Slashdot Mirror
T-Mobile Won't Stop Claiming Its Network Is Faster Than Verizon's (theverge.com)
T-Mobile says it will continue to claim it has the country's fastest LTE network even after the National Advertising Division, a telecom industry watchdog group, "recommended" that it stop doing so in print, TV, and web advertisements. In a statement given to Ars Technica, "NAD previously recognized third-party crowdsourced data as a way to look at network performance, so we looked at the latest results, and verified what we already knew. T-Mobile is still the fastest LTE network and we'll continue to let consumers know that." The Verge reports: The dispute arose earlier this year as part of a T-Mobile ad campaign that insinuated that Verizon's network was older and slower, and that its service did not feature unlimited plans. Verizon then filed a complaint with the NAD, which is a self-regulatory body of the telecom industry designed to settle disputes, avoid litigation, and protect against unwanted government regulation. Verizon said at the time that because T-Mobile was relying on crowdsourced data from third-party speed test providers Ookla and OpenSignal, the data was skewed in favor of T-Mobile. The data was pulled from a one-month period after Verizon first reintroduced its unlimited plans. Verizon's logic wasn't super bulletproof: the company claimed that because it had never before offered unlimited plans, T-Mobile customers -- who were familiar with the concept of throttling after a certain data threshold -- were more likely to be sampled in the crowdsourced data set provided to the NAD. Still, T-Mobile discontinued the disputed commercial, and the NAD felt the need to offer guidelines last week, advising the company not to claim its network was faster or newer. In addition, the NAD also told T-Mobile to modify its claim that it covered 99.7 percent of Verizon customers to make clear that the coverage is by population and not geographic area.
https://www.t-mobile.com/offers/t-mobile-one-unlimited-55
In my experience T-Mobile was faster, but you had to find the single cell tower in the country and stand right under it. That's a bit of a downside.
In addition, the NAD also told T-Mobile to modify its claim that it covered 99.7 percent of Verizon customers to make clear that the coverage is by population and not geographic area.
Why would anyone think anything else? X percent of Y refers to Y, not Z.
In addition, the NAD also told T-Mobile to modify its claim that it covered 99.7 percent of Verizon customers to make clear that the coverage is by population and not geographic area.
That sentence made me do a double-take. How can T-Mobile cover 99,7 percent of Verizon customers without also covering 99,7 percent (or something very close) of the same geographic area? I suppose if Verizon had a small number of customers spread over a very large area it would be possible to cover 99,7 percent of Verizon customers but only, for example, 75 percent of the geographic area covered by Verizon. Still, I just do not understand how it can be in practice anything beyond a rounding error difference. If Verizon's customer base were so skewed in that way, they would be spending very large amounts of money to serve a very small customer population across a very large area. It does not strike me as something that a big company like Verizon would do.
"How can T-Mobile cover 99,7 percent of Verizon customers without also covering 99,7 percent (or something very close) of the same geographic area?"
Because people don't use their cell phones exclusively at their billing address, which is what TMo is looking at. When I travel out of an urban area, I'm still covered by VZW. Not so much for those I know with TMo.
"National Security is the chief cause of national insecurity." - Celine's First Law
The difference between just covering homes and offices, and covering their commutes and daily travels about. Such as whether they get cell coverage at a cousin's farm.
If a cell phone only had coverage in a metropolitan area where phone service is normally had by many means, a cell phone has little use.
However, if the cell phone has coverage out beyond the city limits, or areas where one is likely to be without access to a phone otherwise, then the cellular phone has significant value. So it is something Verizon would do.
I mean, hey, if Our Fearless Leader can do it and get away with it, why not everyone else?
[End Of Line]
I kind of doubt it, because users on all carriers would be doing the same thing. Anyways, I tend to think the third party crowdsourced data is probably more likely to reflect real world usage than Verizon's source, rootmetrics. Third party crowdsourced data looks at where the users are (i.e. in their houses, in their offices, or otherwise on private property, in addition to public property) while rootmetrics just measures places only accessible to the public.
The summary says T-Mobile is looking at crowdsourced data, not billing addresses.
Any sufficiently unpopular but cohesive argument is indistinguishable from trolling.
> If Verizon's customer base were so skewed in that way, they would be spending very large amounts of money to serve a very small customer population across a very large area.
Just the other day we had a couple stories about a carrier sending notices to something like 0.3% of their customers, who live out in the boonies and are roaming on towers owned by another carrier, but they are streaming TV shows.
Consider some Texas counties. Harris county (Houston) and Dallas county each have millions of people. Loving county has 100 people. They are roughly the same size in terms of geographic area.
Suppose Verizon covers Dallas county and Loving county. T-Mobile covers Dallas county, but not Loving. T-Mobile would then cover roughly HALF the geographic area that Verizon does, while covering 99.996% of the people.
when i called my cable company to complain about the speed of my home internet service (after stuttering on youtube prompting me to do a speed test), they told me that they would only accept a complaint like that if it came from a specific speed testing service, and to try it again with that one. what a surprise that this speed test registered my connection as being ~10x faster.
utterly shameless.
