BFD echo mode interoperability

[[e-mail]]

Colleagues,

I've been looking at the BFD base draft and am struggling to understand how interoperable BFD solutions can be developed using echo mode.

The draft states "The means by which the Echo function failures are detected is outside of the scope of this specification. Any means which will detect a communication failure is acceptable." The draft also states "A means of detecting missing Echo packets must be implemented, which most likely involves processing of the Echo packets that are received. The processing of received Echo packets is otherwise outside the scope of this specification."

This is in contrast to the asynchronous mode which uses the detect multiplier to define how many consecutive packets must be missed in order to declare a session down. I have seen an implementation that uses the detect multiplier for echo packets as well as control packets, but this behaviour is not specified in the draft.

If two vendors decide to use different methods/settings for fault detection using echo mode, this will be a real pain to configure/manage, and will also require translation between different methods in order to ensure symmetry in detecting failures in order to avoid black holing traffic.

Unless I've missed something, using echo mode in a multivendor environment does not look like an attractive solution.

Regards,
Richard