Is it a long road to get everybody to recognize that there is indeed an issue. Rogers is at the forefront of this investigation having collected data on this issue for multiple months now and this is now proving really useful in demonstrating that the issue goes beyond simple ICMP improvement.
There is no doubt that the recent development in multiple forums has put more pressure on the various parties and I am hoping that Hitron in collaboration with Intel will be able to fix these issues quickly.
I unfortunately have no confirmation of if and when such fix will be ready but as soon as it is available to me, it will be part of the firmware trial program and all participants will receive it.
Also, we are counting on the new modem to be released in the coming weeks and that seems extremely promising.
Dave
@jalalinator wrote:Hey DataLink / Rogers Dave, please explain these pathetic speeds / latency on the new google speedtest, peered to a Toronto Data Center..... it's amazing how speedtest.net / dslreports are sooooo skewed in the tests, could it be that rogers is filtering / prioritizing / SHAPING traffic to skew the results? Because that is exactly what it looks like.
Jalalinator,
I will need to perform further testing on this specific case. I can assure you that we are not doing any form of shaping of prioritization of speedtests.net / dslreports. In fact, with over 9000 servers around the world hosted by various ISP for speedtest.net, it would be impossible to do so even if we wanted. It would also contravene net neutrality rules.
I ran some additional tests on an Internet connection where I generally get results in the 700-900 Mbps using Google speedtest and was getting results below 100 Mbps. I know for a fact that the 700 Mbps is real as I test this connection over and over, using various test sites but also simple file downloads to characterize the performance of various modems and firmware.
There is an issue in the way Google performs speedtests and I'll have to investigate further to understand where the limitation exists. It may be realted to the use of HTML5 or the location of the server but something is definitely wrong.
Dave
Locations and what network, can make night and day results sometimes.
Testing to the Toronto set of servers from where i am.. i can get anything from above my package speeds, to about 30mbps less.
Why such varriations, even though they are all in Toronto?
One may be on or just off the rogers backbone there.. other ones, you may have to go from Rogers, to bell, to another, etc.
All those extra jumps, could make a difference, each extra hop possibly slowing things down.
It gets worse, when you get to speed tests, which dont give you the option of your end test location.
The google one for example is like that.
(EG: At work here, are on Bell Business Fiber into the building. Its 150/150. On bells speedtest, or even rogers, i show pretty much 150/150 or pretty close. Run the google one? 40/3 :P)
@Hybrid_Noodle wrote:Hey @RogersDave any news on what type/model modem that is expecting to be released or is that all still hush hush?
I would make the assumption that it would be the Hitron CODA, or CODA-45? You can just give a simple cough once for yes and twice for no 😉
LOL, I have never seen a CODA-45 device.
@JohnBeaudin wrote:
Can you give us a hint.. is there an intel puma7 chipset in there? lol
I can't unfortunately answer that just yet but there should be an official answer fairly soon.
Nope, not at all. What are you running, wired or wireless? Beyond the modem, out to the next IP address which is the CMTS, you could see packet loss if there is an issue with the external cables or connectors.
-
917394 views
-
669 Likes
-
-