"They were pure niggers." – Noam Chomsky
Yes, they could be gaming the speed tests by looking at the destination, but it is reasonable that they'd want a known source to test your speeds against, not something they don't know.
anecdotally, i have a great experience with t-mobile while in nyc (even better than my friends with verizon or at&t), but it absolutely degrades to shit when i'm in the middle of nowhere.
statistically, i can't offer a thorough analysis. however, it's worth noting that i can cover 99.77% of the population of the entire United States while covering only 84% of its area. it is quite simple: omit Alaska entirely. similarly, i can cover 90% of the population while covering 51% of area by picking the top 25 most densely-populated states.
hopefully this will at least suggest that games can be played by cherry-picking major metropolitan areas, even more easily than picking states. data is available here based on 2013 census: https://pastebin.com/nDsVCK9U
"They were pure niggers." – Noam Chomsky
If it wasn't already verified numerous times over that the speed tests are being gamed, you'd have a point. Your comment made sense 15 years ago. Now it is just a lie.
There are lots of examples online using traceroute. It isn't a hypothesis. It is a well established fact that cheating is the whole speed test market. The reason people still use them is that it does tell you the relative speed of your connection from one day to the next, as long as you ignore the actual numbers given which are horseshit.
i understand that they would want a known source. i even acknowledge that my comment is almost off-topic, because hopefully the self-regulatory body of mobile providers would not green-light blatantly compromised speed tests for one provider.
however, your argument by hypothetical contradiction aside, yes, i will give more credence to the speedtest which reflects the poor performance i observe in practice rather than the suspiciously specific speedtest my provider tells me to use, especially in light of all the other bullshit my de facto monopoly cable provider does.
"They were pure niggers." – Noam Chomsky
Your comment made sense 15 years ago. Now it is just a lie.
No, it is quite true that it reasonable to use a standard measure and not an unknown. Maybe experiment design is a skill that isn't taught anymore, but when it is, removing extraneous effects that will lead to incorrect results is always a major design element. That means not testing against an unknown.
Since I didn't claim that they didn't game anything, calling me a liar for denying it is a bit, well ... ad hominemish.
There are lots of examples online using traceroute.
Traceroute measures latency, not bandwidth. In many cases not even that, given the routine blocking of ICMP TTL expired response packets. The fact that it takes 30s to get a '*' back from a distant server during a traceroute doesn't mean it is taking 30s for packets to get there and back, you know.
Depending on the endpoints of the bandwidth tool, you can be measuring your connection, your ISPs network, the network of someone unrelated to your ISP, the bandwidth of the source, or a combination of all four. Usually it is the latter. And I forgot to include "border gateway available bandwidth".
Yes, I've seen people say "well, I've measured the throughput from site A somewhere in the world to my local system and it proves that Comcast is deliberately throttling Netflix", or whatever they think the target dejoure might be. It doesn't, but that doesn't stop it from happening over and over again. The only way to get a real number for your connection to your ISP is to use a server on the ISPs network that is set up to perform such tests. The fact that it is set up to do those tests and belongs to the ISP proves, to some, that it's a rigged game. Yeah, rigged to give more accurate numbers for your connection and try to ignore other networks' issues.
The real gem during one of these discussions was the fellow who used his home connection to his business connection to measure the bandwidth of his business connection. Even though he was paying more for business class speeds, and the ISP was the same for both, the results for his business connection were never better than those for his home. That proved that the ISP was throttling his business connection. Pure gold. You can't make that kind of stuff up.
So, yeah, the ISP can demand speed test results from a specific source so they could game the system. They could also demand speed test results from a known source so they can actually determine if the problem is on their network or somewhere else. Telling the difference between the two is hard, but assuming the worst is not always the right answer.
yes, i will give more credence to the speedtest which reflects the poor performance i observe in practice rather than the suspiciously specific speedtest my provider tells me to use,
When I get home tonight I will set up a data source on one of my Raspberry Pi systems that connects via a 10MBps link to my home router. Please set up a conference call including me when you call your ISP and complain that you aren't getting the 100MBps throughput you are paying for based on testing with my indisputably unbiased data source. I'd almost pay to hear that conversation. I'll mute my mic because I will certainly be laughing when you try to get them to act based on that data, and I'll laugh even harder if they do act on it.
It is not suspicious to anyone who understands networks that they would want you to use a known source when testing your network connection. On the extreme, it would be like complaining that your network connection is down because you get no response when you ping 192.168.20.20. The fact that the "test source" happens to be unroutable if it is not on your own internal network is significant, just like the fact that the speed test site you prefer to use is on someone else's network that your ISP has no control over.
> I get this and I get how in certain localized examples it might be the case. However, I don't really see how this can be the case on a national scale
Nationally, almost two-thirds of the population lives in only 3.5% of the land area, according to one government publication. Another set of census data has 80% of Americans living in "urban areas", which total 3% of the land.
So it's very easy for the percentage of people covered to be very different from the percentage of land area covered.
This fact makes all kinds of infrastructure in the US very different than other countries. We have vast areas, millions of square miles, with very few people. France has 1,717 people per square mile, the US has 85. With 25 times as many people in a given area, communication, transportation methods and other infrastructure that makes sense in France or other European countries doesn't work at all in the US, which has 96% less people per area.
A signal isn't important if you can't make out what anyone's saying. I prefer a quality network to Verizon's.
And the coverage argument is pretty much dead. Anyone who claims T-Mobile doesn't have coverage rivaling Verizon hasn't used it recently. It's very, very, rare I see no bars, and if I do, it's usually because I'm somewhere where nobody else seems to get coverage either.
