Reference for ISP: How can I prove my connection is unstable to my ISP?

Kai

Hi,
I am having issues with my cable ISP in Toronto (ISP Tek Savvy). My connection is highly unstable, the monthly reports shows this clearly - however the tech support has never heard of fing and will not accept the report as evidence of service problems. They want me to use speedtest.net and check the results while on the call - this obviously does not tell the whole story. They question that your service may use a test server far away that may affect the results of the report. How can I prove the accuracy and relevance of your test results to them?
Kai

Pooh

Hello @Kai and welcome to the community!

I'm so sorry to hear your tale, but honestly, you quite possibly can't - at least to whatever level 1 person you're probably currently speaking to.

You may want to get this elevated up a level or 2 first then try, try and try again.

True story: a good 10-15 years ago I was having some very odd issues with just some of the digital channels on our TWC cable service (e.g. massive amounts of digital artifact breakup on BBC America but crystal clear pictures on SyFy). I spent months working my way up the chain trying to convince people the issue was real. Each time they'd come out, measure the signal, fritz with amplifiers inside the house and invariably change out the DVR (this losing whatever shows we had stored again).

Finally I hit paydirt and got a hold of someone who listened to AND understood my tale. He then went back deep into their systems and found that the real underlying issue was actually a bad QAM convertor that affected only certain channels (like BBC America) in their distribution system. The moment that was fixed there, everything was normal again.

But yeah, I went through months of people thinking I was just another crazy woman who had no idea what she was talking about.

Don't give up, fight, but be prepared to get very very annoyed...

VioletChepil

We also have this service here which is detecting Internet Outages: 
https://app.fing.com/internet 
When you Login/you should see your ISP score (from other user ratings) and also any widespread outages we detected. 
@Carlo_from_Fing any ideas to add on this one on how @Kai can convince ISP of the results?
@Kai perhaps we could start another thread - "how to persuade ISP to give a discount etc. for loss of speed". I know some users have achieved this...
The other option (if you know the service is not what you pay for) is changing provider? You could also ask for recommendation from the community too...

kltaylor

My first suggestion is to NOT use speedtest.net.  Mostly because of how it relates to us here in the states, but Comcast owns them and I would gather a guess that it could be padded from time to time.

www.testmy.net uses HTML5, so no Flash running on your system at all.  You can also set it to test periodically over the course of some time, this way you can receive an accurate depiction of the peaks and valley's that you're experiencing. Take the time to talk to your ISP with someone other than Tier 1 support.  It takes some work, but once you speak with an engineer and present the information, they (again, here with Comcast) have a tendency to listen more intently.

Carlo_from_Fing

Speedtest/OOKLA is placing servers inside the ISP. Thus it's very good to measure your last mile speed.
Then when you use M-Lab, and the server is not far from you, you know your real Internet browsing speed.
E.g. in one case in Italy a convenient ISP with fiber is telling you with OOKLA that you are achieving something like 500 Mbps download.
But when you try with M-Lab or also other speed tests, you are not going further than 30-50 Mbit.
This is because OOKLA measures the speed from your home to the ISP core network.
Then when you reach outside Internet, which is the same when you are browsing, you leave their core network and go in the world via their interconnect channels and backbones.
Full explanation from our Knowledge Base and also from M-Lab in chapter '
What is the Science Behind the Fingbox Internet Speed Test' at https://help.fing.com/knowledge-base/internet-speed/

The main question here is, where is the M-Lab server that your Fingbox has been using? Is it far or near, like e.g. in same city?

atomicboy

I live East of Toronto and typically my reports run against the Toronto Server. See screen capture above.

Carlo_from_Fing

How big is the difference with OOKLA?
I believe your server is near and is also a good one (North America is the best covered continent by M-Lab).

kltaylor

atomicboy said:

I live East of Toronto and typically my reports run against the Toronto Server. See screen capture above.

That seems pretty consistent.  What are the speeds that you expect to receive?

atomicboy

In my case I am on a 25 down and 1 up fixed wireless 4G LTE A plan with Bell in Canada. On some occasions the server has been in another province but that has really made no difference.

