From: Devi Mallampalli (Devi.Mallampalli@chubb.com.au)
Date: Mon May 03 2004 - 10:44:06 GMT-3
Ken,
I am afraid , this is a debatable issue. I think this is ( Modular Qos)
one of those grey areas where lot of advanced stuff is being written at
"CCO" and unfortunately with different interpretations. And for few of
us mortal souls who got to absorb & digest and put the config to the
"real use" on production networks , it is a bit daunting.
But benefit of doubt which I can give here is to the "IOS code"
variation. Particularly with Qos , I feel IOS version does make lot of
difference. And when you put this IOS issue to multiple platforms like
"VIP"based distributed CEFed 7500s to the modular non VIP 2600s
platforms, we got to be really careful on what are the capabilities of
both "queuing & shaping engines".
I agree with Bob and yours opinion on the need ( and benefit) to
practice each Mod Qos feature and find out the behavior ourselves.
Regards
Devi...
-----Original Message-----
From: Kenneth Wygand [mailto:KWygand@customonline.com]
Sent: Monday, 3 May 2004 11:01 PM
To: Devi Mallampalli; Bob Sinclair; GAD; ccielab@groupstudy.com
Subject: RE: CBWFQ Question
Devi,
I understand what you are saying, however according to the Cisco
documentation Gad and myself found, the default queueing for the
class-default class is _not_ FIFO, but flow-based WFQ. It also
explicitly states that when you configure a defined amount of bandwidth
for any particular class (class-default included), this does _not_
support flow-based WFQ, rather FIFO is the only queueing mechanism Cisco
supports when issuing the "bandwidth" command.
<snip>
If no default class is configured, then by default the traffic that does
not match any of the configured classes is flow classified and given
best-effort treatment. By default, the class-default class is defined as
flow-based WFQ. However, configuring the default class with the
bandwidth policy-map class configuration command disqualifies the
default class as flow-based WFQ. </snip>
http://www.cisco.com/en/US/products/sw/iosswrel/ps1830/products_feature_
guide09186a0080087a84.html#25297
Any thoughts?
Ken
-----Original Message-----
From: nobody@groupstudy.com on behalf of Devi Mallampalli
Sent: Mon 5/3/2004 8:46 AM
To: Bob Sinclair; GAD; ccielab@groupstudy.com
Cc:
Subject: RE: CBWFQ Question
Hi GAD,
I agree with Bob, the class of class-default will use "FIFO" as
its
default queuing mechanism. My understanding of CBWFQ's policy
maps, even
though class class-default will be activated as soon as you
create a
policy map for each kind of traffic as I have done at below
example ,
the default nature ( FIFO) will not be altered until , you "set"
the
condition with either "DSCP"/Ip precedence value or even
"bandwidth
command value".
In other words, where I have defined "fair-queue" on
class-default of
WAN-EDGE-DATA policy, the queuing for all NON matching traffic
other
than "Class DATA" will be put under "class-default" and assign a
flow
based fair queue". And on the other hand , if I do not "set" the
condition, under class-default, the default of "FIFO" will kick
in.
I tend to always define certain "set" condition under
class-default of
each "policy-map" with in CBWFQ so that "no single session or
user can
dominate the "wire" which can happen with in default "FIFO".
Also please note that when you enable FRTS on serial interface ,
she
will automatically "disable" the "fair queue" on the interface.
So it is
always a good idea to deploy some kind of congestion management
queuing
technique ( ex :CBWFQ , CQ ..etc)so that you are not allowing
any single
conversation on the link to consume the majority of the B/W.
policy-map WAN-EDGE-VOICE
class VOIP-RTP
priority 70
class VOIP-SIG
bandwidth 8
class class-default
bandwidth 40
policy-map WAN-EDGE-DATA
class DATA
priority 400
class class-default
fair-queue
>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>
random-detect
policy-map LAN-EDGE
class VOIP-RTP
set cos 5
class VOIP-SIG
set cos 3
class class-default >>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>
Router# sh frame pvc <dlci>
PVC Statistics for interface Serial0/0 (Frame Relay DTE)
DLCI = 200, DLCI USAGE = LOCAL, PVC STATUS = ACTIVE, INTERFACE =
Serial0/0.200
input pkts 220815371 output pkts 260090287 in bytes
4092929658
out bytes 1169002120 dropped pkts 114 in pkts
dropped 0
out pkts dropped 290002 out bytes dropped 432641264
late-dropped out pkts 290002 late-dropped out bytes
432641264
in FECN pkts 30 in BECN pkts 5 out FECN
pkts 0
out BECN pkts 0 in DE pkts 83264630 out DE pkts
0
out bcast pkts 625444 out bcast bytes 50327356
pvc create time 25w4d, last time pvc status changed 2w1d
cir 1792000 bc 1792000 be 0 byte limit 7168
interval 32
mincir 1792000 byte increment 7168 Adaptive Shaping none
pkts 260090184 bytes 1168895936 pkts delayed 151016040 bytes
delayed
243187049
shaping active
traffic shaping drops 289888
service policy WAN-EDGE-DATA
Serial0/0.200: DLCI 200 -
Service-policy output: WAN-EDGE-DATA
Class-map: DATA (match-any)
17897677 packets, 818527532 bytes
30 second offered rate 0 bps, drop rate 0 bps
Match: access-group name ACL-DATA-TELNET
17897635 packets, 818521832 bytes
30 second rate 0 bps
Match: protocol telnet
40 packets, 5612 bytes
30 second rate 0 bps
Queueing
Strict Priority
Output Queue: Conversation 72
Bandwidth 400 (kbps) Burst 10000 (Bytes)
(pkts matched/bytes matched) 681512/30904515
(total drops/bytes drops) 0/0
Class-map: class-default (match-any)
202495162 packets, 174811720189 bytes
30 second offered rate 1791000 bps, drop rate 2000 bps
Match: any
Queueing
Flow Based Fair Queueing
>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>
Maximum Number of Hashed Queues 64
(total queued/total drops/no-buffer drops) 23/277209/0
exponential weight: 9
-----Original Message-----
From: Bob Sinclair [mailto:bsin@cox.net]
Sent: Monday, 3 May 2004 10:40 AM
To: GAD; ccielab@groupstudy.com
Subject: Re: CBWFQ Question
Gad,
Class-default can use FIFO or WFQ. I believe FIFO is the
default.
