From: Lupi, Guy (Guy.Lupi@eurekanetworks.net)
Date: Mon May 10 2004 - 08:35:40 GMT-3
From:
http://www.cisco.com/univercd/cc/td/doc/product/software/ios123/123cgcr/qos_
r/qos_s1g.htm#1085303
Usage Guidelines
Traffic shaping limits the rate of transmission of data. In addition to
using a specifically configured transmission rate, you can use Generic
Traffic Shaping (GTS) to specify a derived transmission rate based on the
level of congestion.
You can specify two types of traffic shaping; average rate shaping and peak
rate shaping. Average rate shaping limits the transmission rate to the CIR.
Using the CIR ensures that the average amount of traffic being sent conforms
to the rate expected by the network.
Peak rate shaping configures the router to send more traffic than the CIR.
To determine the peak rate, the router uses the following formula:
peak rate = CIR(1 + Be / Bc)
where:
Be is the Excess Burst size.
Bc is the Committed Burst size.
Peak rate shaping allows the router to burst higher than average rate
shaping. However, using peak rate shaping, the traffic sent above the CIR
(the delta) could be dropped if the network becomes congested.
If your network has additional bandwidth available (over the provisioned
CIR) and the application or class can tolerate occasional packet loss, that
extra bandwidth can be exploited through the use of peak rate shaping.
However, there may be occasional packet drops when network congestion
occurs. If the traffic being sent to the network must strictly conform to
the configured network provisioned CIR, then you should use average traffic
shaping.
-----Original Message-----
From: MMoniz [mailto:ccie2002@tampabay.rr.com]
Sent: Sunday, May 09, 2004 10:14 PM
To: Kenneth Wygand; ccielab@groupstudy.com
Subject: RE: Shaping to Peak?
Well as I said..QOS is not my strongest area...
But I think the first option will not invoke shaping parameters until the BW
is reached or carrier becns are received, of course if it exceeds your CIR
it is very possible traffic will be dropped. This will allow the PVC to use
whatever is available until the carrier starts dropping your traffic. Not
good for time sensitive type traffic but for fire and forget it is good.
The second option will start to invoke shaping parameters at 64 k if it is
reached.
This statement from the QOS doc is what I am trying to explain.
Frame Relay Adaptability to Congestion Example
This example does not restrict flow across a Frame Relay subinterface that
has been layered onto a single data-link connection identifier (DLCI).
However, in the presence of BECN bits from the network, the flow is
throttled back to the committed information rate (CIR). The access rate of
the interface is assumed to be 1544 kbps, and the CIR is 64 kbps.
interface serial 2
traffic-shape rate 1544000
traffic-shape adaptive 64000
traffic-shape fecn-adapt
-----Original Message-----
From: nobody@groupstudy.com [mailto:nobody@groupstudy.com]On Behalf Of
Kenneth Wygand
Sent: Sunday, May 09, 2004 9:16 PM
To: MMoniz; ccielab@groupstudy.com
Subject: RE: Shaping to Peak?
Mike,
Thanks for your input. Let's use your case for example. What is the
_functional_ (or technical) difference between the way these two options
will shape?
OPTION 1:
policy-map POLICE
class class-default
shape average 1536000
OPTION 2:
policy-map POLICE
class class-default
shape peak 64000 6400 147200
Thanks!
Ken
-----Original Message-----
From: MMoniz [mailto:ccie2002@tampabay.rr.com]
Sent: Sun 5/9/2004 8:18 PM
To: Kenneth Wygand; ccielab@groupstudy.com
Cc:
Subject: RE: Shaping to Peak?
Kenneth, the only reason I can think of is that a lot of carriers
never enforce CIR!!
For instance, I have had circuits that were TI access but suppose to
be like 256 CIR. By shaping to
the max T1 you still have some control of throttling if congestion
occurs.
And at many times the traffic rate was well above the CIR but didn't
get throttled because we were shaping to
much higher. It was only degraded when the carrier was
oversubscribed or degraded.
Does that make sense? Kind of hard to explain
mike
BTW...QOS is one of my weakest subjects as is with lot's of other
people.
-----Original Message-----
From: nobody@groupstudy.com [mailto:nobody@groupstudy.com]On Behalf
Of
Kenneth Wygand
Sent: Sunday, May 09, 2004 7:58 PM
To: ccielab@groupstudy.com
Subject: Shaping to Peak?
Hey everyone,
I'm really trying to understand all the options for QoS and I'm
currently looking at CB Shaping. Why would one want to shape to a _peak_
instead of shaping to an _average_? I understand that shaping to a _peak_
will try to send Bc + Be bits during every Tc, but doesn't that defeat the
whole purpose of shaping? Why not just shape to an _average_ but just make
the Bc the total combined value of the Bc and Be values used when shaping to
a _peak_, and not configure a Be?
Am I missing something?
TIA,
Ken
This archive was generated by hypermail 2.1.4 : Wed Jun 02 2004 - 11:12:08 GMT-3