Open niloda opened 1 year ago
A proposal to complete the functionality could be:
If we wanted to allow LOS forwarding in a chain of radio links (last figure of the previous comment), the device would have to know the status of the remote ports. Currently the TR-532 model does not provide this, so important changes would be required. Therefore, we could postpone the solution of this part of the issue to a possible step 2, of course if it were worth it.
Alternative proposal - "On top of EthernetContainer":
The following proposal requires that the operational state of a client layer interface is regarding the operational state of its serving layer interface(s).
If there is a 1:1 relationship, it is required that the operational state of a client LTP is down, whenever the operational state of its server LTP is down.
If there is a 1:n relationship (e.g. physical link aggregation), it is required that the operational state of the client LTP is down, whenever the operational states of all its server LTPs is down.
This modelling shall be supported on management interfaces of devices that are supporting LinkLossForwarding configuration. If some device would not support LinkLossForwarding the MacInterface would keep being client of the EthernetContainer LTP.
Many thanks again!!! Topic is on the agenda for tomorrow. Cheers,MartinGesendet von Outlook für Android
Decision during the 5G-xhaul call to the 13th of December 2023:
The following proposal would have to be implemented, if the device supports flexibly defining relationships between interfaces that are suffering from link loss and interfaces shall shall be switched off because of this event.
It requires that the mediator implementation can handle multiple LPs at an LTP.
The operational state information would be propagated in lateral direction between interfaces of different kinds (AirInterface and WireInterface).
LinkLossForwardingInterface
LinkLossForwardingFd
LinkLossForwardingFc
The proposal for next Dec6th should be postponed to Dec13 to gather more feedback.
In addition a discussion is beneficial about the creation and deletion of LinkLossForwardingFc . RPC?
Proposal above has been complemented accordingly.
With link loss forwarding enabled, the mac interface must be shut down when the radio link is down. This is easy to implement when the device has only one radio link. In the figure below, when the radio link goes down, the three interfaces should be shut down, stopping the three services (red, blue and green)
With multiple radio links, you should specify which one shuts down the specific mac interface, otherwise any alarmed link could shut down all mac interfaces on the device, even those not receiving traffic from the alarmed link. In the figure below, in a device with 3 radio interfaces, the two services (red and green) outgoing from LAN-3 flow over two different microwave links, while the service (blue) outgoing from LAN-1 flows over a third microwave link.
The question is which LAN interface should be shut down when which radio link is down. Of course, a problem on MW-LINK-3 should shut down LAN-1 only, but a problem on MW-LINK-1 and MW-LINK-2 (one or both?) should shut down LAN-3. LAN-2 should not be affected by any radio link even if link loss forwarding is enabled.
In addition, considering the complete link, not just one end-point, should the los alarm of the remote interface from which the service enters be forwarded? For example, in the figure below, should LAN-1 be shut down in response to a los alarm on the remote interface LAN-3? This allows to forward the los in a chain of radio links.
The problem is that the current model: