@Eniex you're signal levels are low, but they are still in spec. Rogers uses -10 dBmV as a cutoff point on the downstream before they will do anything about it. Your pingplot to google shows the typical latency with a Puma 6MG modem. The hop that you really want to look at is hop #2 which is the CMTS. Select that hop to bring up the plot to the CMTS and right click on the google plot and select hide. That plot to the CMTS will show what affects all traffic to any destination.
You could request V4.5.8.27 which will take care of the IVP4 ICMP (ping) latency, but, it will limit the downstream data rate to 20 Mb/s as a side effect. For that reason, V4.5.8.27 was not distributed network wide to the CGNM-3552s.
Have a look at the following thread regarding V4.5.8.27:
You could also update to the CODA-4582 but that in itself might present technical challenges. Some users report no issues while others are reporting data rate and wifi issues. I run a CODA-4582 in Bridge mode and haven't noticed any issues so far, so it depends on what mode you would be using, Gateway or Bridge mode.
At the present time, the modem and ISP world is waiting for Intel and Arris to complete the firmware that will reduce the latency that is currently caused by the software based CPU processing of the data packets. Its been pretty quiet on that front recently, but, its the Christmas and New Years holiday time as well. So, we're all waiting.....
The two interesting threads in question for the Puma 6/6MG latency are:
http://www.dslreports.com/forum/r31122204-SB6190-Puma6-TCP-UDP-Network-Latency-Issue-Discussion
This last thread is the current discussion thread.
Speeds on .19 seem fine for me, but latency in online games is not amazing, 50-70 ms to East NA servers:
@Lumute your downstream signal levels are high, but within spec. The signal to noise is ok. The upstream signal levels are fine. The only thing that a tech would do is install an attenuator to drop the downstream levels. That would also push up the upstream levels. if you're not having any problems I would leave it alone for now. Over the next few years, without any changes, those levels will drop on their own as the external cable and connectors age.
If that's an ICMP ping to the CMTS it would have been 4.5.8.27 on the modem. V4.6.8.22 and earlier would have resulted in a much worse plot. What that plot really shows is that V4.5.8.27 on probably both CGN3ACSMR and CGNM-3552 is better than the current version of CODA-4582, ergo, my question, does the CODA-4582 use the hardware accelerator. From your plot, I would say that the answer is no. If it does, then I would say that we have yet another problem to solve.
My area has had terrible upload speeds for months, even right now (~1pm) I'm getting ~3Mbps up.
I'm on the D3.1 modem but maintainence siad there are "no problems" in my area. I've had 5 modems (3 unique types) All equipment is connected by Cat5e etc. New rg6 has been ran too, though no change.
If you need more info feel free to PM me!
@VivienM @RogersDave @Datalink I ran a speedtest on my CODA modem and see that the upload speed took a nosedive today. Did someone change my speed profile?
Wanted to add a data point.
Running 2.0.10.19 on CODA with gigabit plan, in bridge mode with ER-X router (had to enable hardware offload for NAT otherwise I was getting about 450 down, also had to remove a rogue cat-5 cable from the mix).
Everything seems to be better than I would have expected.
@Lumute you can find attenuators online, but honestly, with your signal levels where they are, its not worth doing. It won't increase your upload rates. I suspect that the issue is as you indicated, oversold capacity. The question at this point is whether or not these low rates can be resolved with DOCSIS 3.1 on the upload side alone, or, whether it will require node splits. DOCSIS 3.1 on the upload side is still weeks or months away. From what I understand its still in development.
All cable systems are built on numbers. For example, X number of homes in a neighborhood will subscribe to a given ISP, of which, only some, Y homes require service during the day, and Z homes will require service during the evening hours. I think the numbers are all estimates, however by now I suspect that cable companies have this down a science. Crunch the numbers and you come up with an estimated throughput requirement for a CMTS. I suspect that the arrival of internet TV (Netflix and others) along with much higher internet speeds are throwing some of those calculations out the window, leading to unhappy subscribers. How that can be answered is up to individuals such as @RogersDave to determine. Maybe he can shed some light on what seems to be a dilemma for a growing number of customers.
-
942972 views
-
669 Likes
-
-
