Re: Traffic Shaping question

From: Joe Chang (changjoe@earthlink.net)
Date: Sun May 23 2004 - 19:31:17 GMT-3


> I don't follow what you mean here.
> Bucket size will be Bc + Be, but only Bc applies to "refill every Tc".

The bucket is given another Bc tokens at the beginning of a Tc. Unused
tokens at the end of Tc are not discarded but carried over to the next
period; the bucket can store up to Bc + Be tokens

> Peak rate is something that only makes sense when you shape to peak
> (i.e. an analogy to Be=0) or else is something very lousy.

By peak rate I meant the access rate of the interface. Sorry for the
confusion. .

>> Don't follow you here either. Shaping and policing are quite different
> animals, even though their implementation use quite simmilar tools.
> Cisco tends to explain them based on simmilarities, and that sometimes
> generates some confusion.

I'm just pointing out one way to describe that difference. If it doesn't
work for you, don't worry about it.

> I like to think of shaping as what you do not to be policed. With same
> parameters, a shaped traffic will not be dropped/marked by a policer.
> In that sense, "save now" is just "have up to Be available for later
> use", and "spend more now" is "use that saved Be". Same thing...
>

Not true. Unlike traffic shaping, traffic policing has the accumulated debt
and compounded debt features. The way compounded debt is handled gives
traffic shaping a RED type behavior when traffic exceeds Bc but not the Be.



This archive was generated by hypermail 2.1.4 : Wed Jun 02 2004 - 11:12:16 GMT-3