Downstream - Wideband Resiliency

[Ivan Fernando]

Dear all,

I want to ask about Cable broadband, using UBR10012.

I have try to use DS Resiliency (Using cable rf-change-trigger command with cable resiliency ds-bonding) in IOS SCG3 version and wideband resiliency in IOS SCC7 version, but i can't find a different impact when I try it.

Anyone have some experience about different of wideband resiliency and downstream resiliency. Have a scenario to find it?

I have read a document in http://www.cisco.com/en/US/partner/docs/ios/cable/configuration/guide/ubr_ds_resiliency.html#wp1054188
but, i can't simulate it, to find the different.

Regads,

[Manuel Rodriguez · Accepted solution]

Hi Ivan,

The main difference between the Wideband Modem Resiliency and the Resiliency Bonding Group features is that the first, once the trigger is reached, will disable channels that are reported as impaired. This has the potential of impacting non-affected modems for which tht channel may not be impaired. With RBG, you can define a bonding group that will act as resilient. This bonding group will be created dinamically once impairments are reported and will contain a subset of channel which satisfy the impairmnet. The CMs who reported impairment are assigned to that resilient bonding group while other keep on using the normal bonding group. This way, you don't reduce service for modems who don't need it.

Regarding how to test it, the best would be to simulate some RF impairment so your D3.0 modems would send CM-status messages to report the impairments. Otherwise you could use the "test cable cm-status" command to simulate modems sending CM-status messages. Here is a brief explanation on how to use it:

#test cable cm-status   2 01030401

"Transaction ID" should begin from 1, increase when you repeat same type testing; "2" is for testing "QAM/FEC lock failure"; 01030401 is the TLV;

For the TLV format, you can refer to D3.0 MULPI Spec "Table 6-52 -
CM-STATUS TLV Encodings" and "Table 10-3 - CM-STATUS Event Type Codes
and Status Events". 

The TLV begin with 01, the length and sub TLVs depend on the event need
be tested, refer to "Table 10-3".

Taking "QAM/FEC lock failure" for example:

In table 10-3: "QAM/FEC lock failure" need DCID in the sub TLV; Type is
4 (refer to table 6-52), length is 1 (one octet for DCID), value is
DCID;

So the sub TLV for DCID: 0401; Length of this sub tlv is 3;

So entire TLV for "QAM/FEC lock failure" is 01030401;

The DCID can be get from running config, for example:

doc-ubr10k-2#show cable modem 4458.2945.2ade wideband rcs-status 
CM  : 4458.2945.2ade
RF  : 7/0/0 1
 Status                   : UP
 FEC/QAM Failure          : 0         
 Dup FEC/QAM Failure      : 0         
 FEC/QAM Recovery         : 0         
 Dup FEC/QAM Recovery     : 0         
 MDD Failure              : 0         
 Dup MDD Failure          : 0         
 MDD Recovery             : 0         
 Dup MDD Recovery         : 0         
 Flaps                    : 0         
 Flap Duration            : 00:00
RF  : 7/0/0 2
 Status                   : UP
 FEC/QAM Failure          : 0         
 Dup FEC/QAM Failure      : 0         
 FEC/QAM Recovery         : 0         
 Dup FEC/QAM Recovery     : 0         
 MDD Failure              : 0         
 Dup MDD Failure          : 0         
 MDD Recovery             : 0         
 Dup MDD Recovery         : 0         
 Flaps                    : 0         
 Flap Duration            : 00:00

doc-ubr10k-2#

Now I want to test a "QAM/FEC lock failure" for "RF  7/0/0 1" of modem "4458.2945.2ade":

From running config:

doc-ubr10k-2#sh run | b 7/0/0
controller Integrated-Cable 7/0/0
rf-channel 0 cable downstream channel-id 97
rf-channel 0 frequency 555000000 annex B modulation 64qam interleave 32
rf-channel 0 rf-power 52.0
no rf-channel 0 rf-shutdown
rf-channel 1 cable downstream channel-id 98


DCID for 7/0/0 rf 1 is 98, hex value is 62.

So the test CLI should be:

doc-ubr10k-2#test cab cm-status 4458.2945.2ade 1 2 0103040162

After testing, we can see the rf status changed:

doc-ubr10k-2#show cable modem 4458.2945.2ade wideband rcs-status 
CM  : 4458.2945.2ade
RF  : 7/0/0 1
 Status                   : DOWN_PENDING
 FEC/QAM Failure          : 1         Mar 8  08:20:01
 Dup FEC/QAM Failure      : 0         
 FEC/QAM Recovery         : 0         
 Dup FEC/QAM Recovery     : 0         
 MDD Failure              : 0         
 Dup MDD Failure          : 0         
 MDD Recovery             : 0         
 Dup MDD Recovery         : 0         
 Flaps                    : 0         
 Flap Duration            : 00:01
RF  : 7/0/0 2
 Status                   : UP
 FEC/QAM Failure          : 0         
 Dup FEC/QAM Failure      : 0         
 FEC/QAM Recovery         : 0         
 Dup FEC/QAM Recovery     : 0         
 MDD Failure              : 0         
 Dup MDD Failure          : 0         
 MDD Recovery             : 0         
 Dup MDD Recovery         : 0         
 Flaps                    : 0         
 Flap Duration            : 00:00

doc-ubr10k-2#


And finally the modem went to p-online (which is expected since there is no resiliency configured for this example):



doc-ubr10k-2#show cable modem 4458.2945.2ade            
D
MAC Address    IP Address     I/F           MAC           Prim RxPwr Timing Num I
                                           State         Sid  (dBmv) Offset CPE P

4458.2945.2ade 40.4.58.18     C7/0/0/UB     p-online(pt)  21   0.00   1211   0   N

doc-ubr10k-2#

Hope this helps.

Best regards.

[Ivan Fernando]

Hi Manuel,

thank you for your answer. I'm glad to read it. It's very helpfully.

Once more, I want to ask about the triggered value. The count is means cable modem that will be report impairment, right?

how about percentage? because it set in global configuration.

The RBG works if it is below of threshold. So, any problem if I set with maximum value? Any consideration?

I can't get the different performance between wideband resilliency and downstream resilliency, when i use ping test in client side.  

thank you for your help.

regards,

[Manuel Rodriguez · Accepted solution]

Hi Ivan,

Once more, I want to ask about the triggered value. The count is means cable modem that will be report impairment, right?

Correct

how about percentage? because it set in global configuration.

The percentage option keeps track of the percentage of modems reporting impairments on each DS channel in the CMTS.

The RBG works if it is below of threshold. So, any problem if I set with maximum value? Any consideration?

The consideration would be that, if you have a channel which is imparied from the headend, it would be better to simply let the WB resiliency to kick in and bring the channel down. That way, any possible impact is simply avoided until the channel is recovered. If you rely only on RBG, you may have situation where, for example, a modem doesn't report the impairment but the channel is indeed impaired. In that case, the modem will have degraded service. If the channel is down from the CMTS by WB resiliency, this would not happen.

I can't get the different performance between wideband resilliency and  downstream resilliency, when i use ping test in client side.   

The difference is that, when below threshold, WB resiliency will move service flows (primary only or secondary as well, depending on your config) from the bonding group to the primary channel while RBG will move the service flows to the RBG intead than moving them to the primary channel. This way, a better performance can still be provided for the affected modems. I don't think you will notice any difference with a ping test since the difference will be more on a throughput performance side.

Best regards.