[j-nsp] Difference between shaping-rate vs. transmit-rate?

[Pyxis LX]

Hi, All.

I'm quite confused about the differences between:
  shaping-rate
  transmit-rate (with rate-limit option turned on)

Here's my understanding of transmit-rate under class-of-service (I'm
not sure whether it's correct or not):

  1. The default setting is able to use the unused bw.

  2. If the "exact" option is turned on, the unused bw cannot be used
by this queue. When the queue is transmitting over the transmit-rate,
excess packets will be buffered. (Just like a policer with
burst-size-limit != 0?)

  3. If the "rate-limit" option is turned on, the unused bw cannot be
used. When the queue is transmitting over the transmit-rate, excess
packets will be dropped. (Just like a policer with burst-size-limit =
0?)

The shaping-rate configures PIR, so it should act like variant 3?

Thanks!