[Nea] full duplex NEA PB-TNC

["Sahita, Ravi" <ravi.sahita at intel.com>]

Hello,

As per the feedback received on the NEA list, here are some simple
suggestions that the PB-TNC editors would like feedback on for adding
full-duplex functionality into PB-TNC. These changes are proposed to
address some of the requests on this list for supporting such operation.
We didn't go for a totally full duplex approach as this adds substantial
additional complexity for the Posture Collectors and Posture Validators
with little value-add. These changes will allow PB-TNC to take advantage
of full-duplex transports but still ensure that it functions over
half-duplex transports as currently defined. 

1. Unsolicited Retry from NEA Client or Server in between a NEA session:
The proposed change here is that an S-RETRY or C-RETRY message can be
sent at any time during the NEA session to allow the server or client to
cause the session to be reset (and restarted). A new state will be
needed in the PB-TNC state machine that indicates a restart has been
received.

2. Matching requests to response messages: Since a full-duplex C/S-RETRY
message can cross on the wire with S/C-DATA in flight, we need to be
able to match client messages to server messages. The proposed change is
to add a Message Id in the PB messages, which can be used by the Server
to ignore in-flight client messages if it has issued a retry.

3. Updates to error reporting parameters: We would need the Message id
in the error parameters message so that the specific PB-TNC message in
which an error occurred can be indicated by the NEA server or client.

Please provide feedback on these proposed changes.

Thanks,
Ravi

_______________________________________________
Nea mailing list
Nea at ietf.org
https://www.ietf.org/mailman/listinfo/nea