As has been mentioned before in this thread.. tests are hard to compare from one to another.. and will never be a 1:1
Why? Distance, location, the amount of network hops between, etc.
The rogers test is taken as the benchmark, at least to see if the modem is working fine and getting enough POSSIBLE throughput. That the end test point is there, without having to go out to the internet elsewhere, it removes any of the extra hops, etc.. that could slow it down.
As you start doing other tests.. you are then jumping from one carriers network to another possibly.. or even 100 different ones, depending on on where you are testing to. Even 3 tests points in one city, can yeild different results, based on how the rogers network connects to them.
And the further away you go, the more hops you are adding.. each of them will slow things down slightly, add a little more latency, etc.
Realistically, from a single sourse out on the internet.. are you ever likely to max out your connection (either up or down)? Not likely.. just to to the slight slowdowns as a whole, depending on where that endpoint is. Also dependant on what the other side can FEED you. Great you can download 100mbps, but if the other side cant feed it to you at that speed.
Havent really tested/seen since i moved to the 100u internet... but i do know its possible to max a connection.
ON the 60mbps package prior, on a torrent, i was pulling 7.8MBps (Mega BYTES per second), which equates to 62mbps. But thats comming from multiple sources.
Why is there a large dicrepency between the speed test from the rogers website as opposed to many differerent indepentdant speed tests? Rogers website always reports much higer speeds. somtimes as much as 5 times higer. Even if you test directely from the OOkla website and compare the test through the rogers website (i think they use the same ) the speed have large discrepancies. ?????
The descrepancies all have to do with the ENDPOINT. Where the speedtest is testing to.
The rogers speedtest, is a good overall generally good test, to see then if the hardware between your house, the street, etc and rogers is good. The endpoint is within rogers.. so your never going really out to the internet proper. But it will show, if your capable of getting up to those speeds.
Speedtest.net.. and many others.. give you the option of one or many specific endpoints to test with. These are all out there on the internet.
Every endpoint can be a varring different routes/interconnects from rogers. One could be rogers and then right to them. Others could be rogers, to bell, to someone esle, then to that endpoint.
The more nodes you put in, the more potential jumps, the more likely is that you may see some speed loss.
I live in Georgina (SE side of lake simcoe). CLOSEST wise, it says for me to do the barrie test spot. But its much slower, why? It goes from me, down to toronto, then back up to barrie. I am better to test toronto. KW is much the same. The KW servers, to test against them, goes back to toronto then back to them, due to the way that rogers connects to the networks those guys are on.
If your in the toronto area, the best server to test with is the TELUS one, it tends to give pretty close to the rogers test results. (though now that I look, i dont see it anymore? Prob the best now would be the Tekksavy one)
100u internet plan
Speedtest Tekksavy Toronto Results:
Is that measured via wifi or ethernet? If you used wifi, can you run the same test via ethernet?
I agree with @Gdkitty, "The descrepancies all have to do with the ENDPOINT. Where the speedtest is testing to."
A speedtest should prove that the last mile, ie, the copper cabling from the neighborhood node can support the data rate that you are paying for. To do that you need to choose a server that is close enough, and that has enough horsepower to run that test. Using the speedtest.net default server will suggest a site based on ping time, but, that server doesn't necessarily have the high data rate capability that other sites have. If you're closer to Toronto than you are to Montreal, I would suggest using the www.speedtest.net Toronto Rogers or Beanfield site. To use the Rogers site, when you're on the www.speedtest.net site, select "Settings" from the top menu. Then select Toronto Rogers from the drop down menu at the lower right. Save the setting and then select "Take a Speedtest" Choose the Preferred site. If you're close to Montreal, select either Montreal Rogers or Montreal Fibrenoire Internet from the dropdown menu. Save the setting and run the test.
The other issue that goes along with this is the type of test that the various services are providing. Some use single stream, some use multi-stream which is user selectable. A good many of those test sites are scattered across the U.S. and further abroad, so, you will run into slow data rates from those servers. Thats a matter of choice and really proves nothing other than what you might expect if you're running a server in one of those places. You probably can't get a fast data rate out of those servers to the point where it stresses the last mile, and your own network, which is what you should really be interested in.
