InPatching is an in-network fast fail-over approach to INT-based monitoring mechanisms. It allows to recover from link failures in the data plane by performing detours directly in the data plane (with no control plane intervention). Follow the example:
Fig: InPatching schema
The figure above summarizes the InPatching procedure. We assume a link may failure due to several reasons (e.g., port misconfiguration, port down, cabble issues, etc.). To detect these failures, we designate a set of master data planes - in this case, switch A - where the main logic is applied. The master data plane is responsible for detecting the link failures by setting a timeout between each packet of the same flow (Step 1). Then (Steps 2-6), it designates the next switch upon which a detour must be performed in an attempt to discover and resolve the network's point of failure on the original path until it finds a path detour where the timeout threshold remains low (Step 6 - link G-H) - i.e., we maintain the performed detour which deviated the link failure.
PRs are very much appreciated. For bugs/features, consider creating an issue before sending a PR.
For further instructions, please read our wiki.