Greg Mirsky
2018-09-24 21:53:27 UTC
Dear All,
a number of drafts that define encapsulation of iOAM data for the various
overlay networks had been presented and requested adoption by the IPPM WG.
I have one scenario to discuss:
- all proposed encapsulations of iOAM data use the field that explicitly
indicates the type of the payload that immediately follows the overlay
header. In all proposals iOAM data identified either as End-to-end iOAM or
Hop-by-hop iOAM;
- iOAM data uses its own shim that includes the Next Protocol field. The
value of the Next protocol field identifies the type of the payload, the
user payload;
- now consider the situation when a node in the overlay network that
doesn't support iOAM receives an iOAM data, i.e., doesn't recognize iOAM
payload types. The node will drop the iOAM data and the user packet with it.
If my analysis is correct, then here are a couple of questions:
- how useful is the technology that requires an upgrade of the whole
network in order to work;
- should we return the discussion of iOAM data encapsulation in overlay
networks to the WGs that define these overlay networks.
Regards,
Greg
a number of drafts that define encapsulation of iOAM data for the various
overlay networks had been presented and requested adoption by the IPPM WG.
I have one scenario to discuss:
- all proposed encapsulations of iOAM data use the field that explicitly
indicates the type of the payload that immediately follows the overlay
header. In all proposals iOAM data identified either as End-to-end iOAM or
Hop-by-hop iOAM;
- iOAM data uses its own shim that includes the Next Protocol field. The
value of the Next protocol field identifies the type of the payload, the
user payload;
- now consider the situation when a node in the overlay network that
doesn't support iOAM receives an iOAM data, i.e., doesn't recognize iOAM
payload types. The node will drop the iOAM data and the user packet with it.
If my analysis is correct, then here are a couple of questions:
- how useful is the technology that requires an upgrade of the whole
network in order to work;
- should we return the discussion of iOAM data encapsulation in overlay
networks to the WGs that define these overlay networks.
Regards,
Greg