I'm not sure this works very well. I tried running the test on T-mobile's LTE which has free Spotify streaming. It basically said there was "no differentiation" implying there was no packet inspection?
EDIT: Re-read the technical details. I think it can only measure one aspect of Net Neutrality I.e. Throttling one in favor of the other. Which seems reasonable. But I don't think it should be marketed as detecting "Net Neutrality Violation".
> I'm not sure this works very well. I tried running the test on T-mobile's LTE which has free Spotify streaming. It basically said there was "no differentiation" implying there was no packet inspection?
t-mobile, for instance, can't MITM TLS1.2 or TLS1.3 connections between the youtube client on your ios or android phone, and google.
they have direct peering connections all over the place with the google AS and can manipulate traffic in other ways, as it hits the edge of their network.
but they absolutely can recognize traffic on source/origin, or per-flow basis and rate limit it so that you get a 480p stream instead of 1080p (nevermind 4K).
Even this isn't enough anymore. About one or two years ago I was fooling around with DNS queries out of boredom and I did a simple nslookup on a domain I knew would be NX (it was a .ss address), lo and behold it returned an IP and the responding DNS server was not what I use, tried another domain I knew would be non-existent... Same IP, browsed to it and I got a Spectrum "search" (it literally showed results for related sites I had been browsing) page.
The real problem with this is that I never use the ISP supplied DNS servers. I double checked everything to ensure something in the network chain didn't default to the ISP DNS, turned out either Spectrum was either hijacking every DNS query but passing valid replies or running DPI and hijacking certain queries. I reported all this to EFF but never heard back. I switched to DNSCrypt to mitigate.
On newer Android versions there's a setting called "Private DNS", I usually set it to Cloudflare IPv6 DNS.
There is 100% some funky shit going on in Android though; I was working on a project the other night and my phone absolutely refused to connect to the service on my PC. My phone and tablet both on Android simply gave me "no route to host" or timeouts even with data disabled and WiFi only set; this is connecting directly to IPs no hosts/DNS. Of course I assumed the firewall was up until I confirmed every other computing device I can access was able to hit it.
Why is my phone messing with my networking on my private network?
You can edit the wifi network and change from DHCP to static IP. Then you can set the DNS for that network right there.
You run the risk where your phone is set to always use 192.168.1.147 but another device @ .147 is already connected via DHCP, but it's not really a thing that will happen because the chances are so low. And it's a bit hard on public WiFi where finding the gateway, subnet mask, etc. could be hard, or there might be some access control mechanism on the network side.
But for a large amount of people, setting the DNS on their home WiFi with this method would cover 50-90%+ of network usage. No root required.
> cue in google apologists telling me how to root or install hacks to do that and that it is totally possible
I'm not a google apologist btw. Running lineageos with no gapps. With a root-based workaround to get around Google's stupid/illogical/incorrect geography-based call recording restrictions.
> You can edit the wifi network and change from DHCP to static IP. Then you can set the DNS for that network right there.
Gross. I want to accept an IP address without accepting anything else. Just the IP address. Nothing else. Let me choose to turn off the default gateway on this network or even do something dumb like assign a different default gateway. Let me choose what DNS server to use. Let me opt-in to any other routes the network thinks I need. Let me opt-in to a time server. Give control of my device back!
If it is just checking for speed differences than it useless at checking for "Net Neutrality Violations". This is because Net Neutrality would mean treating all traffic the same.
As bigger players use more bandwidth, if treated the same they will have congestion at peak hours as they use more bandwidth per visitor than smaller sites.
For someone like Netflix to be as fast as a smaller website, they would need a peering arrangment that is special for them.
This definition of net neutrality seems to rule out even offering connections with different speeds. If I'm paying for the 100Mb plan and someone else is on the 1Gb one, my downloads will be slower than theirs. And Netflix here would be making a deal for a really big pipe (oversimplification, obviously).
I don't see a need for net neutrality to say you can't pay for more if you need more. I think "we're gonna charge you extra because of the TYPE of data you are" or "we're gonna charge you extra because we have our own competing service" or "we're gonna throttle you [for those same reasons]" are much bigger concerns.
What's the big concern with Netflix operating as their own CDN vs paying a third party one?
Net neutrality seems like an irrelevant discussion now to anyone interested in "radical" internet models (read: anything that's not deeply entrenched in the client-server paradigm). Why should we care about the battles between corporate titans over how they treat each other? Frankly, I hope the ISPs and carriers are able to squeeze them, the enemy of my enemy is my friend.
Maybe it is not so radical. The original, pre-web internet was not client-server. Each end of the connection potentially had something the other wanted. IMO, that's a truer representation of the real world. Today's internet is entirely web and mobile app centric, as if the world is nothing more than a feedlot, with only a small number of large-scale "farmers".
ironically, what you describe today happened because those central points were the ones producing the content (portals) that everyone wanted.