You are not alone. This is not normal. None of this is normal.
You make a good argument, but it has been proven that T-Mobile has been gaming the system of speed tests.
Not that Verizon is any better by going through National Advertising Division (NAD), which belongs to the BBB. Verizon is rated A+ by the BBB and T-Mobile is rated F. And I'm sorry, Better Business Bureau, but if you're going to lie for your paid member, you should at least try to make your lie somewhat halfway believable.
There is no way in hell that anyone would believe Verizon Communications would be rated A+. I would know. I was a Verizon customer for a bit. The coverage was great, but the nickel and diming, the bait and switching, and the fraudulent charges. There is no possible way that they could get a higher rating than T-Mobile (which is what I am currently using right now and which has never tried to rip me off like Verizon was doing).
That's for speed, not coverage. They're two different metrics. That should be clear, since following the summary's quote from the article is an addition, starting with "In addition,". Duh.
"National Security is the chief cause of national insecurity." - Celine's First Law
It is not suspicious to anyone who understands networks that they would want you to use a known source when testing your network connection.
Yes, it is. The ISP having control over it means that it's likely within the provider's network. That is, no peering agreements involved. That's not even really "Internet" speed. That's just private network speed - the speed you may be able to connect to other users of that ISP with, were it not for asymmetric plans being standard.
And they may even further cheat by geo-routing DNS for the test to a server at the nearest hop. How fast can you connect within your ISP's network to the nearest city? Not really relevant to real-world usage.
The only way to get a real number for your connection to your ISP is to use a server on the ISPs network that is set up to perform such tests.
Who cares about that though? My ISP doesn't host anything I'm interested in. They could deliver me gigabit speeds to their network and it won't do me any good unless the speeds from outside their network are good too.
RUSSIA RUSSIA RUSSIA
I turned off WiFi on my phone and fired up speedtest. And it turns out T-Mobie is actually better than my home ISP: 101Mbps download and 24Mbps up. I'm not sure what Verizon would be here, but why would I care? 101Mbps is fast enough for everything I use the phone for, and TMO is a lot cheaper than Verizon, includes more features for my money, and unlike Verizon is not run and staffed exclusively by slimy assholes. I know people do like to go on about Verizon's supposedly superior coverage. But I've only ever noticed one place that I go where tmo tends to fade out: The Costco down by SFO. And even then, it's just data that goes. I can still call and text just fine.
I know... the plural of anecdote is not data, and all that. But T-Mobile has done fine by me. I'm happy to be paying about half what I was before for better (for me) service. And, Frankly, after having *been* a customer of theirs once, and having declared "never again" Verizon can bite my shiny metal ass. I'd even go back to AT&T before I'd deal with those scum again.
Imagine all the people...
You don't even comprehend the role that traceroute plays in investigating their reports, so why pretend you have something relevant to say? You don't understand the things you already said, so there is no rebuttal.
If you really cared to correct your blatant ignorance, you'd look it up on that cat/tube thing.
yeah, although i will accept his narrow nitpicking point about my claim, this autistic prick is missing the forest for the trees.
the ISPs, by and large, lobbied to have the definition of "internet" changed to include preferential peering, traffic shaping, and literally every hypothetically possible trick imaginable, and yet they want to advertise "internet speed" in the most advantageous way (to them) possible. nope. if "internet" now includes gaming the shit out of connection metrics, then the private network speed is not the same thing as internet speed, and the advertising is fraudulent.
"They were pure niggers." – Noam Chomsky
I don't see the issue there. They can only guarantee speed on their own network. The second you leave their network, they don't have control of those other factors. Worked for a cable company years ago. School district customer with a fiber connection complained about speeds. Seems he thought he'd be able to download at 1000Mbit from anyone he liked. We went out and put the light gear on his connection and tested it. Speeds were perfect. He was seeing slowing off network. We don't control the routing of traffic outside our network nor do we control the speed of the servers you connect to. Sorry, that's just not how the internet works.
Headline:
"T-Mobile Won't Stop Claiming Its Network Is Faster Than Verizon's "
Article:
"T-Mobile says it will continue to claim it has the country’s fastest LTE network even after the National Advertising Division, a telecom industry watchdog group, “recommended” that it stop doing so in print, TV, and web advertisements."
Those are two *very different* assertions.
Saying your "the best" in some way means absolutely nothing.
Saying you're "better than some specific competitor" is comparative advertising, and does actually mean something. Which is why you see it far less often than the other.
Presenting a misleading headline to an audience that you *know* is already sensitized to BS just tells us to ignore "The Verge" as a source of useful information.
"Reality is that which, when you stop believing in it, doesn't go away." - Philip K. Dick
I don't know. I need to answer for the things I say, so I'm even shy of simplifying claims about complex processes because they can imply odd things in simplified form.
Then again, I'm an odd sort for a politician.
Support my political activism on Patreon.
"It is not suspicious to anyone who understands networks that they would want you to use a known source when testing your network connection"
Correct, it is ,however, suspicious to anyone who is familiar with the concept of impartiality that they get to choose their own reference. Double suspicious that they choose references that are not standard for their end users own personal tests. Triple suspicious that the results never align with real world experiences.
