Showing results for 
Search instead for 
Did you mean: 
Join Customer Connection to register!

3750 MLS to 3850 MQC Conversion of QoS Configuration

Hi everyone,

I am sorry to bother you with this matter but I am struggling to adapt one part inside the QoS configuration

I had on a Catalyst WS-C3750X-48 (version 15.2(4)E9) to a new switch 3850X-48 (version 16.12.05b)


I have been able to translate the class-map, policy-map and queue-set. However for below part I am stuck.

Here is what I had inside the old config:


mls qos queue-set output 2 threshold 1 100 100 100 400
mls qos queue-set output 2 threshold 2 200 200 100 400
mls qos queue-set output 2 threshold 3 100 100 20 400
mls qos queue-set output 2 threshold 4 100 100 100 400


According to below Cisco configuration:

Example 10


We should get something like this on 3850 (off course by changing numbers inside):


queue-limit cos  2 percent 50
queue-limit cos 3 percent 50
queue-limit cos 6 percent 70
queue-limit cos 7 percent 70

but I am completely lost on how I can translate threshold I had previously to something similar into the new syntax. Cisco documentation is not very clear to me on how to adapt queues and buffers policy to 3850.

Is the logic different, with calculation or else to do ?

Has someone any idea about this ?



Best regards,


Georg Pauwen
VIP Expert



tough one, this has come up quite a bit in the past. There is not really a 1 to 1 conversion as far as I recall. It is mostly class-based, and you need to decide which traffic you want to prioritize, and assign that traffic to classes...



Joseph W. Doherty
Hall of Fame Expert

The 3850's QoS architecture is both similar and in some ways "improved/better" than the 3750's.

As, at least in my approach, platform QoS implementations are often just a "subset" of what my logical QoS model/policy is.  For any new or different platform, rather than just trying to "replicate" one platform's QoS implementation to another (different) platform, I try to get the new/different platform to best implement my logical QoS model/policy.

Might this be possible for you too?

One reason I ask, you seem to be asking (mostly) about replication of WTD, which, can be a bit more advanced part of a QoS model/policy.  If you don't really have your own logical QoS model/policy dealing with WTD, and/or early dropping, you might be better off just using the 3850's defaults or set WTD to all be max.

I have used enabling Auto QoS on a single port on a new platform to get some of the basic stuff created. If the values from Auto QoS meet your needs, great! If not, you can use the platform specific entries created by Auto QoS as a template.