domain routing information to boundary devices, preventing state propagation into intermediate domains. In-band control communication is realized using Open vSwitch’s Nicira extensions with NSH encapsulation, and neighbor discovery is accomplished via Controller Advertisement messages. The approach requires no switch firmware modifications and incurs flow table overhead independent of the number of controllers, maintaining constant space complexity. Experiments in a Mininet environment with 96 switches and 5 controllers demonstrate that internal switches exhibit fixed flow table occupancy, enabling network scalability to hundreds of nodes with controller discovery convergence times on the order of seconds.
📝 Abstract
In-band SDN control planes, where control traffic shares the data-plane infrastructure, suit wide-area, resource-constrained deployments -- such as rural backbones -- that cannot afford a dedicated control network. Partitioning such a network across multiple controllers improves scalability but raises a coordination challenge that in-band designs have largely ignored: controllers must discover one another and exchange state in-band, and switches must recover when their controller fails, all without forwarding state that grows with the number of controllers.
This paper presents the multi-controller coordination plane of Periplus, an in-band control plane whose single-controller design is developed in a companion paper. Periplus controllers discover their neighbors through Controller Advertisement (C-Adv) messages and build inter-controller routes incrementally: each border switch inserts a partial forwarding graph covering only the next domain, so per-controller forwarding state is confined to border switches and never distributed across the interior of an intermediate domain. The same C-Adv mechanism reattaches a switch to a surviving controller after a controller failure.
We evaluate a Ryu-based implementation in Mininet, including a 96-switch, 5-controller scenario. Per-switch flow-table state is set by a switch's role rather than by network size -- interior occupancy stays constant as controllers are added -- partitioning scales bootstrap to networks of around a hundred switches, and inter-controller discovery converges within seconds. The design needs no switch-firmware modifications: it runs on stock Open vSwitch, using only its built-in Nicira extensions for Network Service Header (NSH) encapsulation.