From: gladston@br.ibm.com
Date: Wed Jul 27 2005 - 17:45:22 GMT-3
Thanks a lot Chris,
I did the test again today. Although it was not possible to get a small
delay, changing the priority kbps helped to not lost packets. That was for
LLQ.
With PQ, packets are lost.
Debug shows that PQ matches the packet, but the frequency of the message
is very low, which makes me think IOS is not matching all priority
packets; if that is true, it would explain packets lost:
.Feb 28 22:28:58: PQ: Serial0/0 dlci 503 : ip (defaulting) -> low
.Feb 28 22:28:58: PQ: Serial0/0 dlci 503 : ip (defaulting) -> low
.Feb 28 22:28:58: PQ: Serial0/0 dlci 503 : ip (defaulting) -> low
.Feb 28 22:28:58: PQ: Serial0/0 dlci 503: ip (s=148.5.15.10, d=148.5.3.1)
-> high
.Feb 28 22:28:58: PQ: Serial0/0 dlci 503 : ip (defaulting) -> low
.Feb 28 22:28:58: PQ: Serial0/0 dlci 503 : ip (defaulting) -> low
.Feb 28 22:28:58: PQ: Serial0/0 dlci 503 : ip (defaulting) -> low
.Feb 28 22:28:58: PQ: Serial0/0 dlci 503 : ip (defaulting) -> low
.Feb 28 22:28:58: PQ: Serial0/0 dlci 503 : ip (defaulting) -> low
.Feb 28 22:28:58: PQ: Serial0/0 dlci 503 : ip (defaulting) -> low
.Feb 28 22:28:58: PQ: Serial0/0 dlci 503 : ip (defaulting) -> low
.Feb 28 22:28:58: PQ: Serial0/0 dlci 503 : ip (defaulting) -> low
.Feb 28 22:28:58: PQ: Serial0/0 dlci 503 : ip (defaulting) -> low
.Feb 28 22:28:59: PQ: Serial0/0 dlci 503 : ip (defaulting) -> low
.Feb 28 22:28:59: PQ: Serial0/0 dlci 503 : ip (defaulting) -> low
.Feb 28 22:28:59: PQ: Serial0/0 dlci 503 : ip (defaulting) -> low
.Feb 28 22:28:59: PQ: Serial0/0 dlci 503 : ip (defaulting) -> low
.Feb 28 22:28:59: PQ: Serial0/0 dlci 503 : ip (defaulting) -> low
.Feb 28 22:28:59: PQ: Serial0/0 dlci 503 : ip (defaulting) -> low
.Feb 28 22:28:59: PQ: Serial0/0 dlci 503 : ip (defaulting) -> low
.Feb 28 22:28:59: PQ: Serial0/0 dlci 503 : ip (defaulting) -> low
.Feb 28 22:28:59: PQ: Serial0/0 dlci 503 : ip (defaulting) -> low
.Feb 28 22:28:59: PQ: Serial0/0 dlci 503 : ip (defaulting) -> low
.Feb 28 22:29:00: PQ: Serial0/0 dlci 503 : ip (defaulting) -> low
.Feb 28 22:29:00: PQ: Serial0/0 dlci 503 : ip (defaulting) -> low
.Feb 28 22:29:00: PQ: Serial0/0 dlci 503 : ip (defaulting) -> low
.Feb 28 22:29:00: PQ: Serial0/0 dlci 503 : ip (defaulting) -> low
.Feb 28 22:29:00: PQ: Serial0/0 dlci 503 : ip (defaulting) -> low
.Feb 28 22:29:00: PQ: Serial0/0 dlci 503 : ip (defaulting) -> low
.Feb 28 22:29:00: PQ: Serial0/0 dlci 503 : ip (defaulting) -> low
.Feb 28 22:29:00: PQ: Serial0/0 dlci 503 : ip (defaulting) -> low
.Feb 28 22:29:00: PQ: Serial0/0 dlci 503 : ip (defaulting) -> low
.Feb 28 22:29:00: PQ: Serial0/0 dlci 503 : ip (defaulting) -> low
.Feb 28 22:29:02: PQ: Serial0/0 dlci 503 : ip (defaulting) -> low
.Feb 28 22:29:02: PQ: Serial0/0 dlci 503 : ip (defaulting) -> low
.Feb 28 22:29:02: PQ: Serial0/0 dlci 503 : ip (defaulting) -> low
.Feb 28 22:29:02: PQ: Serial0/0 dlci 503 : ip (defaulting) -> low
.Feb 28 22:29:02: PQ: Serial0/0 dlci 503 : ip (defaulting) -> low
.Feb 28 22:29:02: PQ: Serial0/0 dlci 503 : ip (defaulting) -> low
.Feb 28 22:29:02: PQ: Serial0/0 dlci 503 : ip (defaulting) -> low
.Feb 28 22:29:02: PQ: Serial0/0 dlci 503 : ip (defaulting) -> low
.Feb 28 22:29:02: PQ: Serial0/0 dlci 503 : ip (defaulting) -> low
.Feb 28 22:29:02: PQ: Serial0/0 dlci 503 : ip (defaulting) -> low
.Feb 28 22:29:02: PQ: Serial0/0 dlci 503 : ip (defaulting) -> low
.Feb 28 22:29:02: PQ: Serial0/0 dlci 503 : ip (defaulting) -> low
