Note: This kind of scenario is not possible as other interfaces can also use the buffer, but, this can definitely help to reduce the packet drops to a certain level.
Laugh - perhaps TechNote needs to be read a third time?
Right above the that note is (which I believe is the object of the note):
The Softmax for queue-0 can go up to 10,000 buffers and as a result, the drops are Zero.
I.e. I believe the "scenario" is actually obtaining 10,000 buffers. Which cannot actually happen, or "is not possible as other interfaces can also use the buffer".
Further, that note ends with:
but, this can definitely help to reduce the packet drops to a certain level.
I.e. in reality, the drops are Zero. is unlikely or not guaranteed with this setting.
What the TechNote doesn't have, is anything like, for this Case Study example is "DON'T TRY THIS AT HOME". ; )
At to your statements:
"10000 buffer meaning that all softmax will be use by only this interface. Other interface will face huge drop issue."
I would revise those to be:
10000 buffer meaning that all softmax might be used by such an interface. Other interfaces might have increased drops.
Assuming my revised statement are correct, often the only way to know for sure the impact, good or bad, is to try a change.
PS:
BTW, my intent is not to totally discount @MHM Cisco World 's concern, as there really is a possibility of these changes creating an overall negative impact. Even if that happens, negative results, can lead to later positive results (assuming we learn from the negative results).
