RE: QOS question on Bc with CAR

From: Jonathan V Hays (jhays@jtan.com)
Date: Tue Apr 08 2003 - 00:50:17 GMT-3


> -----Original Message-----
> From: nobody@groupstudy.com [mailto:nobody@groupstudy.com] On
> Behalf Of Richard Davidson
> Sent: Monday, April 07, 2003 6:48 PM
> To: wsqccie@hotnail.com; Rick; ccielab@groupstudy.com; Scott
> M. Livingston
> Subject: Re: QOS question on Bc with CAR

[snip]

> course of time, you want to transfer data. The T in
> this case time, represents 1 second. This one second
> is broken in what you know as Tc. Tc is an increment
> of the T which is in msec. 1 second / 8 = .125. This
> is what is usually used as the Tc (.125). If you are
> using Voice, the T (1 second) is recommended to be
> broken into 10 time slots. 1 second / 10 = .1.

No. For voice you would have one hundred (100) intervals in a second if
you use Tc=10 ms.

The recommendation for voice is to configure so that the interval (Tc)
is 10 milliseconds, rather than 125 milliseconds. This is simply because
voice cannot tolerate much more of a delay than 10 ms. Since you cannot
configure Tc directly, you need to configure Bc, the committed burst
size (in bits).

Bc = CIR*Tc
   = CIR*(10 ms)
   = CIR*(.010)
   = CIR/100



This archive was generated by hypermail 2.1.4 : Thu May 01 2003 - 13:35:48 GMT-3