I don't see the issue there. They can only guarantee speed on their own network. The second you leave their network, they don't have control of those other factors. Worked for a cable company years ago. School district customer with a fiber connection complained about speeds. Seems he thought he'd be able to download at 1000Mbit from anyone he liked. We went out and put the light gear on his connection and tested it. Speeds were perfect. He was seeing slowing off network. We don't control the routing of traffic outside our network nor do we control the speed of the servers you connect to. Sorry, that's just not how the internet works.
I agree that they lose network speed control outside of their network. That doesn't mean the way they advertise on Internet speed is acceptable. The meaning of the word "Internet" gives you the sense of global (around the world). If they can't do it as global but still want to advertise their own network speed, then the word should "Intranet" speed instead.
Verizon's customer base were so skewed in that way, they would be spending very large amounts of money to serve a very small customer population across a very large area"
That is exactly what they are doing.
The result is a stupidly high price which is basically a subsidy from urban populations to the backwater.
They do that because the average American is easily manipulated into spending more for services and features they don't need.
Do you know what the NYCer's solution is? Don't go to the middle of nowhere. There is nothing of value in the middle of nowhere anyway and thus they get to cut their cellphone bill in half.
i try. believe me, i try.
"They were pure niggers." – Noam Chomsky
Who cares about that though?
Anyone who is honestly trying to measure the bandwidth of their current network connection.
My ISP doesn't host anything I'm interested in. They could deliver me gigabit speeds to their network and it won't do me any good unless the speeds from outside their network are good too.
Do you want to verify that you have a gigabit connection, or just whine about it not being fast enough because other networks are slowing you down?
Yes, it is. The ISP having control over it means that it's likely within the provider's network.
Yes, I think we've said as much. Why is this a problem?
That is, no peering agreements involved. That's not even really "Internet" speed. That's just private network speed
Yes, no peering gateways that connect to other networks. I think we've said as much. Yes, it's "private network speed", but it also the speed that the ISP has sold you. They haven't sold you "gigabit to every place on the planet".
And they may even further cheat by geo-routing DNS for the test to a server at the nearest hop.
As long as the limiting factor is your connection to their network, then it doesn't matter where in their network the server is.
How fast can you connect within your ISP's network to the nearest city?
No faster than I can connect to it in this city. If the test shows I have a connection that manages 100Mbps to the local city, then I know that any throughput less than that is caused by something else.
this autistic prick
We're discussing the bandwidth of your connection to the ISP. That you want to turn this into personal insult shows a lack of support for your argument, not your superiority.
the ISPs, by and large, lobbied to have the definition of "internet" changed
Of, for goodness sake. No, they haven't.
to include preferential peering, traffic shaping,
By "preferential peering" you mean gateway congestion, of course. And "traffic shaping" has always been a part of the TCP/IP standard. "Internet", as you call it.
and literally every hypothetically possible trick imaginable,
Hyperbole is your friend.
and yet they want to advertise "internet speed" in the most advantageous way (to them) possible
The do not advertise "internet speed". They have always advertised the speed of your connection. They cannot advertise "internet speed" because there is no such measurement. What is "internet speed"? Is it 100MBps? Gigabit? 10Gbps? Different parts of the internet run at different speed. No ISP can advertise speeds for any other network than their own, and they don't even do that. They are always referring to your little connection to their network. If you don't understand that, well, there are books and stuff to look things up in.
if "internet" now includes gaming the shit out of connection metrics,
You can run your own software to measure the speed, and how they can game that is anyone's guess.
Now, if you can show that the number their speedtest sites report do not match the measurement you make during the same test, you'd have an argument. But somehow the ISP is gaming your software and you can't.
They said "Internet speed" in their claim. You sound like a Verizon lawyer. If private network speed is all that counts, they could just sell you a gigabit LAN router with a dial up modem attached.
They said "Internet speed" in their claim.
Here is what retchdog wrote in the GGGP of this posting: "when i called my cable company to complain about the speed of my home internet service". "They" didn't say nothing about "internet speed".
You sound like a Verizon lawyer.
Where have I been talking about Verizon? Are they retchdog's cable company?
If private network speed is all that counts,
I didn't say that, and you know it. What I did say is that when testing your connection to the internet, measuring your connection to the internet is what is important, not measuring the output of someone's web server somewhere on the other side of the planet.
they could just sell you a gigabit LAN router with a dial up modem attached.
Then your connection to the internet would be at dial-up modem speeds, now wouldn't it? Or do you really believe that because your router is gigabit on LAN it will always do gigabit on WAN, and so such a connection would fulfill the actual promises made about your connection to the internet? Yes, they COULD just sell you that kind of connection, but it is irrelevant because that is not what they do and it is not what they claim.
Correct, it is ,however, suspicious to anyone who is familiar with the concept of impartiality that they get to choose their own reference.
It isn't supposed to be impartial, it is supposed to be accurate, and there are technical reasons why using a server on their own network results in more accurate measurements. That removes the suspicion.
Double suspicious that they choose references that are not standard for their end users own personal tests.
I don't understand what you mean by this. If you mean it is suspicious because they aren't using the same distant servers from other people's networks to test their own network performance, well, no, it isn't. We've been over that point.
Triple suspicious that the results never align with real world experiences.