Now that users produce the content, they kept the distribution and revenue model (i.e. you go to the central places and they sell you to advertisers) but they have zero cost for content since everyone is a producer and consumer, which was the use case for the non-centralized portals in the first place.
An astute observation. In an ideal "web" (as imagined in the 1990's), every business might have a website, there would be news sources[1], and many nerds would have websites, but beyond that, to the non-nerd, after a while, it's not very interesting. Despite what people once might have thought in the early-mid 90's, every living person is not going to create their own website. (Or even a blog, as people thought in the 2000's). The web is finite and that is bad news for search engines.[2] UGC and "social media" have been a way for certain companies to mask this truth.
1. As I remember it, news was one of the early internet centralisation points. As dial-up telephone charges were expensive, we patiently waited for someone at a large university to download the news and forward it. I am not a great source of internet history, others will may correct me here, but one of the largest operations like this, downloading Usenet news and making it available, ended up becoming what some called the first "ISP". That was UUNet. The takeway from this footnote is that "news" showed to be an early centralisation point, high traffic. Everyone wants "the news".
2. The trend today with Google and Bing, and those who use their feeds, is to limit the number of unique search results any user can retrieve. Around 250-300 max but with many searches one is lucky to get 50-100. The search engines are trying to market themselves as a way to "get answers" instead of a way to discover what websites exist on the web today. We all know what this looks like on Google and Bing. The companies place their own "web properties" in the results, i.e., many of the "results" are links to the companies own servers, and they scrape other websites to provide "instant" answers. The user never leaves the search results page, never even visits another website. DDG, following the lead of Google, calls this "instant answers" and "zero-click info". This statement from DDG sums up the present day popular search engines:
"When people search, we believe they're really looking for answers, as opposed to just links."
(Personally I do want "just links". I have written scripts to get them.)
It is up to the reader to decide whether this is intentional or not, but either way, unlike in the 1990's and early 2000's, search engines are limiting how much of the web users can actually "see" at one time. Regardless of intent, that is the effect. If, hypothetically, the web was not growing very much, no one could detect that using a search engine. The web today is portrayed by search engines as some sort of oracle that can provide answers. For easy questions, sure. For more difficult ones, we can fabricate answers but that does not mean they are good ones. Than add in "AI" hype. What happens when people lose all critical thinking ability. They just accept the oracle's answer as "good enough". We can already see this happenig with young people. You can end up with a Wizard of Oz scenario, but no one ever discovers the tiny man behind the curtain. The truth is that the web is still a motley collection of websites, along with some very large "walled gardens" of UGC that draw the lion's share of daily traffic.
I knew the guy who wrote the very first web crawler/search engine, and even then, the intent was to find answers, not websites.
Moreover, I don't see any limit on what google returns when you search. I just searched for "the trend with google today" and it told me there were 927,000,000 results. When I got to page 22 of those results, it told me that the rest were all very similar but gave me a link to fetch them anyway. I could get to an arbitrary Nth result.
The reality is that search engines actually are, by any historical standard "some sort of oracle that can provide answers".
And when they can't/don't, they still function as tools that provide you with links to help you explore the question more.
Recognizing whether or not you've actually found the answer (from a book, from deductive processes, from a human oracle or from a search engine) has always required critical thinking skills, and that has not changed.
Yes, I can imagine that search engine folks always thought the web would be some oracle. But if you have been using the web since its debut you know that it isn't. Search results are a list of results (links, i.e., addresses/locations of the sources), not an "answer". The oracle idea (that all searches are questions with an "answer") is only realistic to some programmers. When I go to the library and search for sources (books, journals, etc), I might have a question in mind to which I seek an answer, but at the library I am only searching for the location of the sources. I do not expect an "instant answer" from the library's search terminals. In any event, not all searches are questions. Does Google Scholar return "instant answers".
For the query you tested, I could only retrieve 101 results. Could you get more than that. If you can retrieve more than 300 results (the actual results pages with the links, not just a line about how many results were "found"), I would like to know what headers you sent. I do not think this is possible anymore.
Interesting if you believe critical thinking skills are not on the decline.
Google does provide answers to simple questions, and its capabilities grow over time. Wolfram Alpha has operated in a similar domain, but using different technology (i.e. not inherently a search engine) for some time now.
I didn't suggest that critical thinking was declining. I don't have a position on whether it is or is not - I can think of several different factors that would (collectively) push in both directions.
Does it? I'm interested in radical internet models. I worry that without net neutrality we're on a gradual slippery slope towards an internet whose "slow lane" goes away entirely and I have to apply for a permit so that my IPFS node can be appropriately prioritized.
