I'm glad you offered to go in detail I'm specifically looking if you can go into more detail on how OFDM multi-profile support works to improve stability (i.e what are these various "thresholds" set in the modem?), and what happens when the modem encounters an impairment and automatically adjusts? What consists inside a separate profile?
Also, if you would like you can also go into how 3.1 works and how it is implemented at Rogers (Feel free to PM me the response if you like )
Count me in on this discussion, especially in the area of "what happens when the modem encounters an impairment". That's where we typically become involved, when a customer posts a problem report.
I'd like to see better reporting for the OFDM channel so that we can assess the performance of that channel. That's fairly easy to do with the DOCSIS 3.0 channels, impossible to do with the 3.1 OFDM channel at the present time.
Lol, what a bunch of night owls....
If you don't mind a bit of light reading i recommend checking this pdf out. It goes through some of the features of docsis 3.1 and how they work in depth.
Is active queue management already implemented? I see it being talked about as a major stepping stone for docsis 3.1 for improving gaming and overall latency sensitive applications.
Looking at the D3.1 DS profile on the PDF and it makes sense. The modem can switch to different profiles depending on it's SNR for the best performance. This means that modems can be using up to 4K QAM (if Rogers allows it ). Wouldn't this allow for more bandwidth?
Also, cablelabs has mandated AQM in the 3.1 spec (it required it to be present at the CMTS and CPE level). So it should be implemented.
I'm connected with rfog which could have some effect on it? Also my area is fairly new construction so could have something to do with newer hardware? Are you seeing some sort of problem related to TCP packets?
Is the update for AC Modems delayed also? Did not get my update yet, was it not suppose to be done by Thursday?
@xenoguy that Puma 6 test is highly dependent on the distance to, and reaction time from the speedtest server as well as the load on the CMTS, which your modem is connected to. Personally I don't put a lot of faith in that test, primarily as I'm located in Ottawa and the nearest Speedtest servers for that test are located in Toronto or Montreal. That's not an ideal situation for either server. While a test result late at night or very early in the morning can show a pretty good result, running a test throughout the day can show a reasonable to good result one minute, and a fair to middling test the next minute.
To see reasonable results, all of your return times have to be under 50 milli-seconds. If you get anything above that, the visual result will look fair to poor.
Personal opinion, to really test the latency thru any modem, you need to run and plot your own ping tests, which include:
1. IPV4 and IPV6 ICMP and TCP/IP ping test to the modem
2. IPV4 and IPV6 ICMP ping to the CMTS. With the Puma 7 modems, thats not possible due to a high return time, which is a modem firmware issue introduced in version .27. Not sure about the Puma 6 modems at the present time.
3. IPV4 and IPV6 ICMP, TCP/IP and UDP ping tests to the DNS, both Rogers and a chosen external DNS to really see the difference between internal ISP and external ISP results.