"Real world experiences" are mostly subjective, and depend on all kinds of factors that are not under the control of the ISP. They also change from minute to minute based on network usage. Why is it suspicious that quantitative answers from "calibrated" sources differ from subjective experience with arbitrary data generators?
"That removes the suspicion" No it does not. I am going to go out on a limb here and use hyperbole. But no one even here gives a shit about what you are calling "accuracy" What people want it an impartial metric to compare the performance of different ISPs to. If google.com takes 5 minutes to load on my PC I could not give a crap that your internal metrics tell you your network is amazing, because it is not, it is shit.
There are two roads, from your house, both go a bar . Your dipshit friend says "The road on the right is faster I can get to my house in exactly 10 minutes!!!!" Your not moron friend says the road on the left is faster it would take me about 15 minutes to get to the bar!!
The dipshit friend of yours is incorrect to use the standard of how fast it takes him to get home even if he is more accurate
Why? Because that standard is not useful in this context.
You make a good argument, but it has been proven that T-Mobile has been gaming the system of speed tests.
Did you even bother to read what that link points to? The article you linked is a story about someone finding a hole in a network access filter. T-Mobile allowed access to websites that contained the string "/speedtest" even after paid service had run out.
That's not "gaming the system of speed tests". That's allowing people to run speed tests using devices that currently don't have service, as a way of evaluating the network performance. The only "gaming the system" going on was when the person who found this error started hosting web data with "/speedtest" in the URL so he could get access through T-Mobile without paying.
What people want it an impartial metric to compare the performance of different ISPs to.
No. Most people have one ISP. How do they compare two if they don't have service from both?
But even if they do have two ISPs, then comparing between the two is something very different than verifying if the service you are paying for is what you are getting. If you want to know if your connection to ISP A is running at "up to 100MBps", you test that connection using a source on ISP A's network. You don't use a source on ISP B, because that makes it trivial for ISP B to "game the result" by being slow in sending you data. You don't trust A to be honest, but you trust B -- who could benefit by lying to you and getting you to switch to B.
If google.com takes 5 minutes to load on my PC I could not give a crap that your internal metrics tell you your network is amazing, because it is not, it is shit.
And you will ignore the fact that the google.com servers could be suffering a meltdown right at that moment, and you will blame your ISP for every fault you find in any data service you try to connect to. This is "impartial" and "accurate". Sure.
The dipshit friend of yours is incorrect to use the standard of how fast it takes him to get home even if he is more accurate
Not when you're testing the claim of which road is faster for him to get home.
Why? Because that standard is not useful in this context.
Exactly what standard is useful when you want to test whether your connection to your ISP is providing the bandwidth you are paying for? No, I'm sorry, but "google.com loads slow" isn't it. There are simply too many outside actors involved that all can slow things down. If you don't understand why you try to eliminate all the interfering effects from irrelevant sources when making a measurement, then there is nothing left to explain.
You don't even comprehend the role that traceroute plays in investigating their reports,
I know what traceroute is, what it does, and what it measures. I don't have to "comprehend" what role "they" thought traceroute plays in "their reports", I know what role it CAN play, and measuring bandwidth is not it. I also understand, apparently better than you, that latency and bandwidth are not the same metric.
Your insults notwithstanding, the fact remains. Proving that someone is "gaming speed tests" based on traceroute is nonsense.
I agree that they lose network speed control outside of their network.
Then how can using a speedtest server outside of their network be suspicious, unfair, unethical, or "cheating"? That's what the claim is.
That doesn't mean the way they advertise on Internet speed is acceptable.
Here's how CenturyLink, as one example, talks about their "internet speed" (which they don't actually say): "Private, Direct Connection and Speed Claims: Private, direct connection and/or speed claims are based on providing High-Speed Internet customers with a dedicated, virtual-circuit connection between their homes and the CenturyLink central office."
Between their homes and the central office. Nothing about "internet globally". Also: "Speed Demo Chart: Speeds provided above may not be typical or standard and each user may not experience the exact same results as depicted for each category. Listed broadband speeds vary due to conditions outside of network control, including customer location and equipment, and are not guaranteed."
It seems like CenturyLink is quite explicit in saying "not guaranteed" and "due to conditions outside of network control". I've see the same kind of disclaimers on other ISPs ads, so CL is not unique in this honesty.
The meaning of the word "Internet" gives you the sense of global (around the world).
Oh, please. Anyone who thinks that the use of the word "internet" in some advertising makes every claim about the service being sold a guarantee that applies to everywhere on the Internet is just, well ... there has to be a reasonable person test to every ad, and if a reasonable person would not think that his 100MBps fiber internet meant he could get data from a server in Lower Elbonia on a 56k dialup at 100MBps, then that passes the test.
Anyone who is honestly trying to measure the bandwidth of their current network connection.
It could be useful in determining where a problem is if you're not getting the speeds you want. Beyond that I don't see the use. What do you use it for?
Do you want to verify that you have a gigabit connection, or just whine about it not being fast enough because other networks are slowing you down?
I want a useful measure of how fast my internet connection is. As I said, I don't care how fast my connection to my ISP is, because I don't pay them so I can get a connection to their network. I pay for a connection to the internet. That's what I care about.
"They" didn't say nothing about "internet speed".
That is what they ISP calls it in their advertising material - they call it Internet speed.
Then your connection to the internet would be at dial-up modem speeds, now wouldn't it?
