Local Fault Signal on 1GE/10GE/40GE/100GE client port (Link Fault Signaling)

Created: Jan 22, 2021 06:38:09Latest reply: Aug 12, 2021 12:28:59 2467 136 33 0 2

Hello, everyone!

I would like to share with you some Q&A about "Local Fault Signal on 1GE/10GE/40GE/100GE client port (Link Fault Signaling)" queried with the Huawei TAC and solved.

Our customer is asking to clarify a topic about ODUk SNCP switching in OptiX OSN 8800 nodes.

Consider a situation in which we haven't activated ALS functionality in customer ports (third-party equipments) and, during some milliseconds (<50ms) an ODUk AIS is generated because there has been a break in the link. The scenario is below:

Then, before switching to the protection link occurs, we have next questions:

How is AIS alarm propagated to the Ethernet side?

The AIS alarm will not send to Ethernet side, it will terminate on our device (client port).

Maybe Ethernet Frames are not sent meanwhile?

If the AIS is continuous, that means the WDM link was down, the live service was down, but we will send a Local Fault Signal to Ethernet Frame.

Does this event go through the cross-connection matrix?

The Local Fault Signal is inserted by Tributary Board (which is connected to customer third-party equipment), it will not go through cross-connection matrix.

Which is the signal flow to have this Local Fault Signal to Ethernet Frame?

The Local Fault Signal is a particular bit stream sequence that consist of 66 bits. The preamble is the first two bits 10 of the sequence (sync header), then followed by 8 bytes stream.

Does this feature come configured by default in the nodes or it must be configured?

The Local Fault Signal is not configurable, by default it is there. When e.g. AIS alarm arose on TSC tributary board, the 100GE client port will send out the Local Fault Signal automatically.

Which is the structure of this bit stream? Is the Local Fault Signal bit stream the same for all type of ETH cliente ports?

For 1GE port, there is no LF signal.

For 10GE port, the IEEE 802.3ae-2002 protocol is followed and the LF signal bit stream is as follows:

66B = 10 0x55 0x00 0x00 0x01 0x00 0x00 0x00 0x01

For 40GE & 100GE port, the IEEE 802.3ba-2010 protocol is followed and the LF signal bit stream is as follows:

66B = 10 0x4B 0x00 0x00 0x01 0x00 0x00 0x00 0x00

The IEEE 802.3ae-2002 and IEEE 802.3ba-2010 protocol whitepaper are attached on this post.

Finally, what about the details of the two encapsulation methods to the OTN world such as GMP and GFP? Is it necessary to cross the rules between IEEE802.3 and ITU-T G.709?

For the WDM device we have two sides, Client side and WDM side. For WDM side we have GMP and GFP two kinds of encapsulation which follow ITU-T Recommendation G.709. This encapsulation only used for WDM internal network, i.e. when we received signal from Client side which need to send to opposite site via OTN network, we will use GMP or GFP to do a encapsulation then send to WDM side. When we send signal out from WMD side to Client side, there had no GMP or GFP encapsulation. Like we receive a packet from 3rd party device A on site A, we will add GMP or GFP encapsulation on site A then send it to site B via OTN network, on site B we will remove the encapsulation then send it to 3rd party device B.

For the Local Fault Signal, when the WDM side detect fault, on Client side we will send out Local Fault Signal from client port to 3rd party device, the Local Fault Signal follow IEEE 802.3 protocol, there will have no GMP or GFP encapsulation at all.