QoS Marking of BGP Traffic

[Robert Juric]

We have an ASR1001X running BGP over an ethernet link to our MPLS provider. We have had occasional outages where BGP bounces and we believe this is being caused by micro-bursts (historical logs don't show pattern of over-utilization).

The only drops were are seeing are in the Best Effort queue but it made me question the marking of our BGP traffic. When I look at our Netflow records it is showing as CS0 both inbound and outbound.

I thought that RP generated traffic should be marked with IP Precedence 6 by default. However that doesn't look like the case based on Netflow. Is netflow incorrect or would we be missing any config necessary?

[Harold Ritter]

Hi Robert,

BGP packets, like most control protocol traffic should definitely be marked as CS6.

Quote from the following document:

https://www.cisco.com/c/en/us/td/docs/solutions/Enterprise/WAN_and_MAN/QoS_SRND/QoS-SRND-Book/QoSIntro.html

"IP Routing traffic should be marked to DSCP CS6; this is default behavior on Cisco IOS platforms."

Regards,

[Robert Juric]

Yes, that was my understanding as well. I thought I saw the traffic between endpoints as CS0, but upon further investigation it was ICMP traffic used during troubleshooting at the time of the BGP flap. The BGP traffic doesn't show up in Netflow at all.

Thanks for taking a look though.

[Joseph W. Doherty]

Having BGP tagged with CS6 is good, but tagging it alone doesn't guarantee such tagged traffic gets preferential treatment. You'll need to check that, end-to-end.