You said that the only speed that matters is your connection to the private network provided by the ISP. In that case, that would gigabit, not dial-up speed. And if your ISP is the one providing it, what difference does it make what its reach is? You said that any private network provided by the ISP is "The Internet."
Without peering agreements, most of "the Internet" (the actual service connection on offer) is inaccessible. So being able to connect to your ISP's private network at 100Mbps means absolutely nothing at all.
measuring your connection to the internet is what is important, not measuring the output of someone's web server somewhere on the other side of the planet.
So let's assume they're picking some private speed test server in Korea, and not something like Ookla's Speedtest.net that has endpoints literally everywhere around the country.
The use-case for 99.9999% of consumers is to compare ISPs. I am sorry that is just the reality. Not in having both at once, but to see if a switch is worth it between different providers. "Hey I just switched to FIOS can I now download my porn faster? Y/N if so how much faster"
You are right that the the system can be gamed, that is exactly the problem everyone else is pointing out here.
When ISP's force their customers to try and use their own in-house test to validate their they are basically pissing on their own clients as rightfully no one trust their in house tests to be honest.
The solution is to use some 3rd party trusted appraiser. No one should trust A or B. In virtually every situation where side A and B can not be trusted and you can't afford to do a rigorous system test yourself , the solution is always to find an independent source C who you can put some faith in.
Oh, you meant bandwidth?
Then how can using a speedtest server outside of their network be suspicious, unfair, unethical, or "cheating"? That's what the claim is.
Who said I claim that? Where in my post say that it is suspicious, unfair, unethical, or cheating to use speed test server "outside" of their network? Reread again. I simply agree with the parent that speed test using server within the network could, of course, be guaranteed for the advertising speed. And that's cheating.
Here's how CenturyLink, as one example, talks about their "internet speed" (which they don't actually say): "Private, Direct Connection and Speed Claims: Private, direct connection and/or speed claims are based on providing High-Speed Internet customers with a dedicated, virtual-circuit connection between their homes and the CenturyLink central office."
And that's exactly what I am against -- wrongly use and/or associate the word Internet. They don't do it directly but attempt to play with words. Humans mind will attempt to associate what they see with what they know. If they don't know, they will associate whatever closest to their experience with those words. In other words, assumption happens, and providers know exactly what they are doing.
Playing around the edge of the meaning by associating words with something humans are familiar with, but not quite know exactly what those words are, could (and would) psychologically mislead their mind. No need to "explicitly state" that it is not about global Internet in this case. Just use the word "speed" and may add "streaming faster" would be suffice to make most people associate that with the "real" Internet. Why? Because they don't know or understand how Internet works. They all are not engineers (especially have interest in networking) and their experiences are from using computer and get on world wide web / email only.
It seems like CenturyLink is quite explicit in saying "not guaranteed" and "due to conditions outside of network control". I've see the same kind of disclaimers on other ISPs ads, so CL is not unique in this honesty.
Oh please, honesty! I wouldn't say "quite explicit" at all. In every commercial and/or advertising, this kind of definition is "always" either in a small print or a fast talking at the end of a commercial on TV while other things else are flashing by in order to distract audiences. Please don't exaggerate or conflate this kind of behavior as "explicit" because it is not.
Oh, please. Anyone who thinks that the use of the word "internet" in some advertising makes every claim about the service being sold a guarantee that applies to everywhere on the Internet is just, well ... there has to be a reasonable person test to every ad, and if a reasonable person would not think that his 100MBps fiber internet meant he could get data from a server in Lower Elbonia on a 56k dialup at 100MBps, then that passes the test.
Oh please. Reasonable person? Let me see... Do you believe the majority of people are reasonable? Do you think laymen are reasonable? Don't use yourself as a standard. You should have known it by now that average people are NOT reasonable. That doesn't count those who are reasonable at a certain degree but then let their bias/faith/belief blind their reasoning, and as a result, they become unreasonable. Do you want an obvious example? Deity religious is one of them. So stop using "reasonable person" as an excuse to say that misleading/falsify ads are OK because consumers should already know if they are reasonable.
I understand that you come from technical point of view that there is no guarantee on Internet speed. But at the same time, you completely ignore psychological point of view that providers play on their consumers. If there is total honesty in any business, there shouldn't be advantages or disadvantages but rather a co-exist that help each other to grow. I highly doubt that ISP and consumers are in that kind of relationship because all ISP are taking advantage of consumers in some way along the way (later or sooner).
Who said I claim that?
Nobody, including me. I said that was the claim.
They don't do it directly but attempt to play with words.
So they don't say it, but you want them to have said it, so you will assume they did and go from there. They didn't say it directly or indirectly. Nobody advertises what speed you will be able to get data from arbitrary sources across the internet. They can't. THAT would be false advertising, and their lawyers are smarter than that. No, instead, they are pretty explicit in telling you that the speeds they quote are for your connection to their network. Nothing past that. And they can't guarantee anything past that. Nobody can, and it is lunacy to assume that they could.
I wouldn't say "quite explicit" at all.
I quoted the statement from CenturyLink example that the service speed they were offering was from your end of the connection TO THEIR CENTRAL OFFICE. That's pretty explicit. Not "from you to the world", and not even "from you to our network". It was limited to the central office -- a specific place.
Oh please. Reasonable person?
