From: Lupi, Guy (Guy.Lupi@xxxxxxxxxxxxx)
Date: Tue Apr 02 2002 - 23:05:14 GMT-3
Your calculations match up to theirs, thanks to you and Bob I actually
understand this now. This is unbelievable, just when you thought you
understood something. Well, better before the test than after........
~-----Original Message-----
~From: Darren Ward [mailto:dward@pla.net.au]
~Sent: Tuesday, April 02, 2002 9:08 PM
~To: Lupi, Guy
~Cc: 'ccielab@groupstudy.com'
~Subject: Re: FRTS
~
~
~Not really.
~
~Time = bc/cir so that 12,500/100,000 gives a sampling time of 125msec.
~
~this means the committed rate is 12.5Kb per 125msec which adds
~up to 100Kb
~per second.
~
~burst excess is over the same sampling period defined by the
~ration of cir
~and bc which is 1/8th second or 125msec so it's not 28Kb per
~second you've
~defined but 28Kb per 125msec or 204Kb per second :)
~
~The correct value for 'be' per 125msec to get 28Kbps is 3500
~
~Darren Ward
~(PGradCS, CCIE #8245, CCNP, CCDP, MCP)
~
~
~On Tue, 2 Apr 2002, Lupi, Guy wrote:
~
~> If I want to send traffic at an average rate of 100K, being
~able to burst to
~> 128K if credit is built up and dropping to a minimum of 56K during
~> congestion, does this map class look correct? I am trying
~to figure out if
~> I have this down or not, I am trying to follow the
~guidelines I found on
~> CCO. Thanks.
~>
~>
~> map-class frame-relay tor1
~> frame-relay cir 100000 (send traffic
~at an average
~> of 100000)
~> frame-relay bc 12500 (cir/8)
~> frame-relay be in 56000 (added automatically)
~> frame-relay be out 28000 (burst in
~first interval)
~> frame-relay mincir 56000 (drop to a
~minimum of
~> 56K)
~> frame-relay adaptive-shaping becn
This archive was generated by hypermail 2.1.4 : Thu Jun 13 2002 - 10:57:52 GMT-3