CODA-4582 - Open Issues for Investigation

Need Help?

That's what we're here for! The goal of the Rogers Community is to help you find answers on everything Rogers. Can't find what you're looking for? Just ask!
cancel
Showing results for 
Search instead for 
Did you mean: 
Reply
I Plan to Stick Around
Posts: 154

Re: CODA-4582 - Open Issues for Investigation


@RyzenFX wrote:

This bufferbloat can be attributable to the speedboost feature on all the Ignite plans EXCEPT for Ignite Gigabit. It makes a lot of sense. For instance, @4kgigabitguy is on the Ignite Gigabit where  speedboost is not enabled and gets 0.9 packets lost/minute in League of Legends. In my case, I'm receiving 6.7 packets lost/min on the 100u plan. If you read here, the Casa Systems CMTS could also be at the source of the problem. In Comcast D3.1 land, they aren't experiencing any bufferbloat issues which makes sense, because they're on a Cisco CMTS (or because they don't have a speedboost feature). So the problem has to be attributable to the speedboost feature + the CMTS.

 


I thought there were a lot of people here on the gigabit plan also complaining about UDP and packet loss issues. I don't think it's limited to just those below gigabit.

 

Occam's razor here -- Cisco equipment is probably a LOT better...

I Plan to Stick Around
Posts: 154

Re: CODA-4582 - Open Issues for Investigation


@RyzenFX wrote:

@4kgigabitguy Thanks for replying. Your upstream signal strength is much stronger than mine, and when looking back at your speedtests (specifically the 2015 macbook air) on the Rogers speedtest site, your upload graph looks really similar to my upload graph on the CGNM-3552. Also, your bufferbloat looks good on the upload and download, and ideally that is what I should be getting. I'm on .24, but hopefully I'll get upgraded to .26 tomorrow (sent a PM to Dave requesting upgrade). I'll let you know how things go.

Didn't you say that you're also on D3.0?  You're comparing your results to D3.1 -- don't think that's really fair.

Resident Expert
Resident Expert
Posts: 1,066

Re: CODA-4582 - Open Issues for Investigation


Telek wrote:

 

Occam's razor here -- Cisco equipment is probably a LOT better...


Cisco hardware is definitely better, but it costs a lot more. Regardless they just bought all this CASA hardware so we're stuck with it for a few years until they switch to FTTH



I'm a Reliable Contributor
Posts: 577

Re: CODA-4582 - Open Issues for Investigation

@Telek I already mentioned that I'm on D3.1 downstream and D3.0 upstream- same configuration as @4kgigabitguy, and ideally I should be seeing similar results to him, but that's not the case. 

 

(I'm comparing D3.0 upstream results from the 3552 to the D3.0 upstream results on the CODA-4582; it's a fair comparison. I should be seeing really similar upload graphs between the town modems, but once again it's not the case.) 

 

 The buffer bloat on the CODA-4582 is much worse than 3552, as pointed out here. I've stated that the bufferbloat could be the reason for all the high packet loss in League of Legends, and random ping spikes in other games. I'm not sure if this is attributable to the Puma 7 chipset, but I'm confident that the Speedboost feature and Casa Systems CMTS is playing a big role in this problem.

I Plan to Stick Around
Posts: 154

Re: CODA-4582 - Open Issues for Investigation


@RyzenFX wrote:

@Telek I already mentioned that I'm on D3.1 downstream and D3.0 upstream- same configuration as @4kgigabitguy, and ideally I should be seeing similar results to him, but that's not the case. 

 

(I'm comparing D3.0 upstream results from the 3552 to the D3.0 upstream results on the CODA-4582; it's a fair comparison. I should be seeing really similar upload graphs between the town modems, but once again it's not the case.) 

 

 


Gotcha -- sorry, my bad.

 


@RyzenFX wrote:

The buffer bloat on the CODA-4582 is much worse than 3552, as pointed out here. I've stated that the bufferbloat could be the reason for all the high packet loss in League of Legends, and random ping spikes in other games. I'm not sure if this is attributable to the Puma 7 chipset, but I'm confident that the Speedboost feature and Casa Systems CMTS is playing a big role in this problem.


