After the router in a softswitch office is changed from NE40 to NE40E, packet loss occurs when the EFGE electrical Ethernet port of the NE40E is connected to the IFM board of the MSOFTX3000 and the packet loss rate is 86%.
1. Check the working mode of the egress GE electrical port of the NE40E. The working mode is auto negotiation mode and the negotiated result is 100 Mbit/s full-duplex. The link status is normal. Set the network port at the two ends to 100 Mbit/s full-duplex mode by force. The fault still persists and the link status is normal. Therefore, the packet loss is not caused by the port negotiation failure.
2. Directly connect the NE40E to the PC, and PING NE40E on the PC. The ping is successful and no packet is lost. The NE40E can be successfully pinged on the PC, indicating that the NE40E can normally receive and send packets and process PING packets normally.
3. Once the original network cable is added with prolonged cables to perform the ping test, the connector on the MSOFTX3000 remains unchanged and the connector on the NE40E side is replaced with the connector on the prolonged cable. The ping test is performed successfully. Therefore, the original cable connecting the NE40E and the MSOFTX3000 is normal.
4. The same problem also occurs after the BFI board is replaced with many other BFI boards with different manufacturing dates (including the BFI board manufactured on August of 2008). Therefore, the fault is not related to the BFI board manufacture date. The LSW is added between the BFI board and the NE40E, neither does underlayer error packet nor ping packet loss occurs and the SCTP alarm is no longer generated at the background. In addition, the BFI can receive packets from the peer end. After the NE40E is replaced with the NE40, the SCTP alarm is not generated either. Therefore, the problem is caused by the improper mapping of the MSOFTX3000, the connecting cable, and the NE40E.
5. Check the difference between the original cable and the new cable. It is found that the code of the new cable is 19-04070012-0218260830, whereas the code of the original cable is 19-04040101-0217480750V4. Therefore, the problem may be caused by the type of the new cable.
6. Use the new cable to connect a normal BFI board with the NE40E, the ping test is successful no matter to which port on the NE40E is connected to the cable.
7. Use the new cable to connect the BFI board brought from the live network with the NE40E, the ping test is successful no matter to which port on the NE40E is connected to the cable.
Therefore, packet loss does not always occur when the NE40E is connected to the MSOFTX3000 using the new cable.
8. Use the new cable to connect the NE40E with the MSOFTX3000 through the S3526E, the MSOFTX3000 is pinged through on the NE40E.
The preceding tests show that the problem may be caused by the network cable type. The new cable brought from the live network has trapezium connectors and the trapezium connectors had brought problems before.
The narrowest place of a trapezium connector is 438 um and the broadest place is 743um, whereas the width of a common rectangle connector is 640 um. However, the width of the contact part of RJ-45 ports from almost all manufactures is about 450 um. That is, if the contact part of the RJ-45 port contacts the insulation part of the trapezium RJ-45 connector, the transmission is affected. In addition, BFI boards are different from each other, so this problem may occurs when the cable is connected to some BFI boards whose RJ-45 port cannot properly connect to the trapezium connector.
Replace the BFI board or the connecting cable to address this problem.
The trapezium RJ-45 connector is forbidden to be delivered. In this way, problems caused by the trapezium RJ-45 connector will never occur.