[Case Show] Noise Occurs in the VoIP Service of the ONU Because the OLT Suffers

小小小飞鱼
小小小飞鱼  Platinum  (1)
6 years 11 months ago  View: 1253  Reply: 0
1F

Fault Description
Network topology: VoIP user -> MA5606T (ONU) -> MA5600T (OLT) -> SS -> UMG -> PSTN
Noise often occurs in the VoIP service of the ONU (MA5606T).

Alarm Information
None

Cause Analysis
-The quality of the subscriber line deteriorates.
-Interference source exists in the telecommunications room.
-The DDF grounding is improper.
-The service boards are faulty.
-The upstream transmission is faulty.

Procedure
1.Make calls for test on the DDF and the device separately. It is found that noise exists. This indicates that the noise is not caused by the poor quality of the subscriber line.
2.Check the environment of the telecommunications room. It is found that both cabling and the power supply meet the specifications and there is no strong electromagnetic interference. This indicates that the noise is not caused by the environment.
3.Test the grounding resistance. It is found that the grounding resistance is 3.48 ohms, which is in the normal range. This indicates that the noise is not caused by improper grounding of the DDF in the telecommunications room.
4.Test all the boards. It is found they are all faulty. This indicates that the noise is not caused by service boards.
5.Check the upper-layer device. It is found that an alarm is generated on the OLT indicating that the CPU usage is high (90%). Ping the gateway from the ONU. It is found that packets with small sizes are not lost but a long delay exists, and certain packets with large sizes are lost. It is suspected that the noise occurs because of the high CPU usage of the OLT. After checking, it is found that the OLT suffers the ARP attack, which leads to high CPU usage. Configure the firewall for the OLT. It is found that the CPU usage recovers and the noise occurred in the VoIP service of the ONU disappears.

Suggestion and Summary
To rectify the noise fault, follow the method of checking from nearness to farness and from the hardware to data. After an uncomplicated hardware component is replaced, if the fault persists, compare data to find out certain differences. That is, do not replace a complicated hardware component such as the backplane. This minimizes the effect on the current service.