Disclaimer
The Author of this posting offers the information contained within this posting without consideration and with the reader's understanding that there's no implied or expressed suitability or fitness for any purpose. Information provided is for informational purposes only and should not be construed as rendering professional advice of any kind. Usage of this posting's information is solely at reader's own risk.
Liability Disclaimer
In no event shall Author be liable for any damages whatsoever (including, without limitation, damages for loss of use, data or profit) arising out of the use or inability to use the posting's information even if Author has been advised of the possibility of such damage.
Posting
1. can it be in both classes? If yes, Are there any problems with this?'2. Should it only be in one class?
3. should it be in the Parent policy TEST_ETH10000 under class-default so it applies random-detect to all classes in the child policy TEST-MPLS-QoS ??
How are other enterprises implementing random-detect?
#1 Sure, it can be used in both classes or multiple classes. Possible problems, sure too - it's an "it depends" kind of answer.
#2 Another "it depends". Doesn't have to be. If FQ is supported on the platform I recommend that over RED.
#3 No, as parent policy is pushing queuing to child policy.
Re: other Enterprises
Don't know, but most, I think, don't really understand how to use RED effectively, which can be surprising difficult to accomplish (except for tiered drop points). For example, RED was designed to interact with (equal valued) flow rate adaptive traffic, such as TCP. If your class-default or scavenger has other than flow adaptive traffic, you're probably suboptimally using RED. Or, for example, do you allow RED to target non-piggy-backed ACK packets? If so, are you doing that by design or just allowing RED to target all class traffic? If the latter, again, you're probably suboptimally using RED.
Besides many not really understanding RED, many don't, I also think, understand the finer nuisances of QoS. For example, your scavenger class is provided 1% remaining bandwidth but is also policed at 40%. Why the policer if the scheduler effectively is providing this class available bandwidth. I.e. why limit available bandwidth, unless you pay extra for exceeding 40% bandwidth. But if if you did, why use a policer and not a shaper?