The point of the saying is to help you realize that the enemy of your enemy has at least one shared incentive, which can be very useful. It doesn’t mean you literally become their friend.
Well, until proven otherwise, the answer to that question is: centralization and trust is a huge efficiency optimization. Just look at how wildly inefficient Bitcoin is as compared to all other forms of ... well anything. The reason we see such centralization is because it represents a huge saving in energy, in time, in money. That market incentive will continue to exist into the future. Even if alternative strategies take hold, there will be a massive market for centralized solutions - they will serve the majority of the population - so we will need to make this push one way or the other for that faction to be well served.
If that ever changes, we can stop the fight. To do so now is, to say the least, premature.
What makes you think P2P will be unaffected by traffic shaping? If anything I could see sites that pay getting prioritization far above edge to edge traffic.
> This material is based upon work supported by the National Science Foundation under Grant No. (CNS-1617728), a Google Faculty Research Award, Arcep (Autorité de Régulation des Communications Électroniques et des Postes), Verizon Labs and Amazon. Any opinions, findings, and conclusions or recommendations expressed in this material are those of the author(s) and do not necessarily reflect the views of the National Science Foundation, Google, Arcep, Verizon Labs or Amazon.
The Net Neutrality debate seems so quaint now. In January 2021, ISPs are pretty much the least likely level of the stack to interfere with who can be on the internet. Hosting companies, cloud providers, payment providers, and big media platforms are where decisions are made.
NN went away in 2017. Reading the replies on reddit is troubling, thinking we need something that we actually don't. With plenty of data to back it up.
I still don't understand "net neutrality". Further, reading other people's definitions and opinions, it feels like the pro and con advocates are talking past each other.
Here's my current stab at understanding each side:
Customer perspective: Rent seeking bad. Don't want to be gouged.
Broadband providers perspective: Mitigate the free rider problem(s), make the system fair.
Examples would be very helpful. Of course there are bad actors. I'm trying to not get sucked into that food fight. My hope is that we can curtail bad behavior (cheating) with better policies.
Surely we can design markets to satisfy the needs and concerns of both sides. Right?
--
Mea culpa: My bro created and runs the ISP portions of a cable company. He relates examples of their struggles. Stuff like spending millions of dollars on gear and having other companies abusing it.
We pretty much don't agree on anything. But my bros examples and concerns are legit. So I've been trying to dig into his positions on net neutrality, to better understand the larger impasse.
I pay my ISP for bandwidth. WHAT I choose to download with that bandwidth has no impact on their bottom line, unless they were counting on me not actually using it.
Websites and web apps pay for bandwidth. Anyone who has paid an AWS bill knows there's nothing free about it.
The only group of companies that are guaranteed to make money in any internet venture are the ISPs. The rest of us are gambling, hoping that our investments will be worth it. On top of that, we're all competing with a global market, whiles ISPs tend to only compete with a handful of local providers at a time, allowing them a lot more control over prices.
Meanwhile, they're selling pick axes in a gold rush and crying victim?
You pay your ISP for their transit and peering blend. The same as I pay my Web host for access to their transit blend.
This is a seprate issue from what most people see as net neutrality. If true net neutrality was enforced, then there would be no caching boxes for Google,Facebook or Netflix and they would have to rely on public peering with the congestion that comes with.
> This is a seprate issue from what most people see as net neutrality.
Exactly.
It really feels like the debate over "net neutrality" is conflating multiple issues.
I kinda get the ISP biz model. Overprovisioning and so forth.
I have no clue about the backbone biz model. How the transit fees work. I want the ELI5 (Ray Dalio, Courtney Love) covering how broadband works. What the basics? Who's screwing who? How companies deal with each other. How they deal with content publishers like Netflix, TikTok, etc.
Consumers are not the free riders my bro cares about. It's other companies. Like when he builds out new edge servers which get filled up by YouTube, TikTok, etc. He thinks the sources should pay their fare share.
I'd like to learn what fare means, how to curtail consumer throttling, how yo curtail free loaders (bypassing transit fees), what system would be more fair.
Surely we can design market mechanisms to balance these concerns.
> Consumers are not the free riders my bro cares about. It's other companies. Like when he builds out new edge servers which get filled up by YouTube, TikTok, etc. He thinks the sources should pay their fare share.
Is he letting those sources push content to him for free?
Or are his customers determining what's going through his network?
I don't get it. How do the sources not pay? Don't they pay for hosting and bandwidth? Aren't edge servers an implementation detail that's inherent in providing bandwidth?
Me neither. I don't consider my bro a reliable narrator. What I've gleaned is the transit fees were regulated for a while.
