AAR - Application Probe Class - BFD question
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
04-10-2025 07:41 AM
Hello,
Please check the attachment for current settings, where Application Probe Class is not enabled yet.
I have simple question about BFD.
How many BFD packets will be in the output "show sdwan app-route stats summary" TOTAL PACKETS, when I enable Application Probe Class for SLA class Bulk-Data_wide, Video_wide, Voice_wide ?
Please note, there is a formula in Cisco documentation : "When BFD packets are sent with app-probe-class along with SLA class, the BFD packets are sent for each SLA app-probe-class separately in a round-robin manner."
I am only curios about the amount of BFD packets. Nothing more.
Many thanks,
Jan
- Labels:
-
Other SD-WAN
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
04-10-2025 11:10 AM
The BFD probes are set at multiplier 3 and poll interval of 180 seconds. The first table shows mpls transport with 7 classes. There is the default class(None) along with user configured classes of BE,AF1,AF2,AF3,AF4,EF. Add up all the probes on this mpls -> private transport for each index which roughly comes out to 28 or 29 probles on each index0 and there are 7 index zeros which comes out to about 200 probes total for index0 on the first table . The second table is another transport that only has the default BFD probe and no app-probe-class associated for a total of 200 bfd probes per index.
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
04-19-2025 01:24 AM
Hello Dan,
thank you for the reply.
I checked your output.
If you use an "app probe class", then you send fewer BFD packets overall, because BFD packets are sent for each SLA app-probe class separately in a round-robin manner—correct?
I'm trying to figure out if there's any additional BFD overhead when using app probe classes.
Many thanks for the feedback.
