The MPLS WG Archive

[Shahram Davari <Shahram_Davari@pmc-sierra.com>]

>
>         I agree with Curtis; please keep the original text in 
>the draft.
>
>         --Tom

Tom,

Could you please clarify why you think LSP-ping is Simple and efficient? 

The reasons that I think it is not simple and efficient are:

1) Before sending Echo request, you need to run a complicated traceroute to determine the
hash keys (127/8 addresses) that covers all the ECMP points in the path. Also all intermediate nodes need to assign a series of 127/8 addresses to their ECMP path selections.

2) The fact that there are many TLVs defined, means processing the ping/traceroute messages
requires lots of processing power. 

3) It is not clear in the event of a failure how can the Pinged node process all the incoming echo or traceroute messages simultaneously. In other words it seems that it is not
scalable.

4) LSP ping is a bidirectional transaction and therefore requires 2x BW compared to
a unidirectional transaction. So it is not BW efficient either.


Yours,
-Shahram