It seems to me that Generic ISP Inc would meter all the Netflix traffic and then send them a bill. And if Netflix wants to ensure their end users are getting whatever quantity and quality of bandwidth, the two parties will negotiate.
> I pay my ISP for bandwidth. WHAT I choose to download with that bandwidth has no impact on their bottom line
This isn't true in reality. For example, it's significantly more expensive to deliver video traffic from say, my computer to yours, than it would be to deliver from Netflix to your computer.
> This isn't true in reality. For example, it's significantly more expensive to deliver video traffic from say, my computer to yours, than it would be to deliver from Netflix to your computer.
At which side of the setup? To get the data from the neighborhood hub to my computer? To get the data from the start of the ISP's network to the neighborhood? To get the data from you to my ISP's network?
And are you meaning actually from Netflix or from a CDN or from a Netflix node in an ISP's location or what?
> At which side of the setup? To get the data from the neighborhood hub to my computer? To get the data from the start of the ISP's network to the neighborhood? To get the data from you to my ISP's network?
> And are you meaning actually from Netflix or from a CDN or from a Netflix node in an ISP's location or what?
Netflix has a proxy node on your ISP's network, or at least somewhere closer to it than where my computer sits. The video makes a smaller trip to your computer.
That's one optimization, among others, utilized by CPs.
You pay your ISP for bandwidth, but not for full and constant utilization of said bandwidth. Check the acceptable use policy of any major ISP, and you'll find they can throttle you whenever and however at their sole discretion.
Possibly helpful example: the classic net neutrality example, in my mind, is a decade or so ago when Time-Warner Cable (in some markets) was throttling Craigslist to be painfully slow, because Craigslist was eating into the classified-ads business of other parts of the Time-Warner group.
Of course. I'll try harder to ferret out my bro's grievances. Get some IRL details and not just his arm wavey stuff.
Most of our policy discussions quickly devolve into partisan bullshit. It's a lot of work to keep convos on track. And when he talks about topics I do know about, a lot of the stuff he says (repeats) isn't even wrong.
He is very good at his job, and knows the technical stuff cold, which is why I'm more open to his input on net neutrality.
No PC options at all. Interesting choice, and perhaps a JS browser version wouldn't have worked properly.
Maybe I'll try it, and maybe my PC VM would have caused issues, but doing a full mobile install is a big ask.
This appears to be specifically looking at cellular data, and cellular carrier’s tendency to throttle video services. It’s not a generic service to check generic ISP throttling. Perhaps updating the title might help.
That being said, I have cellular data on my laptop and I’ve always been curious if it gets throttled the same way they throttle mobile, since I’m sure there some differences in the connection that may or may not fit the content filters the carrier has set.
When I download games from Steam I can only pull around 25.5MB/s, but when I turn on my wireguard VPN I suddenly can pull 95MB/s from Steam. The same happens with Netflix because fast.com is throttled. Comcast ISP.
From what I remember reading when Wehe came out, ISPs in Spain and Portugal (and Brazil?) were already providing data capped plans with options to uncap particular services for a small fee. So you'd get an (example) 100GB plan and pay $1-2/mo for each service you want excluded from the cap (Netflix, or whatever).
IANAL but I'm sure this was/is legal in those jurisdictions, which indicates a total disregard for net neutrality there.
So, I can have 100Gb of data for X or 100Gb of data plus all-I-can-eat from 'PopularService' for X+Y. If I choose to pay extra, that's my choice. Seems reasonable, after all it's my own bandwidth that I'm buying with my own money.
It's dodgy if the ISP serves it at lower speed for either capped or uncapped. An ISP might throttle uncapped to force the service to use lower quality and less bandwidth, or they might throttle capped to encourage folk to pay for uncapped.
It seems neutral at first glance, but the side effect is a barrier to market for competitors not on the ISP offers.
As for legalities until it's tested in court it's hard to know if any law stands up to scrutiny especially in an area as grey as this. My gut feel is that ISPs would get away with it unless they are abusing a monopoly, but hey IANAL either.
In the US most mobile carries will hijack your traffic, either slow it down(they call it whatever saver), or replace the data with their cache.
I tried a T-Mobile MVNO or T-Mobile(?) phone a few years ago, noticed regardless whatever resolution I set inside the Youtube app, it is super blur on Cellular but it works fine on Wi-Fi. The I found that the carrier setup its own cache of low resolution youtube videos, it will hijack all your requests to youtube and replacing the content to its own cache (360p videos).
That's how they give you "unlimited" data as you will never able to use it.
EDIT: Re-read the technical details. I think it can only measure one aspect of Net Neutrality I.e. Throttling one in favor of the other. Which seems reasonable. But I don't think it should be marketed as detecting "Net Neutrality Violation".