dslreports logo
site
 
    All Forums Hot Topics Gallery
spc

spacer




how-to block ads


Search Topic:
uniqs
1528
share rss forum feed


Lifeforce
Heard, not seen

join:2002-05-27
canada

[Cable] Node congestion or NOT!?

Lately, it seems, that every time I mention there is a problem with my speed(s) I am told it is node congestion. I'm reaching out to you guys on cable to explain to me how node congestion works when my speeds vary to different servers? I will post to you some speed test results to different (geographically) servers. Please keep in mind that I'm in Woodstock:

Teksavvy (Toronto):



NRBN (St. Catharines):



Megawire Inc. (Kitchener):



Nexicom (Toronto):



If we said that my poor speeds to some of the servers were caused by node congestion wouldn't that hold true for all servers? Node congestion isn't server-biased.

I noticed that my tests to servers in Toronto were all yielding results below 15Mbps. While, tests to servers in Clifford, ON and Barrie, ON (which yielded the highest speed) were higher.

What gives?

JMJimmy

join:2008-07-23
Reviews:
·TekSavvy DSL
congestion is not constant. If you want a statistically valid sample then run 16 tests to a single server over a period during prime time. Do the same at a typically non-congested time. Compare the two. That will show you congestion or not. If you want to verify, do the same thing with a second server at the same two times.

bjlockie

join:2007-12-16
Ottawa, DSL
reply to Lifeforce
Do you have a D3 or D2 modem?

yyzlhr

join:2012-09-03
Scarborough, ON
kudos:4
reply to Lifeforce
Don't know if woodstock is cogeco or rogers, but if you have neighbours who have service with the incumbent, you can ask them if they're having performance issues. If they don't have issues, then it's likely not node congestion.


Lifeforce
Heard, not seen

join:2002-05-27
canada
reply to bjlockie
D3 modem.

I don't give and take with my neighbours much as I don't see them. Most of the people in the area are seniors and it's usually lights out for them by around 9:00 pm.

Congestion aside, how do you explain the variance in speeds between servers? Especially since my best speed is from the farthest server?

Also, if congestion isn't constant then shouldn't I be able to at least see HALF of my advertised bandwidth at least some time during prime time?


TypeS

join:2012-12-17
London, ON
kudos:1
Reviews:
·TekSavvy DSL
·TekSavvy Cable

4 edits
TSI's own speedtest server is known to get hit too much during peak hours, it's an unreliable test for checking your speed level.

Likewise,going to many different servers doesn't really give much info,there are too many variables that could affect any server. My best speeds come from Beansfield or the servers in Hamilton, which logically going over the internet, are farther away then TSI's in Toronto (your data goes to Toronto first and the out the internet).

You should pick on server and due multiple test @ multiple times of the evening to set a pattern to see if congestion is occurring.

Tests done on the same servers in the same order as you posted:





technocar2

join:2009-05-29
Brampton, ON
kudos:2
reply to Lifeforce
Try these servers as well
»support.bell.ca/Internet/Internet-Speed-Test
»speedtest.bellaliant.net/

And also post a tracert to google.ca or some other website. (Start-->Accessories-->Command prompt and the type "tracert www.google.ca"


d4m1r

join:2011-08-25
Reviews:
·Start Communicat..
reply to Lifeforce
FYI guys, those flash based speedtests were never accurate....Go download a 100mb test file instead and pay attention to your download speed after the first ~25mb or so.
--
www.613websites.com Budget Canadian Web Design and Hosting

InvalidError

join:2008-02-03
kudos:5
said by d4m1r:

FYI guys, those flash based speedtests were never accurate...

They seem plenty accurate as long as the servers and other intermediate links aren't overloaded.

A large file download is not a perfect test either unless you are 100% certain there are no bottlenecks between your ISP and the remote host's storage subsystem(s) aside from your ISP's own peering arrangements and internal network structure.


d4m1r

join:2011-08-25
Reviews:
·Start Communicat..
said by InvalidError:

said by d4m1r:

FYI guys, those flash based speedtests were never accurate...

They seem plenty accurate as long as the servers and other intermediate links aren't overloaded.

You are thinking too network...I said they are never accurate because of the fact they are based on flash:

»testmy.net/ipb/topic/28902-why-d···d-tests/

As for 100MB test files, I mean grabbing them from hosting providers that often store them on premium bandwidth servers so their own internal network capacity is rarely the issue. Cachlyfly comes to mind and their file gets crazy amounts of traffic...All with minimal to no slowdowns because they have the capacity.
--
www.613websites.com Budget Canadian Web Design and Hosting


TypeS

join:2012-12-17
London, ON
kudos:1
Reviews:
·TekSavvy DSL
·TekSavvy Cable
Even with the inaccuracy of the speedtests from Speedtest.net, if everyone uses it, it can be used as a benchmark still to compare speeds (since everyone will doing the same flash based test). The OP is trying to find out if there's congestion or not, so the speedtests are perfectly valid.


Lifeforce
Heard, not seen

join:2002-05-27
canada
I have been continually running tests and we'll see where I end up.

In the meantime can someone confirm that congestion, per say, is not specific to a provider? For example, the node which is congested should be congested for both Rogers and Teksavvy customers?


TypeS

join:2012-12-17
London, ON
kudos:1
Reviews:
·TekSavvy DSL
·TekSavvy Cable
said by Lifeforce:

I have been continually running tests and we'll see where I end up.

In the meantime can someone confirm that congestion, per say, is not specific to a provider? For example, the node which is congested should be congested for both Rogers and Teksavvy customers?

Someone from TSI could confirm for their customers if there's congestion there, but for Rogers I don't think they'd tell TSI if anything is affecting their customers or not. Your best bet is to poke a neighbour with Rogers or as in the Rogers forum. You could ask in Start's forum too.

technocar2

join:2009-05-29
Brampton, ON
kudos:2
reply to Lifeforce
said by Lifeforce:

I have been continually running tests and we'll see where I end up.

In the meantime can someone confirm that congestion, per say, is not specific to a provider? For example, the node which is congested should be congested for both Rogers and Teksavvy customers?

I highly doubt its node congestion, with 12 downstream channels equalling 514mbps per mode I don't think the node will get congested any time soon and you should do a "tracert" it will tell you the ping to the node and if its too high then its node congestion and if its around 10-15ms on the 2nd hop (node) then the node is fine.