.Feb 28 22:29:02: PQ: Serial0/0 dlci 503 : ip (defaulting) -> low
.Feb 28 22:29:02: PQ: Serial0/0 dlci 503 : ip (defaulting) -> low
.Feb 28 22:29:02: PQ: Serial0/0 dlci 503 : ip (defaulting) -> low
.Feb 28 22:29:03: PQ: Serial0/0 dlci 503 : ip (defaulting) -> low
.Feb 28 22:29:03: PQ: Serial0/0 dlci 503 : ip (defaulting) -> low
.Feb 28 22:29:03: PQ: Serial0/0 dlci 503 : ip (defaulting) -> low
.Feb 28 22:29:03: PQ: Serial0/0 dlci 503 : ip (defaulting) -> low
.Feb 28 22:29:03: PQ: Serial0/0 dlci 503 : ip (defaulting) -> low
.Feb 28 22:29:03: PQ: Serial0/0 dlci 503 : ip (defaulting) -> low
.Feb 28 22:29:03: PQ: Serial0/0 dlci 503 : ip (defaulting) -> low
.Feb 28 22:29:03: PQ: Serial0/0 dlci 503 : ip (defaulting) -> low
.Feb 28 22:29:03: PQ: Serial0/0 dlci 503 : ip (defaulting) -> low
.Feb 28 22:29:03: PQ: Serial0/0 dlci 503 : ip (defaulting) -> low
.Feb 28 22:29:04: PQ: Serial0/0 dlci 503 : ip (defaulting) -> low --
congestion drop
.Feb 28 22:29:04: PQ: Serial0/0 dlci 503 : ip (defaulting) -> low --
congestion drop
.Feb 28 22:29:04: PQ: Serial0/0 dlci 503 : ip (defaulting) -> low --
congestion drop
.Feb 28 22:29:04: PQ: Serial0/0 dlci 503 : ip (defaulting) -> low --
congestion drop
Cordialmente,
------------------------------------------------------------------
Alaerte Gladston Vidali
IBM Global Services - SO
Tel.55+11+2121-2879 Fax:55+11+2121-2449
"Chris Lewis \(chrlewis\)" <chrlewis@cisco.com>
27/07/2005 14:45
To
Alaerte Gladston Vidali/Brazil/IBM@IBMBR, <ccielab@groupstudy.com>
cc
Subject
RE: Priority under Shaping
Gladston,
Hopefully some background info will help in understanding the results
you see.
Originally, on 7200 platforms and below due to the fact that the shaper
token bucket was only refilled every 10ms (hence a PQ packet could be
delayed 10ms by the shaper alone), the PQ bypassed the shaper and you
would not see PQ packets delayed by the shaped queue. On 7500 the shaper
refresh was every 4ms and the PQ never bypassed the shaper. A bit later
on, QoS code converged on the 7500 code base and the 7500 behaviour
became universal.
The net result is that with the 7500 behavior being universal,
interfaces on 7200 and below will no longer exceed the shaped rate if PQ
traffic is transmitted (as was the case with the old 7200 and below
behavior), but they can experience some delay as a result of shaping.
Chris
-----Original Message-----
From: nobody@groupstudy.com [mailto:nobody@groupstudy.com] On Behalf Of
gladston@br.ibm.com
Sent: Tuesday, July 26, 2005 2:26 PM
To: ccielab@groupstudy.com
Subject: Priority under Shaping
Hi,
Am I wrong expecting a better round-trip time for traffic that receives
priority on the queue formed by Shaping?
I could not notice it on tests. Tested CB Shaping and FRTS.
Whithout traffic, the round-trip average is 5ms for priority traffic.
When the link is saturated using SAA (jitter), the round-trip average is
109ms and there is no way to get a small time (on the test I did).
I tried to get a better time using these:
-Priority under CB Shaping
-LLQ under map-class frame-relay
-PQ under map-class frame-relay
With CB priority and LLQ I could check that priority packets are
matched.
These are the times:
Type escape sequence to abort.
Sending 5, 100-byte ICMP Echos to 148.5.57.7, timeout is 2 seconds:
!!!!!
Success rate is 100 percent (5/5), round-trip min/avg/max = 48/109/132
ms Rack2R2(config-subif)#do pin 148.5.57.7
Type escape sequence to abort.
