Operation and Maintenance

One of the most important aspects of carrier-grade networks is reliability. This is provided partly by using reliable hardware and properly designing the network. However, sometimes that is not enough and a number of different protocols have been developed to handle those occasions. These protocols are used to troubleshoot the network e.g. by measuring delay and packet loss over certain paths or trace paths through the network. They also handle fault conditions such as broken links very quickly, e.g. by switching over to a pre-calculated path within a few milliseconds from the failure.

These protocols pose a challenge to the split architecture by raising the question of how much of the classical router/switch functionality can actually be moved to a central location. Many of these protocols requiring high accuracy timing, low jitter, high packet rates, making it impractical to handle them with the central controller. In the project we have focused on one of these protocols, namely Bidirectional Forwarding Detection for MPLS-TP which is designed to provide Connectivity Verification and Continuity Check in MPLS-TP networks (MPLS-TP BFD draft-ietf-mpls-tp-cc-cv-rdi-05). We have designed, implemented, and tested a system for configuring, running and using BFD on top of OpenFlow version 1.0 with MPLS extensions.