Kai

My issue is that the speed varies - within three tests you may get discrepancies of up to 100 mb/s. I was told that on cable connections bandwidth is shared, so there would be a variance of performance. The ISP I use (Tek Savvy) does not own the infrastructure - they can only complain with the telco (Rogers) if they have sufficient proof that the service is not working as expected. I was advised that if I upgrade my package to a higher bandwidth, the average would increase, but that is not the case. I have a 250 / 20 mb connection and while I get the advertised 250 mb sometimes, the 'valleys' are pretty low and may be as long as 2-3 min. working from home via VPN on a RDP connection means at least 6-7 disconnections per hour, at peak times (after 7:00 pm almost impossible to have a consistent connetion. when I call the ISP they have me do speedtests and based on the outcome they will say that the result is reasonable for the service that I am subscribed to - if 3 out of 6 tests get close to 200 mb/s the result is deemed OK, but that does not tell the whole story. Listening to music streams means at least 2-3 disconnections per hour, a 2 hour movie on Netflix times out for buffering at least 3-4 times. when I was on a slow 7mb/s dsl I did not have these problems. Last resolution is that they want a call with me running speedtest.net and mLab from fing subsequently to proof that the results between both methods are in line - if this is the case they will accept the monthly report that they can then use to complain with the Telco (Rogers)

kltaylor

@Kai looks to see what the cost would be (if this option even exists for your provider) to change from residential service to business.  Here in the US, I have Comcast as my provider, however, I pay for Business Class internet so that:

* I connect to their VLAN with dedicated bandwidth

* I do not share my bandwidth with residential/consumer customers in the neighborhood

* Connection stays consistent

Gidster

atomicboy said:

In my case I am on a 25 down and 1 up fixed wireless 4G LTE A plan with Bell in Canada. On some occasions the server has been in another province but that has really made no difference.

Wireless is always going to be impacted by varying environmental conditions and contention with phone user. It's never going to be as stable as wires in the ground

SimoneSpinozzi

now that i think about it we have a reliable way in italy to prove our ISPs that there are problems with the service.
https://www.agcom.it/misura-la-qualita-del-servizio
however doing this means that the house/work computers disconnect from the internet for 24 up to 48 hours, which... many find unacceptable
https://www.agcom.it/misura-internet-da-postazione-fissa
https://www.agcom.it/il-progetto-misurainternetmobile.it
would it be possible to use fingbox for this so that it checks the quality of the service with the italian agcom on user demand? Like, leaving the service for a monday or a weekend, and telling everybody that internet won't be available that day... and locking out every user that tries to sneak in for that "just 5 seconds won't do any harm"?

Pooh

@SimoneSpinozzi speaking just from myself, that sounds like an extraordinary niche need there! Given the far more global requests out there are are down to basic functionality, what you're asking for sounds like you may be better off getting something like a Pi and doing some of this yourself via that.

kltaylor

SimoneSpinozzi said:

now that i think about it we have a reliable way in italy to prove our ISPs that there are problems with the service.
https://www.agcom.it/misura-la-qualita-del-servizio
however doing this means that the house/work computers disconnect from the internet for 24 up to 48 hours, which... many find unacceptable
https://www.agcom.it/misura-internet-da-postazione-fissa
https://www.agcom.it/il-progetto-misurainternetmobile.it
would it be possible to use fingbox for this so that it checks the quality of the service with the italian agcom on user demand? Like, leaving the service for a monday or a weekend, and telling everybody that internet won't be available that day... and locking out every user that tries to sneak in for that "just 5 seconds won't do any harm"?

That's an interesting feature request.

SimoneSpinozzi

Pooh said:

@SimoneSpinozzi speaking just from myself, that sounds like an extraordinary niche need there! Given the far more global requests out there are are down to basic functionality, what you're asking for sounds like you may be better off getting something like a Pi and doing some of this yourself via that.

meh. ¯\_(ツ)_/¯
this whole thread is about proving the ISP are not providing a service.
True. I just listed the italian one.
But i'm sure 99% of the countries where Fingbox is sold have something like that.

Pooh

@SimoneSpinozzi trouble is, it won't prove anything. You can show them results but they don't have to accept them.

VioletChepil

Maybe this is a different thread in itself entirely, but could you ever take any 3rd party action based on those results? Sure maybe they won't accept them - but maybe a governing body would. If enough evidence was collected. 

