EVPN performance and fault management
Operation, Administration, and Maintenance (OAM) plays a critical role in daily network monitoring. ITU-T Y.1731 provides comprehensive guidelines for performance and fault management in Ethernet-based networks. In our testing, we specifically focused on performance management metrics, including frame loss ratio and frame delay.
We set up an EVPN-VPWS service between two DUTs and used Calnex SNE Ignite to introduce controlled packet drops and delays. First, we sent end-to-end unicast traffic with no impairment; the DUT Connectivity Fault Management (CFM) output showed no packet loss and minimal delay. Next, we introduced a 30% bidirectional packet drop and a 30ms bidirectional delay. The measured CFM latency matched the impairment tool’s configuration. For frame loss, although the impairment tool was set to drop 30% of packets in each direction, the cumulative effect resulted in about 49% total frame loss (If we send 100 requests in one direction, the destination will receive only 70 (30% drop). Then the destination will send 70 responses to the source, and the source will receive only 49 (30% drop). So it's 49% in math). The DUT CFM reported a loss ratio consistent with this calculation, around 50%, as expected.
Figure 25: EVPN performance management
Next, we addressed fault management by simulating a fault: we shut down the link between the PE and CE. The CFM output on the DUT confirmed that the EVPN-VPWS service was down. After restoring the link, the EVPN-VPWS was up again. The following is the topology we used in the test.
Figure 26: EVPN fault management
The following table shows the DUT combination in our test.
| Combination | PE1 | PE2 |
|---|---|---|
| 1 | Cisco 8711-48Z-M | HPE MX304 |
| 2 | Ciena 5164 | Nokia 7750SR-1 |
| 3 | Ciena 8192 | Cisco 8711-48Z-M |
| 4 | Ciena 8192 | HPE MX304 |
| 5 | Ciena 5164 | Ericsson RAN Connect 6682 |
Table 2: EVPN fault management participants list
| < Previous | Next > |

