Hello,
Today, I would like to share with you how to solve the Issues that need to be paid attention to when dual-homing protection between Hybrid MSTP OSN box equipment and router.
Problem Description
At present, the scenarios where Hybrid MSTP devices and routers are connected with dual protection are increasingly being applied to existing networks by customers. However, improper use of this kind of networking will cause various unexpected problems.
Since September 2014, in a network that uses Hybrid MSTP equipment and routers to form a dual-protection delivery network, the OSN equipment frequently reports FLOW_OVER alarms, causing service interruptions.
Due to the lease of a third-party network, the bandwidth was small, and after the use of dual-profile protection, there were also services interruptions caused by packet loss and link congestion.
Earlier, at a certain office, dual protection was also used. After the router link was adjusted, the OSN device continued to report the FLOW_OVER alarm, which caused a large-scale service interruption.
The problems of the above several sites are characterized by using Hybrid MSTP and routers to form dual protection. The simple networking is as follows:
Generally, a 10GE port on OSN7500II_01 and a 10GE port on OSN7500II_02 form a 10GE service ring with other devices.
Process
Existing problems:
1- Customers and front-line companies like simple configuration, directly configure an ELAN service on the OSN7500II, form a VRRP protection with the router, and then configure the PW (Pseudo Wire) to the OSN site connected below. Each time you expand or add a new site, you directly add a PW to the services, such as a network diagram, adding a working PW and a protection PW. More than 40 PWs may be configured on each OSN7500II. When the OSN7500II clears the Mac address due to port link status changes or other reasons, all services on the OSN7500II will be broadcast, and the services sent by the router to the NodeB will be broadcast several times the number of PWs, that is, if the router port sends 300M traffic, will generate 300M*PW number (calculated based on 40) = 12G traffic, far exceeding the bandwidth of the OSN10GE port. Therefore, it is recommended to divide the services on the front line. If 40 PWs are divided into 2 services, each with 20 PWs and 150M traffic, the result of broadcast replication is 150M*20*2=6G, which will produce good results. The effect avoids congestion.
2- As shown in the network diagram, when OSN7500II_01 is connected to the main router, the PW101 of OSN7500II_02 will also receive service packets from multiple NodeBs when the Link status of all ports has not changed. Since the Mac address has not been learned, OSN7500II_02 broadcasts these messages after receiving these messages, causing congestion on the 10GE port of the OSN7500II_02 network element to report the FLOW_OVER alarm. This problem has occurred in multiple sites, so it is very typical. After positioning, it is found that the MAC address aging time of the ELAN service on the OSN is 5 minutes, while the default time of some protocols of the router is 20 minutes. When the MAC address learned by the OSN ages, it will take 15 minutes to learn the MAC address again, and during these 15 minutes, the service has been broadcasting. Therefore, when connecting with the router, you need to confirm the time interval of protocol packet sending with the router, and set the aging time on the OSN to be twice that of the router. The services with black lines as shown below.
3- Due to the characteristics of the VRRP protocol of the router, the services from NodeB to the router needs to be forwarded to the VRRP master router, while the services from the router to the NodeB is uncertain. It may be delivered from the master router, and the services may be Issued by the standby router. When the service is delivered by the backup router, the service may also be broadcast for a long time. As shown in the figure below, when the router's transmission and reception are separated, the services sent to the router are only sent to OSN7500II_01, and OSN7500II_02 will not learn the MAC address of the NodeB. At this time, the service sent by the backup router to NodeB is broadcast on OSN7500II_02 because there is no Mac address of NodeB, which causes the 10GE outlet on OSN7500II_02 to be congested and report the FLOW_OVER alarm. Therefore, when configuring the router, avoid this scenario.
Root cause
1- There are too many PWs in the service configuration, which leads to excessive traffic and packet loss during broadcast replication of service packets. The service configuration and quantity need to be adjusted.
2- When working with routers, the protocol time needs to be set reasonably, and the aging time on the OSN should be greater than that configured on the router.
3- When the router is protected by VRRP, there may be a scenario where the transmission and reception are separated. You can configure the router to force the transmission and reception to be consistent, which is completed by the same router.
Solution
1- Reasonably configure services to ensure that bandwidth requirements can be met during broadcasting.
2- Configure the aging time of the ELAN service on the OSN to be longer than the interval time of the router protocol packet, preferably twice the time of the router.
3- When configuring the router, there will be no separation of sending and receiving.
You are welcome to leave a message and exchange in the comment area. Thank you!