Pooh

Unsure what these different tests would mean to an ISP more than the current ones would. If the ISP isn't accepting the results now, why would they accept any others? Remember, ISPs are like insurance companies; they'll do anything to get out of accepting it's their responsibility.

kltaylor

Pooh said:

@SimoneSpinozzi trouble is, it won't prove anything. You can show them results but they don't have to accept them.

Completely true.  Their common phrase here in the U.S. is 'we don't guarantee speeds, we guarantee that you can reach 'up to' the speed that you have.'

It's a cop-out but that's how they protect themselves.

SimoneSpinozzi

VioletChepil said:

Maybe this is a different thread in itself entirely, but could you ever take any 3rd party action based on those results? Sure maybe they won't accept them - but maybe a governing body would. If enough evidence was collected. 

What i provided is exactly what @VioletChepil asks: "links to a government agency who is supposed to check onto the ISPs" i do not know about the situation on other countries though, or how seriously such government agencies are considered abroad.

But: Who else aside from a government is going to make any business listen? Customer associations? More or less... in that those customer associations always rely on the government to enforce said rules and or demands.

Any government... no matter where... always wants a certified information. It is the government who always provides validation to the offices who certify information. So a defeatist attitude i am seeing in these comments seems... excessive and unwarranted.

True. Any government on this earth is made up of humans, and humans do make mistakes, but considering anything "not 100% perfect" as "useless" (as the comments around here following mine seem to imply) seems a stretch too much.

Therefore If people could provide ways to certify the speed of the internet connections in their country... that is what it should be done so that there is hard, certified, and approved proof of any claim and legal actions can start to be moved on solid grounds.

SimoneSpinozzi

kltaylor said:

Pooh said:

@SimoneSpinozzi trouble is, it won't prove anything. You can show them results but they don't have to accept them.

Completely true.  Their common phrase here in the U.S. is 'we don't guarantee speeds, we guarantee that you can reach 'up to' the speed that you have.'

It's a cop-out but that's how they protect themselves.

It is a bad cop-out.

There are rules that need to be followed, usually.

"up to" sets a maximum. Then there is a maximum there is a law-enforced minimum, otherwise you could sell 1000 connections that are linked to a 100 MBps ISP connection... and tell people "you can reach speeds "up to™" 100 MBps" aaaaaaand... usually that is illegal. I never checked U.S. laws though... let me see.

okay, apparently on FCC.gov it says that when people state "up to" they have to provide at least a bare minimum of 25% the declared speed... it is not much and it is worse than the italian version of the law... but it's not as bad as i thought.

Pooh

@SimoneSpinozzi : It's not defeatist, it's being realistic. For the speed tests be taken seriously your ISP would either have to shrug their shoulders and say "you got us" (can happen, but pretty gosh darned rare) or for you to prove that the tests are accurate.

Now, how do you 'prove' this? You think you can use the Fingbox to do so. Here's the issue - your ISP can quite easily turn around and state "Prove that your Fingbox is accurate". Which you can't. The only folks who can do that are the folks at Fing. So, in order to do this you'd need to bring them in to prove to your ISP that their results are accurate. Dog love the Finglish folks, but I'm unsure if they'd be willing to be embroiled into a legal dispute between you and your ISP.

In addition your ISP will most likely start arguing that the issue isn't them - but your WiFi. Or your cable connection. Or Interference.

I get you want to fight them and somehow use the Fingbox to do so, but I fear you may be looking for too much. Dealing with this is exceptionally annoying - trust me I've been where you are, I've had to fight people on something you know absolutely to be true, but they won't listen.

Your best bet here is getting your ISP to plug something in to your network to do their own monitoring. Unsure if they'd be willing to do that, but if you use these informal scans enough they might. But trust me, this isn't going to change the world.

SimoneSpinozzi

@Pooh
no.
all you need to do is prove you followed the rules.

kltaylor

SimoneSpinozzi said:

kltaylor said:

Pooh said:

@SimoneSpinozzi trouble is, it won't prove anything. You can show them results but they don't have to accept them.

Completely true.  Their common phrase here in the U.S. is 'we don't guarantee speeds, we guarantee that you can reach 'up to' the speed that you have.'