If you are running via ethernet and can't get somewhere in the order of 130 down, 12 up, then it might be time to look at your signal levels at the modem to see if they are within spec. If you log into the modem, copy the downstream and upstream tables at the DOCSIS WAN page and paste them into a post. That will paste in the text contents of the tables. With those on hand, we can determine if the signal levels and signal to noise ratios are where they should be.
If you running the test via wifi, then its time to look at your wifi environment, both 2.4 and 5 Ghz to determine if you happen to be competing with other users for a clear channel to operate in. That can be virtually impossible in many locations depending on what frequency band you're operating in.
If you load inSSIDer on a dual band laptop, you can check both bands to determine who else and how many other modems and routers are running nearby. The link below is for the last freebie version of inSSIDer. It doesn't properly show the 802.11ac networks in the 5 Ghz band, but the newer licenced versions do. If you use 5 Ghz networks, its worth the $20 US to purchase so that you see the complete 5 Ghz picture.
Rogers Fixed Internet 10-20x slower then Telus LTE
I did the following test (same location, same time, same device)
1) Used my iPhone on Wifi (over Rogers Broadband 75u service)
- Speedtest.net gives me 100Mb/s down
- Fast.com (Netflix) gives me 20-40Mb/s
- Speedof.me test I get 2-5Mb/s
2) I turn off Wifi on my iPhone and switch to Telus LTE (in a dense urban environment during busy hours)
- Speedtest.net gives me around 100Mb/s
- Fast.com (Netflix) gives me around 100 Mb/s
- Speedof.me gives me 50 Mb/s
We all know Speedtest is hosted by Rogers, and probably prioritized to always be fast. Netflix likely has a CDN in the Rogers Network or nearby peering point, so also fast, but surprising that Telus LTE is double the speed.
But what is really surprising is that Speedof.me which is supposed to simulate regular web surfing performance, is 10-20x faster on Telus LTE.
I did this test after I noticed that certain websites (particularly software updates from smaller sites) were surprisingly slow.
Is Rogers selectively throttling certain sites? Or is there simply a bottleneck somewhere?
I would expect others should be able to replicate this test.
While rogers does have speedtest hosted servers.. its not their product (its ookla).
And depending on where you are, may have extra options (may have to go to their legacy tester to get the server choices)
From my location, to the closest point, gives me about 5 options? (which is Toronto)
Rogers i get above my speed, makes sense, as I am not leaving their network. This shows that at least your connection TO rogers is working, the infrastructure there, can handle those speeds.
I have also just tested with 2 other servers.
Telus - gives me the same results as the rogers server speed wise, but a higher ping.
Bell - again, higher ping, even than telus, down was the same speed, up was a little slower though.
(These are actually much better results than I expected.. normally there is a much bigger drop in the past to other ones).
I just ran the speedof.me test.. and much to my surprise, for me was able to get pretty much the same results as above. (last time I ran it, it was much slower). Same goes for fast.com.
All that being said above..
Is rogers specifically slowing down stuff outside of their network, etc? No. The tests I have run pretty much prove that. (as speedof.me goes to new york as the closest.. have not tested anything further away than new york though)
THAT DOES NOT MEAN THIS ISNT A ROGERS ISSUE.
It is. At least potentially for where YOU yourself is.
Again, my testing, i live in the GTA, north of toronto, everything filters through toronto and out onto other backbones, etc from there.
Where you are may be another story. The routing that yours takes, could be a completely different route.
Being near toronto. netflix, etc all likely have a CDN in or near the Toronto area, etc, or at least in some cases new york, etc.. so getting a fast connection to things, its a very short route and able to get full speeds.
Yours may be taking a round about way to do its routing.. which is causing those speed drops.
Telus may not experience them, as they may be getting onto a completely different backbone, and completely different routes.
(This is quite a number of years back, likely changed now, but always used to have to tell a friend to NOT do speedtests in his town. Lived in Waterloo. Would speed test in waterloo, and get less than he should! Except, he wasnt. If he speed tested ANY toronto servers (even non rogers) would get full speeds. Why? The stupid way things were routed.. the data went from waterloo, back to toronto, then back to waterloo).
I have never had Rogers speed to what I signed for, except when I call their helpdesk,
Agree, this happens to me all the time, and if you call their help desk, they will bump up the speed to justify.