HTH,
Bob Sinclair
CCIE #10427, CISSP, MCSE
www.netmasterclass.net
----- Original Message -----
From: "GAD" <gad@gad.net>
To: <ccielab@groupstudy.com>
Sent: Sunday, May 02, 2004 8:13 PM
Subject: CBWFQ Question
> Question. If you do not specify the "class-default" class in a
> policy-map, the class-default class appears, but it does not
say
> "Weighted fair queueing". Can someone tell me what it IS
doing?
>
> ex:
>
> class-map match-all Prec-5
> match ip precedence 5
>
> policy-map GAD
> class Prec-5
> bandwidth percent 90
>
> int s0
> service-policy outbound GAD
>
> R3#sho policy-map interface s0
>
> Serial0
>
> Service-policy output: GAD
>
> Class-map: Telnet (match-all)
> 0 packets, 0 bytes
> 5 minute offered rate 0 bps, drop rate 0 bps
> Match: ip precedence 5
> Weighted Fair Queueing
> Output Queue: Conversation 265
> Bandwidth 90 (%) Max Threshold 64 (packets)
> (pkts matched/bytes matched) 0/0
> (depth/total drops/no-buffer drops) 0/0/0
>
> Class-map: class-default (match-any)
> 1 packets, 13 bytes
> 5 minute offered rate 0 bps, drop rate 0 bps
> Match: any
>
> Difference when you add:
>
> class class-default
> bandwidth percent 10
>
> To the policy map is:
>
> R3#sho policy-map interface s0
>
> Serial0
>
> Service-policy output: GAD
>
> Class-map: Telnet (match-all)
> 0 packets, 0 bytes
> 5 minute offered rate 0 bps, drop rate 0 bps
> Match: ip precedence 5
> Weighted Fair Queueing
> Output Queue: Conversation 265
> Bandwidth 90 (%) Max Threshold 64 (packets)
> (pkts matched/bytes matched) 0/0
> (depth/total drops/no-buffer drops) 0/0/0
>
> Class-map: class-default (match-any)
> 38 packets, 494 bytes
> 5 minute offered rate 0 bps, drop rate 0 bps
> Match: any
> Weighted Fair Queueing
> Output Queue: Conversation 266
> Bandwidth 10 (%) Max Threshold 64 (packets)
> (pkts matched/bytes matched) 0/0
> (depth/total drops/no-buffer drops) 0/0/0
>
> Notice the addition of the last 5 lines.
>
> So my question is, in the first example, what is the queuing
> mechaninsm if not WFQ? Queueing for the interface is Weighted
Fair as
> shown in in a show
> int:
>
> Serial0 is up, line protocol is up
> Hardware is HD64570
> Internet address is 208.28.1.2/29
> MTU 1500 bytes, BW 1544 Kbit, DLY 20000 usec,
> reliability 255/255, txload 1/255, rxload 1/255
> Encapsulation FRAME-RELAY, loopback not set
> Keepalive set (10 sec)
> LMI enq sent 43526, LMI stat recvd 43527, LMI upd recvd 0,
DTE LMI
up
> LMI enq recvd 0, LMI stat sent 0, LMI upd sent 0
> LMI DLCI 1023 LMI type is CISCO frame relay DTE
> FR SVC disabled, LAPF state down
> Broadcast queue 0/64, broadcasts sent/dropped 2/0, interface
> broadcasts 0
> Last input 00:00:04, output 00:00:04, output hang never
> Last clearing of "show interface" counters 5d01h
> Input queue: 0/75/0/0 (size/max/drops/flushes); Total output
drops:
0
> Queueing strategy: weighted fair
> Output queue: 0/1000/64/0 (size/max total/threshold/drops)
> Conversations 0/1/256 (active/max active/max total)
> Reserved Conversations 2/2 (allocated/max allocated)
> Available Bandwidth 1158 kilobits/sec
> 5 minute input rate 0 bits/sec, 0 packets/sec
> 5 minute output rate 0 bits/sec, 0 packets/sec
> 50821 packets input, 2826502 bytes, 0 no buffer
> Received 0 broadcasts, 0 runts, 0 giants, 0 throttles
> 0 input errors, 0 CRC, 0 frame, 0 overrun, 0 ignored, 0
abort
> 43576 packets output, 570160 bytes, 0 underruns
> 0 output errors, 0 collisions, 3 interface resets
> 0 output buffer failures, 0 output buffers swapped out
> 1 carrier transitions
> DCD=up DSR=up DTR=up RTS=up CTS=up
>
> I'm sure the answer is obvious somehwere, but I'm fried and
it's
> eluding me.
>
> Thanks!
>
> GAD
>
>
This archive was generated by hypermail 2.1.4 : Wed Jun 02 2004 - 11:12:03 GMT-3