It's a cop-out but that's how they protect themselves.

It is a bad cop-out.

There are rules that need to be followed, usually.

"up to" sets a maximum. Then there is a maximum there is a law-enforced minimum, otherwise you could sell 1000 connections that are linked to a 100 MBps ISP connection... and tell people "you can reach speeds "up to™" 100 MBps" aaaaaaand... usually that is illegal. I never checked U.S. laws though... let me see.

okay, apparently on FCC.gov it says that when people state "up to" they have to provide at least a bare minimum of 25% the declared speed... it is not much and it is worse than the italian version of the law... but it's not as bad as i thought.

It's jacked up, isn't it?

There was a huge Bru-ha-ha with the current FCC director, especially when he killed off Net Neutrality as his first duty upon being assigned.

Big box ISP's here in the U.S. have an incredible power and influence over what we can and can't do (sans illegal activity, of course) on the internet.  The regulation is intolerable, and honestly I hope that someone after election season re-institutes Net Neutrality once again.

SimoneSpinozzi

@kltaylor unfortunately... i'm italian i can help very little with this stuff. I know i will be affected eventually since all the major internet infrastructure and companies are in the US and thus any internet problems in the US will eventually affect the rest of the world (also the US is famous for asserting it's "independence" by ignoring and/or UN resolutions) but... honestly i can do nothing. I can just ask you to vote. Because as far as i know in the US turnout is below 50% which means even with a majority party you get at most a quarter of the actual population voting for whomever wins.

kltaylor

You can believe as you live and breathe I will vote ...

keith

Examples of consumer group and government testing that in theory will work.

Both use devices like fingbox connected to your modem and are free with 1000s installed.

They publish monthly reports with isp names.

The only thing missing $s to fix infrastructure in my case, like yours "congestion".

https://www.choice.com.au/electronics-and-technology/internet/connecting-to-the-internet/articles/broadband-provider-performance-measuring-program-faq

The Australian Competition & Consumer Commision can prosecute and fine isps and restrict misleading advertising.

It is not perfect but does help.

As does the consumer advocacy group "Choice" who get national media coverage when they persue any issue.

Hronos

kltaylor said:

Pooh said:

@SimoneSpinozzi trouble is, it won't prove anything. You can show them results but they don't have to accept them.

Completely true.  Their common phrase here in the U.S. is 'we don't guarantee speeds, we guarantee that you can reach 'up to' the speed that you have.'

It's a cop-out but that's how they protect themselves.

Same Here, At Perú (my country) ISP don't "guarantee" speeds... 

SimoneSpinozzi

@Hronos i understand what the companies say. But, you see... there are higher level regulations (usually) against this stuff by the state. While the companies might say X the state might say "Yes, but you still must...." In the case of the US it's (still) 25% of the "Up To" unless they specify a minimum, if they specify a minimum it must still be above that 25% that rule has not been overruled or overturned. But they can work around it by not selling you an internet service but something else. At which point you must confirm you have received an internet service, not the "something else". @kltaylor was frustrated, because many companies do not sell an internet service, but "cable TV with internet access" or other stupid rewordings like that, which are indeed useful to skirm around the minimum 25% of the promised bandwidth. Check the peruvian authorities. I mean... i get it, unless you live in Lima or Iquitos you probably have crappy stuff, especially around Chachapoyas, buuuut yeah, you guys have around the same internet connection one gets here in italy.

TL;DR: check with Osiptel ¯\_(ツ)_/¯, it might help ( ´･･)ﾉ(._.`). https://www.osiptel.gob.pe/

Hronos

@SimoneSpinozzi ,Been there done that, same thing... they get a "penalty fee" if there is an "scandal" if not, nothing happens...
Nevertheless, I was pointing out than we (and most countries) suffer the same issue that they have on USA about the ISP strategies to fulfill just the minimum at a given time (or for a zone of clients)
If I were in that situation, I am pretty sure I will fight it with all the tools at my reach, but some times this is a loss war from the beginning (and many times people don't even try).
Last time my service was faulty, they left me without service for almost a week, cable and internet, (because they say they "have" to be at my house to review, and revert what they have done remotely ...).
The only thing they are supposed to do in this cases is to discount the time I didn't have service, in my next bill...