FRF.12 bug?!?

From: Adam Elghafri (ccie.adam@gmail.com)
Date: Tue Jul 15 2008 - 00:24:13 ART


guys i have question in IEWB vol I Qos

Ar = 64 kbps
CIR = 56 kbps
Tc = 10ms

so bc = 560 or 70byte
and be = 640 or 80 byte

and the equation to find the fragment size for serialization delay of
10 ms (notice that the tc = min serialzation delay)

in FRF.12 we will use the following equation

Portspeed*minimum serialization delay = 64k*10m =640 bites

but in this situation... the 640 bits wont fit in one single interval
.. it has to go to the other interval .. so the delay eventually will
become more that 10 ms !!!!!!!!!!!!!!!!!

please refere to this pic
http://tinypic.com/view.php?pic=2wf5rfc&s=4

ok what i concluded is the following:
as long as the tc is greated than the serialization delay.. then
nothing wronge will happen.. but when tc is as low as the serialzation
delay.. then the serialization will become slitly bigger...

thats why this fix the minimum tc to 10ms

CUTION.. dont fill in the mistake of substituing the CIR in the
equation insted of AR.. coz the interface will send at AR not CIR...
thats why u see in the pic times of max sending and times of compete
scilence... (all during the tc)

correct me if i am wrong..

thanks

Adam Elghafri



This archive was generated by hypermail 2.1.4 : Mon Aug 04 2008 - 06:11:55 ART