[Roll] [ROLL] Projected DAO and Extended MOP

[rabi narayan sahoo <[e-mail]>]

Hello All

If my understanding is not wrong ID draft-ietf-roll-mopex-cap-00 is
introduced to support Projected DAO. I think Extended MOP is not required
for projected DAO.

Objective of P-DAO is to enable a path computing element (PCE) with global
topological view of the network, knowledge of device capabilities and
various network parameters to compute and install optimized (TE) routes to
meet certain application requirements (Application scenarios requiring P2P
path between devices).

Why Projected DAO is more of a Capability Compare to the Mode of Operation?

1-     The core mesh network will always operate either in Storing or
Non-Storing mode. Projected DAO will use the underlying topology to compute
more efficient p-routes and install them either in storing or non-storing
mode. A new LR, irrespective of it supports projected DAO or not it has to
join the network. This joining decision is completely based on which mode
global network is operating and the new LR supports it or not.


2-     P-DAO is more like a capability. This capability mostly includes if
the LR has enough memory to store the P-Routes and support to do SRH based
packet forwarding.

Let’s consider the below scenarios

1—Global Network is Storing


   - P-Route to be installed in Storing mode

For this to work all the LR need to have P-Route table


   -   P-Route to be installed is non-storing

Just the Ingress node must store the route entry in the P-Route table and
all other intermediate LRs in the P-Route process the SRH header and
forward the packet.

2—Global Network is Non-Storing

               Same requirement as above mentioned in storing mode case.
So P-DAO is independent of global mode of operation of the LLN.

What do you all think?

Thanks
Rabi