Yes. Unreasonable people can make all kinds of specious and moronic arguments about anything that is advertised and the law protects the advertiser from them. This is just one example. "Gee, you said "internet" and "20 gigs" in the same ad, so that "20 Gigs" must mean that you can get anything from anywhere on the Internet at "20 Gigs." Wow. That would be so fantastically impossible to guarantee that no reasonable person would believe such a claim was being made, especially given the clear limit of "from your home to our central office".
Before you start with the false advertising claims again, you should look at any of the ridiculously outrageous ad claims that are so obviously parodies that no reasonable person would believe them, and thus no lawsuits are ever filed. Do you remember the "money out the wazoo" ad for eTrade, I believe it was? Do you REALLY think that using the eTrade product will cause hundred dollar bills to spew from your rectum? That's what the ad showed. Then why would you believe that a service that is advertised as "up to 20 Gigs" means "exactly 20Gbps from anyplace on the planet"?
Deity religious is one of them.
Wow, way to turn a technical discussion into an anti-religious rant. When have you seen a company advertising "deity religious", whatever that is, and are they advertising bit rates for a connection to it?
So stop using "reasonable person" as an excuse to say that misleading/falsify ads are OK
It's not false advertising, so I'm not trying to excuse it as such. Don't tell me to stop doing something I'm not doing.
I understand that you come from technical point of view that there is no guarantee on Internet speed.
That is not just a "point of view". That is an indisputable fact. It is true whether you choose to believe it or not.
But at the same time, you completely ignore psychological point of view
Yes, I ignore the false facts from ignorant people who ignore not only the fact that the speed cannot be guaranteed to such an extent, but the fact that they are told the specific limits of that guarantee. Ignore the words at your own peril. Yes, I've been bit by that. I bought a radio from eBay at what I thought was a great price, thinking it was the VHF version. When it showed up and I realized it was UHF, I felt cheated. Then I reread the eBay listing. Yep, the information was there, I just ignored it. That was MY fault, not the seller's. Why is it not YOUR fault if you don't bother reading the information about the service you are buying to see what it actually is?
there shouldn't be advantages or disadvantages but rather a co-exist that help each other to grow.
Here's
I want a useful measure of how fast my internet connection is.
Your connection to the Internet is through that line from you to the ISP. That's the bandwidth YOU pay for. That's the service you need to test if your connection speed is in question.
As I said, I don't care how fast my connection to my ISP is, because I don't pay them so I can get a connection to their network.
That is EXACTLY what you are paying them for. Access to the rest of the internet depends on having that connection to their network. They do not advertise, nor are you paying them for, a specific bandwidth gateway to other people's networks. (Yes, if you talk to Level3 or one of the other major backbone providers you might talk details like that, but not when you talk to Comcast Home or CenturyLink, and those are the kinds of ISPs we're talking about here. If you are talking details like that, they will be spelled out in the contract you sign for service and you damn well better be reading that contract.)
But let's remember the context here. Someone said that his ISP was not accepting speed test data from arbitrary outside sources when he was calling them about the internet connection to his home. That is specifically an issue of his connection to the ISP. While it might be nice to know how fast you can get data from Lower Elbonia servers, it is irrelevant when determining if the connection to your home is providing the service you are paying for.
That is what they ISP calls it in their advertising material - they call it Internet speed.
I have never seen an advertisement for "internet speed". I've seen ads for the speed of your connection to the ISP. Can you provide a link to one?
You said that the only speed that matters is your connection to the private network provided by the ISP.
No, actually, I did not make such a blanket statement. What I actually said is that if you are talking about whether your home internet service is providing the bandwidth that the ISP is selling you, THEN the only speed that matters is the speed on the connection that your ISP is selling you.
In your contrived example, the speed of the connection would be 56k since that is the connection from you to the ISP. Putting a gigabit LAN router on the local end of a dial-up modem doesn't change that.
In that case, that would gigabit, not dial-up speed.
You can't be that technically ignorant, so I have to assume you are misstating the situation for a reason. No, putting a gigabit LAN on your end of a dial-up modem doesn't change the service you are paying for into gigabit.
You said that any private network provided by the ISP is "The Internet."
No, I didn't say that, either. Are you sure you are replying to the person you think you are?
So let's assume they're picking some private speed test server in Korea, and not something like Ookla's Speedtest.net that has endpoints literally everywhere around the country.
I'll point out that the discussion has been about the "suspicious" nature of an ISP using a speed test server on their own network, so there is no assumption about Korea. Why should we make such an assumption, and how would it make the measurement any more accurate to do that than to use a server on the same network?
I'm assuming you mean they have physical servers co-located in many ISPs networks when you talk about their numerous "endpoints", since every speed test connection has two endpoints. One of those endpoints is always on the network connection being tested, and thus are "literally everywhere on the planet", not just "around the country". If that's what you mean, then of course an ISP should accept that data since they've explicitly vetted it and it is on their own network. But nobody has said "ookla" until now, and the discussion has been about speed test servers located on other people's networks, not the ISPs.
The use-case for 99.9999% of consumers is to compare ISPs.
Citation required. Since I doubt that 99.9999% of consumers have two ISPs with the same service from both, your claim seems to be outside the norm, and thus requires support. While I have access to several ISPs services, none of them are the same quality or level, so comparing any one of them against another would be useless. I.e., a speed test I do on my phone compared to my home cable internet would be meaningless, as would a comparison of my home cable to my business cable service (ignoring for the sake of argument that I can't switch cable providers even if the comparison was valid.)
