IGX frame relay connection limitation

[sysadmin]

I am running a frame relay connection back-to-back on a IGX between two ports on the same UFMU card. The CIR is 512 kbps and the PIR is 1280 kbps, I have Cisco 7304/7206 VXR connected on the two ends. Regardless if traffic shaping is enabled or not on the router, I cannot peak over approximately 675 kbps. Any ideas?

Jens

[mheusinger]

Hi,

what are the port and connection settings? Is ForeSight enabled? Can you post a "dspcon" and "dspportcnf"?

Regards, Martin

[sysadmin]

No foresight

XXXXX_26 TN Jens:1 IGX 8420 9.3.11 Aug. 8 2006 13:37 GMT

Conn: 10.11.150 XXXXX_26 10.12.150 fr Status:OK

MIR CIR VC Q Depth PIR Cmax ECN QThresh QIR

512/512 512/512 5334/5334 1280/1280 75/75 3100/3100 512/512

Pri: L Test-RTD: 0 msec FST: n % Util: 100/100

Path: Route information not applicable for local connections

XXXXX_26 UFMU: OK XXXXX_26 UFMU: OK

UFI: OK UFI: OK

Last Command: dspcon 10.11.150

XXXXX_26 TN Jens:1 IGX 8420 9.3.11 Aug. 8 2006 13:38 GMT

Port: 10.11 [ACTIVE ]

Interface: V35 DCE Configured Clock: 1536 Kbps

Clocking: Normal Measured Rx Clock: 1536 Kbps

Port ID 150 Min Flags / Frames 1

Port Queue Depth 65535 OAM Pkt Threshold 3 pkts

ECN Queue Threshold 65535 T391 Link Intg Timer 0 sec

DE Threshold 100 % N391 Full Status Poll 0 cyl

Signalling Protocol Cisco LMI EFCI Mapping Enabled Yes

Asynchronous Status No CLLM Enabled/Tx Timer Yes/100 msec

T392 Polling Verif Timer 15 IDE to DE Mapping Yes

N392 Error Threshold 3 Interface Control Template

N393 Monitored Events Count 4 Lead CTS DSR DCD

Communicate Priority No State ON ON ON

Upper/Lower RNR Thresh 75%/ 25%

Last Command: dspport 10.11

XXXXX_26 TN Jens:1 IGX 8420 9.3.11 Aug. 8 2006 13:38 GMT

Port: 10.12 [ACTIVE ]

Interface: V35 DCE Configured Clock: 1536 Kbps

Clocking: Normal Measured Rx Clock: 1536 Kbps

Port ID 151 Min Flags / Frames 1

Port Queue Depth 65535 OAM Pkt Threshold 3 pkts

ECN Queue Threshold 65535 T391 Link Intg Timer 0 sec

DE Threshold 100 % N391 Full Status Poll 0 cyl

Signalling Protocol Cisco LMI EFCI Mapping Enabled Yes

Asynchronous Status No CLLM Enabled/Tx Timer Yes/100 msec

T392 Polling Verif Timer 15 IDE to DE Mapping Yes

N392 Error Threshold 3 Interface Control Template

N393 Monitored Events Count 4 Lead CTS DSR DCD

Communicate Priority No State ON ON ON

Upper/Lower RNR Thresh 75%/ 25%

Last Command: dspport 10.12

[mheusinger · Accepted solution]

Hi,

enable ForeSight (!) or increase CIR or VC Q depth.

The reason is: without FS you will be limited to CIR (plus a little bursting influenced by the VC Q setting).

You should be able to see dropped frames with dspconstat, if you burst too much.

ForeSight will allow to increase throughput up to port speed/PIR.

VC Q depth is a measure for Bc/Be defining how much you can burst.

Hope this helps! Please rate all posts.

Regards, Martin

[sysadmin]

Yes it indeed did. Thanks, now it works in the lab... So lets see if we get something useful running in production!

[sysadmin]

Actually I cried "solved" to early. Switching to ForeSight solved the problem when transferring data from the 7304 to the 7206VXR, not when transferring from the 7206VXR to the 7304. Here I am limited to EXACTLY CIR. It does not go over CIR setting on the router, while it does when I transfer data from the 7304 to the 7206VXR with a similar config and the same IOS.

Any ideas?

