Problem Description

One day, the user replaced the OSN3500 single-master control by himself, and used the "interrupt mode" to download, and the service had a 16S flash break. The user requested to test the download of OSN3500 single master board in the experimental environment.

Alarm information

TU_AIS. T2000 version is V2R6C03, OSN3500 version is 5.21.13.47.

Processing

The download process is analyzed as follows:

1. During the download process in the initialize network element step, the master data is cleared and enters the An state. At this time, the configuration is not sent to the single board, and the service is not interrupted.

2、After creating the single board and dynamic port, create the MSP ring protection group, and then issue the check command, at this time, the configuration will be issued to the single board, and the service starts to interrupt

3, SDH services are downloaded in packets, T2000 sends OSN3500 devices with the number of 200 crosses per packet, and each packet of services is validated once after the download is completed, and the next packet is downloaded after the host gives feedback on the success of the download. With the successful verification of each packet of services, the SDH services involved start to recover. When all SDH services are downloaded, all SDH services are restored.

4. The download order of cross-connections is randomized. There is no guarantee in the network management code that the one with the smaller cross-ID is downloaded first.

For details, see the attachment

Root Cause

Communicate with users, T2000 "interrupt mode" download for 5.0 platform equipment service interruption time is relatively long. Through the download test, the following conclusions:

During the download of OSN3500, if the conventional mode is adopted, the interruption time of SDH service is somewhat random. After downloading the ring multiplexing segment, linear multiplexing segment and other protections, the SDH service begins to be downloaded. the SDH service is downloaded in packets and verified in packets. Each packet is estimated to be around 200 items (estimated value, for reference only), and all SDH services are restored after the last packet is downloaded and verified successfully. Through several download attempts found that the interruption time of SDH service ranges from ten seconds to several minutes, and the download order of the network management is random. Theoretically speaking, if the download mode of OSN3500 is changed to "non-interrupt mode", the SDH service will not be affected. That is to say, if the download mode of OSN3500 is changed to "non-interrupt mode", the interruption time of SDH service is <=50mS.

There are several conditions for the service not to be affected:

The versions of host and service board must be matched.

2. The service configuration of the network management must be consistent with the current configuration of the network element.

3. All protection groups on the network element must not be in the inverted state.

4、Downloading process of the network element should not appear such as reset, fiber breakage and other abnormalities.

The risk of operational failure during the network management download process is high. The program to restore the business in the way of network management download is mainly for failure recovery, and the most common application scenario is that the master control replaces the single board. It is recommended that the master control master and backup dual distribution can effectively reduce the probability of download operation, which will greatly reduce the risk associated with network management download.

Recommendations and Summary

Due to the increasingly stringent assessment of operators on business interruption, it is required to replace the main control board business will not be affected. For Metro equipment, it is recommended to improve the download mechanism to achieve SDH service interruption time <=50mS;

For OSN3500 equipment single master control in addition to improve the download mechanism to achieve SDH service interruption time <= 50mS. In the market configuration is done in accordance with the master double match.