Not in having both at once, but to see if a switch is worth it between different providers.
You can't compare two things if you don't have two things to compare. How can I test ISP B to see if it is worth switching to ISP B unless I already have ISP B to test with? You might try claiming that I can do it from my friend's house, but that is his service from ISP B, not mine. He may have 100Mbps service and I have 30Mbps -- the comparison will be useless because B will always look faster than mine, and I may wind up switching to a service that is actually slower.
You are right that the the system can be gamed, that is exactly the problem everyone else is pointing out here.
And I haven't denied it. I'm pointing out that there are valid and reasonable technical reasons for an ISP to rely on speed test servers that are on their own network and not on servers elsewhere, whether or not they are "gaming the system". In fact, "gaming the system" is one very good reason why they SHOULD ignore outside servers. If you believe that YOUR ISP is gaming you, then why would you accept a speed test run on ISP B's network?
If you use your ISPs speed test server and both your and your ISPs measurement shows that your connection is operating at or above the specified speed, you have corroborated evidence you are getting what you paid for. While the ISP can fudge the number on their end, you are making the same measurement. If you aren't getting what you paid for, then at a minimum your measurement from their server will show that.
IF you use a speed server on someone else's network and it and your own measurement shows a data rate below advertised, then you have learned nothing about your connection. The server could be seeing a connection from a Comcast IP address and deliberately sending slowly, and both the server and your measurement will detect that. But it isn't the fault of your connection, it's the remote server. Calling your ISP to complain about your connection based on that data would be wasting everyone's time. Yours included.
The only useful result from the latter test would be if both the server and your measurement show you are getting rated speeds. In that case you aren't calling your ISP, so it doesn't matter what server they want data from.
When ISP's force their customers to try and use their own in-house test to validate their they are basically pissing on their own clients as rightfully no one trust their in house tests to be honest.
But you will trust someone else. And you won't make your own measurement, or you don't trust your own measurement.
The solution is to use some 3rd party trusted appraiser.
You've missed the point entirely. If that 3rd party uses a server that isn't on the ISPs network, then there are too many places outside the control of the ISP where the speed can be limited. It can be gaming on the part of the other network operator, or it can be simple network congestion or a normally slower network path.
What I will agree is that as long as the 3rd party server is on the ISPs network where that ISP is solely responsible for service to both the server and you, THEN it should be valid for testing. And I've always said that the server nee
Can you see the forest yet or is your head stuck in a tree?
Using Ookla as a test, you can still test to a geographically close network that isn't specifically in the provider's own private network.
Yes, outside networks matter. it's the "inter" part of Internet. Few web sites are actually collocated within Charter or Comcast's network outside of some edge caching services. Just because they don't say "Internet" speeds, it's implied - enough that the FTC would assume they mean Internet. Secretly meaning no high speed access to the outside world is just plain weaselly.
You didn't get the dial up modem story. I won't explain much further. But deciding what counts as uplink and what is local is arbitrary - and saying that you have 50Mbps but only to the node in the alley is not much different. How many hops until it counts as a valid speed to advertise with in your book? The clear assumption is that the internetworking speed is what's relevant here.
That is EXACTLY what you are paying them for.
Nope. If that's all I got, I wouldn't pay for it.
They do not advertise, nor are you paying them for, a specific bandwidth gateway to other people's networks.
I'm not really even paying them for a specific bandwidth to their own network - it's "up to" a number. Though for some reason I have occasionally seen speeds significantly above that number.
But let's remember the context here. Someone said that his ISP was not accepting speed test data from arbitrary outside sources when he was calling them about the internet connection to his home.
There needs to be something between "downloads from this random server are too slow" and "we will only accept speed tests from this service that we have artificially juiced to make sure you always get good speeds". Without actual competition between ISPs we will probably never get that though.
look, if they insist on using specific speed test then it's likely they don't throttle that one AND that they got their transparent proxy tweaked for it.
really it's no use complaining with them, switch the provider if you can.
they only will guarantee the last link speed, which is pretty useless.
world was created 5 seconds before this post as it is.
Yes, I linked exactly to the article I wanted to.
And I'm sorry, but your assumption(s) that non-customers would be allowed to do a speed test is just complete nonsense.
The string "/speedtest" is not some random error. The kid didn't just stumble unto it by accident. Filtering by url to give false readings of speed is a well-known marketing practice that has been going on for decades. And that practice isn't limited to cell phone carriers either, many ISPs have done something similar. If you don't believe me, I can dig up those other examples and even find examples of former ISP employees admitting as much.
Why don't we do that? I'll find examples of what I'm talking about at other companies and other ISPs, and you find examples of what you're talking about at other companies and other ISPs? Then, we can see which one of us found what he was looking for. What do you say?
Yes, you would have to have comprehended the role that the tool plays in a particular context in order to understand what the fuck anybody is talking about. That is always true.
And no, you don't know if you understand some other technical detail better than somebody else, or not. That isn't knowable to you. Therefore you wrong about it, even without knowing how much anybody knows about the detail.
Random words notwithstanding, it is already well proven that the speed tests are being gamed and instead of looking it up somewhere you made up some fake "facts."