Sending 5, 100-byte ICMP Echos to 148.5.57.7, timeout is 2 seconds:
!!!!!
Success rate is 100 percent (5/5), round-trip min/avg/max = 96/117/129
ms Rack2R2(config-subif)#do pin 148.5.57.7
Type escape sequence to abort.
Sending 5, 100-byte ICMP Echos to 148.5.57.7, timeout is 2 seconds:
!!!!!
Success rate is 100 percent (5/5), round-trip min/avg/max = 4/5/8 ms
Rack2R2(config-subif)#do pin 148.5.57.7
Type escape sequence to abort.
Sending 5, 100-byte ICMP Echos to 148.5.57.7, timeout is 2 seconds:
!!!!!
Success rate is 100 percent (5/5), round-trip min/avg/max = 4/5/8 ms
Rack2R2(config-subif)#do pin 148.5.57.7
These are the configs:
For CB Shaping
class-map match-all DLSW-priority
match access-group 114
!
policy-map Child-priority-dlsw
class DLSW-priority
priority 64
policy-map CB-Shaping-and-priority
class class-default
shape average 128000 16000 1000
service-policy Child-priority-dlsw
!
int ser 0/0.235
service-police output CB-Shaping-and-priority
For FRTS
class-map match-all Priority-under-Shaping
match access-group 132
!
policy-map Priority-under-Shaping
class Priority-under-Shaping
priority 64
!
int ser 0/0
frame-relay traffic-shaping
frame interface-dl 502
class Shaping-to-r2
!
map-class fra Shaping-to-r2
service-policy out Priority-under-Shaping
Monitoring for CB Shaping:
Rack2R2#sh pol int ser 0/0.235
Serial0/0.235
Service-policy output: CB-Shaping-and-priority
Class-map: class-default (match-any)
1007047 packets, 64489391 bytes
5 minute offered rate 188000 bps, drop rate 59000 bps
Match: any
Traffic Shaping
Target/Average Byte Sustain Excess Interval
Increment
Rate Limit bits/int bits/int (ms) (bytes)
128000/128000 2125 16000 1000 125 2000
Adapt Queue Packets Bytes Packets Bytes Shaping
Active Depth Delayed Delayed Active
- 0 718864 46045640 682359 43687588 yes
Service-policy : Child-priority-dlsw
Class-map: DLSW-priority (match-all)
755 packets, 49824 bytes
5 minute offered rate 0 bps, drop rate 0 bps
Match: access-group 114
Queueing
Strict Priority
Output Queue: Conversation 24
Bandwidth 64 (kbps) Burst 1600 (Bytes)
(pkts matched/bytes matched) 461/36268
(total drops/bytes drops) 0/0
Class-map: class-default (match-any)
738922 packets, 47304513 bytes
5 minute offered rate 187000 bps, drop rate 59000 bps
Match: any
Monitoring for FRTS:
Rack2R5(config-if)#do sh fram pvc 502
PVC Statistics for interface Serial0/0 (Frame Relay DTE)
DLCI = 502, DLCI USAGE = LOCAL, PVC STATUS = ACTIVE, INTERFACE =
Serial0/0
input pkts 873443 output pkts 889948 in bytes 55946844
out bytes 57035216 dropped pkts 968 in pkts dropped 0
out pkts dropped 968 out bytes dropped 62102
late-dropped out pkts 968 late-dropped out bytes 62102
in FECN pkts 0 in BECN pkts 0 out FECN pkts 0
out BECN pkts 0 in DE pkts 0 out DE pkts 0
out bcast pkts 500 out bcast bytes 72475
5 minute input rate 79000 bits/sec, 155 packets/sec
5 minute output rate 87000 bits/sec, 182 packets/sec
pvc create time 02:41:56, last time pvc status changed 02:41:33
cir 128000 bc 128000 be 0 byte limit 2000 interval 125
mincir 64000 byte increment 2000 Adaptive Shaping none
pkts 16 bytes 1450 pkts delayed 0 bytes delayed 0
shaping inactive
traffic shaping drops 0
service policy Priority-under-Shaping
Serial0/0: DLCI 502 -
Service-policy output: Priority-under-Shaping
Class-map: Priority-under-Shaping (match-all)
12 packets, 1188 bytes
5 minute offered rate 0 bps, drop rate 0 bps
Match: access-group 132
Queueing
Strict Priority
Output Queue: Conversation 24
Bandwidth 64 (kbps) Burst 1600 (Bytes)
(pkts matched/bytes matched) 0/0
(total drops/bytes drops) 0/0
Class-map: class-default (match-any)
5 packets, 320 bytes
5 minute offered rate 0 bps, drop rate 0 bps
Match: any
Output queue size 0/max total 600/drops 0
This archive was generated by hypermail 2.1.4 : Sun Sep 04 2005 - 17:00:31 GMT-3