Abstract

Techniques are presented herein that support a new mechanism that employs a software-defined networking (SDN) controller for detecting paths that are experiencing packet loss. Aspects of the mechanism employ path tracing (PT) probe packets that are generated by a source node towards a sink node and which follow the same path as customer traffic with a matching entropy label value thus providing the path tracing. A transit node that encounters a forwarding issue while processing a PT probe packet arising from a forwarding loop (e.g., detected through a time to live or a hop limit value of zero), a next-hop lookup failure in a forwarding information base (FIB), an interface maximum transmission unit (MTU) exceeded, a Layer 2 (L2) header issue, a reverse-path forwarding (RPF) check failure, an access control list (ACL) drop or a quality of service (QoS) tail drop, a denial-of-service (DoS) attack causing packet drop, etc. can immediately trigger to forward the received packet (including the original header and the collected path tracing data) with a new encapsulation to a controller. Such encapsulation may also include an additional type-length-value (TLV) containing troubleshooting information, such as a reason for a packet being dropped, to be able to quickly diagnose the packet loss on a transit node.

Creative Commons License

Creative Commons License
This work is licensed under a Creative Commons Attribution 4.0 License.

Share

COinS