Output from show frame-relay pvc is interesting:

From 7206VXR (notice the "delayed packets, shaping active and output queue"):

7206#show frame-relay pvc 150

PVC Statistics for interface Serial1/1 (Frame Relay DTE)

DLCI = 150, DLCI USAGE = LOCAL, PVC STATUS = ACTIVE, INTERFACE = Serial1/1.1

input pkts 4269 output pkts 8300 in bytes 194726

out bytes 12170381 dropped pkts 9 in pkts dropped 0

out pkts dropped 9 out bytes dropped 12436

late-dropped out pkts 9 late-dropped out bytes 12436

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 8 out bcast bytes 2621

pvc create time 00:07:46, last time pvc status changed 00:07:46

cir 512000 bc 51200 be 76800 byte limit 16000 interval 100

mincir 256000 byte increment 6400 Adaptive Shaping none

pkts 8320 bytes 12201919 pkts delayed 8219 bytes delayed 12167924

shaping active

traffic shaping drops 0

Queueing strategy: fifo

Output queue 20/40, 11 drop, 8219 dequeued

while from 7304 it looks like this:

7304#show frame-relay pvc 150

PVC Statistics for interface Serial4/2 (Frame Relay DTE)

DLCI = 150, DLCI USAGE = LOCAL, PVC STATUS = ACTIVE, INTERFACE = Serial4/2.1

input pkts 3816 output pkts 6400 in bytes 194570

out bytes 9573275 dropped pkts 0 in pkts dropped 0

out pkts dropped 0 out bytes dropped 0

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 7 out bcast bytes 2254

pvc create time 1w2d, last time pvc status changed 00:14:26

cir 512000 bc 51200 be 76800 byte limit 16000 interval 100

mincir 256000 byte increment 6400 Adaptive Shaping none

pkts 7 bytes 2254 pkts delayed 0 bytes delayed 0

shaping inactive

traffic shaping drops 0

Queueing strategy: fifo

Output queue 0/40, 0 drop, 0 dequeued

[mheusinger]

Hi,

does the router config show something like "frame-relay traffic shaping" and further down "frame-relay cir ..."? Then the router will do traffic shaping to the configured CIR.

Remove the "frame-relay traffic-shaping" command and check throughput again.

To be sure, please post the relevant interface config of the router.

Hope this helps! PLease rate all posts.

Regards, Martin

[sysadmin]

Yes shaping is enabled, but I thought it is still possible to go over CIR as long as there is space on the link, just that it isn't "committed" (but Cisco docs are not consistent here). How else would make a "zero CIR service" make sense.

And on the 7304, it is possible to send more than CIR, just not on the 7206VXR.

Config of 7304 (can peak to 1280 kbps):

interface Serial4/2

no ip address

encapsulation frame-relay

load-interval 30

no fair-queue

frame-relay traffic-shaping

frame-relay lmi-type cisco

no clns route-cache

!

interface Serial4/2.1 point-to-point

ip address 10.6.6.5 255.255.255.252

frame-relay interface-dlci 150

class Intranet2

!

map-class frame-relay Intranet2

frame-relay cir 512000

frame-relay bc 51200

frame-relay be 76800

frame-relay mincir 256000

no frame-relay adaptive-shaping

Config of 7206VXR (cannot peak over 512 kbps):

interface Serial1/1

no ip address

encapsulation frame-relay

load-interval 30

no fair-queue

serial restart-delay 0

frame-relay traffic-shaping

frame-relay lmi-type cisco

no clns route-cache

!

interface Serial1/1.1 point-to-point

ip address 10.6.6.6 255.255.255.252

frame-relay interface-dlci 150

class Intranet

!

map-class frame-relay Intranet

frame-relay cir 512000

frame-relay bc 51200

frame-relay be 76800

frame-relay mincir 256000

no frame-relay adaptive-shaping

!

IOS is 12.2(25)S8 and 12.2(25)S9

So at least there is an inconsistency, because the 7304 CAN burst.

Thanks for your help,

Jens

[mheusinger]

Hm.

And your frame-relay connection settings and FR port settings in the IGX are symmetric? You could have different port speeds or different CIR/PIR settings for transmit and receive. A dspcon ... should reveal the settings.

Just for testing purposes you could remove the FR traffic shaping from the routers. If the limit is still present, then the routers are not the culprits.

Regards, Martin

[sysadmin]

When I remove it, it bursts. So the router is the culprit.

So why does it burst with the statement on the 7304 and not on the 7206VXR?

CEF is enabled on both routers on the Ethernet as well as Serial ports.

[sysadmin]

Any ideas why it cannot burst on the 7206VXR, but on the 7304, even with same IOS when I enable traffic shaping?

I would like to keep it enabled, so the Cisco doesn't even try to send packets at wire speed, and maybe even marks the right packets as DE. Any ideas?