Sorry, I'm trying to follow. Since we see people on pure gigabit that also have the same problem, without speedboost, what's leading you to believe that it is a major factor?

I'm a Reliable Contributor
Posts: 577

Re: CODA-4582 - Open Issues for Investigation

No problem. I'm thinking that the cause for the bufferbloat issue is because of the Speedboost feature and the CMTS. The bufferbloat could be the exact reason why I'm dropping packets.

Resident Expert
Resident Expert
Posts: 6,012

Re: CODA-4582 - Open Issues for Investigation

Somewhere in all of these threads is a discussion on Bufferbloat as calculated by the DSLReports speedtest, and it has to do with the initial speeds versus the speeds as seen during the remainder of the test.  The result is that the slower speeds as seen beyond the initial rise is considered as bufferbloat by the test and results in an inaccurate calculation.  I think in that discussion Dave pointed out that you have to set the speed for the service that you're on, which should result in a more accurate calculation.  Fwiw, my personal opinion is that the DSLReports servers are, for the most part, too far away to result in a useful indication of the last mile capability, that is to say, between the CMTS and the modem, especially for gigabit service.  There are only two local servers, Toronto and Montreal, every other server is fairly distant, and the server speeds in many cases are low, compared to the Rogers, Beanfield and Fibrenoire servers in Toronto and Montreal.  I found that testing somewhere between midnight to 8:00 am is the only time where you might see a useful result.  Outside of those times, you might as well not bother, and that includes both speedtests and the mythical "bufferbloat".  

 

Also fwiw, I see more random packet loss with the CODA-4582 than I have with other modems which I believe is a little unusual.  That is seen during pingtests to the CMTS. 



I'm a Reliable Contributor
Posts: 603

Re: CODA-4582 - Open Issues for Investigation

@gp-se

 

I think we are still several years away from FTTH.. ( 5years+ if not more.. ) 

 

With D3.1 full duplex I don't see any reason to start rolling FTTH any sooner.

Highlighted
I Plan to Stick Around
Posts: 26

Re: CODA-4582 - Open Issues for Investigation

  I can't add much to what actual problem with 4582 and UDP might be, but i can say, since we switched to 3552 i have had zero issues with browsing and playing games. 

 

 

  Like RyzenFX has noted, the speed test on the 3552 are usually steady, on the CODA was always spikes, would hit same speeds but very inconsistent. Bufferbloat was always fail on the CODA, still some on the 3552 but much better.

 

 

Network Architect
Network Architect
Posts: 611

Re: CODA-4582 - Open Issues for Investigation


@RyzenFX wrote:

@4kgigabitguy @RogersDave @Datalink I just switched my CGNM-3552 back to the CODA-4582. I ran some test games of LoL, however it's packet loss is nowhere near the packet loss I averaged with the 3552; it's worse. I averaged  6.7 packets lost/min. However, I do have a suspicion that the CMTS is not communicating with the modem properly (i.e CMTS is not giving the modem enough upstream signal strength, which needs to be adjusted). 

 

Upstream Signal Strength:

Upstream Overview

Port IDFrequency (MHz)ModulationSignal strength (dBmV)Channel IDBandwidth
130596000ATDMA - 64QAM29.25016400000
238595785ATDMA - 64QAM33.25033200000
323700000ATDMA - 64QAM29.25026400000

 

 


The uplink signal strengh on the CODA is measured in a completely different way compared to previous devices and this is due to a change in the DOCSIS specification. The power level now takes into account the bandwidth of the channel.

 

For a 6.4 MHz channel, you signal level will be about 6 dB lower than what would be reported by a D3.0 modem. For a 3.2 MHz channel, the power level will be about 3 dB lower than a D3.0 modem.

 

In the case above, this means that if you were to connect a D3.0 modem in the exact same conditions, the power levels would be 35, 36, 35 wich is perfectly fine. The modem is also instructed by the CMTS how to regulate power level so that the signal arrives at the CMTS with a value close to 0. I checked your modem and all three uplink channels are received at the CMTS at 1 dB which is perfect.

 

Dave