[mpls-tp] Problems with SPME as the path segment monitoring tool in MPLS-TP

[Alexander Vainshtein <[e-mail]>]

Dear Mr. Koike,
I've read your draft on path segment monitoring (http://www.ietf.org/id/draft-koike-mpls-tp-temporal-hitless-psm-00.txt) and I think that there is one more potentially fatal issue with the SPME mechanism if it is used for on-demand OAM. (My reading of the draft is that you see on-demand as the primary mode for path segment monitoring as it will be used to localize an observed fault.)

To the best of my understanding, setting up a temporal SPME will result in the LSRs within the monitoring segment only looking at the added (stacked) labels and not on the labels of the original LSP. (Actually, the label immediately below the SPME one will have to be set to one expected on the tail end node of the segment for the original LSP).

This means that problems stemming from incorrect (or unexpected) treatment of labels of the original LSP by the nodes within the monitored segment may disappear when you set up SPME. This includes HW problems during label look-up, mis-configuration etc. Of course, the inversion of this situation is also possible, .e.g., incorrect or unexpected treatment of SPME labels can result in false detection of a fault where none originally existed.

Does this make sense to you? Do you think that I've missed something important?

Regards,
     Sasha