BFD UDP port ?

[Mario Rosi]

Hi,

my customer is complaining that between ME3600x and 7609 (both in release 15.3.3.S4) when there is congestion, the BFD session fall down and ISIS, BGP etc etc... adj as well.

I want to introduce into QoS also the BFD as match criteria in a class that already exist where there is a 4% of bw guaranteed and my idea is to configure something like this:

ip access-list extended BFD_ECHO

permit udp any any eq 3785

!

...

class-map match-any NETWORK-MPLS

 ...

 match access-group name BFD_ECHO

!

policy-map QOS-BB_MPLS-OUT

 class REALTIME-MPLS

   priority

   ...

 class NETWORK-MPLS

   bandwidth remaining percent 4

 ...

!

interface GigabitEthernet2

 ip address ...

 ...

 bfd interval 100 min_rx 100 multiplier 3

 service-policy output QOS-BB_MPLS-OUT

Now, my questions are:

1) is  3785 actually the right UDP destination port?

Because here https://www.cisco.com/c/en/us/td/docs/switches/metro/me3600x_3800x/software/release/12-2_52_ey/configuration/guide/3800x3600xscg/swiprout.html#90248 i've read that BFD packets are encapsulated in UDP packets with a destination port number of 3784 or 3785."

2) is BFD managed in HW, right? 

Thanks,

Mario

[Mark Malone · Accepted solution]

Hi

haven't done this before for BFD but see this below may be a good example for you , its  on 6500 but theory is the same

https://bst.cloudapps.cisco.com/bugsearch/bug/CSCtd04501/?referring_site=bugquickviewredir

[Mario Rosi]

Thanks Mark,

perfect, we are in the same picture. :)

Mario

[Mario Rosi]

Do you know if BFD packets are marked by the system as CS6 on AS1002x, 7609 and ME3600x?

Because my customer has already a Queue with that traffic and bandwidth enough to be maanged in case of congestion.

My doubt is that not all the previous platfmors mark it as CS6 and so the traffic goes in the default-class with 1% of remainng bw.

Mario

[Mark Malone · Accepted solution]

Hi

Yes CS6 is the marking its used control traffic for BFD that's how Cisco tags them in all platforms