"He is currently Director of Engineering at Netflix, working on Open Connect, Netflix's home CDN also known as the CDN moving about 30% of the total US Internet traffic"

There is something about CDNs and DNS, usually not good. According to Paul Vixie, that is how we ended up with EDNS0 despite the objection of IETF. Wonder if this company gets permission to share data with Netflix. I would read the terms carefully.

Hopefully people will choose to run their own Pi-Holes on their home networks, preferably without pointing them at third party "upstream" DNS providers.

Some CDN rely on DNS for load balancing. They need information on end user network to properly resolve ( sending you to a server in the right geo for instance). If you use a third party resolver and it is not providing enough info, you may get poor performance. That's not some fancy tracking, in the end the CDN will get your IP and traffic. (Not saying it's impossible to use it for tracking purpose)

How does your own DNS server substantially help, given that it needs to get its data somewhere?

By distributing its requests across the many authoritative servers for the domains you visit. The only one that still sees all your DNS requests is the ISP.

In the US at least the major ISPs are likely the worst people to see your dns queries as they are actively selling that data to advertisers.

They see all of your traffic anyway, so it doesn't really matter if they also see your DNS traffic, it's not like HTTPS hides who you are visiting.

> They see all of your traffic anyway, so it doesn't really matter if they also see your DNS traffic, it's not like HTTPS hides who you are visiting.

Given a lot of traffic goes to cloud providers with IP pools that are discriminated largely by the HTTP Host header, it absolutely does somewhat hide "who you are visiting".

In other words, virtual hosting does provide some incidental "privacy". However, looking at the web as whole, not simply focusing on certain large CDNs, most HTTPS websites actually do not require the SNI extension. "Modern" browsers send domain names in the ClientHello plaintext automatically, by default, even though it is not required.

Then there are HTTPS websites who require SNI but do not actually check the name in ClientHello is the same as the name in the Host header.^1 Any name sent in the ClientHello will suffice to retrieve the correct web page. "Modern" browsers again blindly send more than what is required in that situation.

As such, it is the HTTP client, e.g. major browser, that is leaking information in plaintext unecessarily. "Modern" browsers are useful for displaying web content. However when it comes to retrieving it, they are less trustworthy. Too much is happening in these programs outside the user's awareness and control.

1. AWS Cloudfront is one example. It is possible to send a less descriptive, arguably more private, CNAME in the ClientHello whilst sending the known domain name in the Host header. https://news.ycombinator.com/item?id=21977961

Not really, as HTTPS doesn't currently encrypt the HTTP host header. It might someday, but even that wouldn't be foolproof.

I thought https was just http over a socket with ttls? Os, is it the certificate handshake where host header is leaked?

It's sent un-encrypted in the very first handshake setup so the server knows which public key to return. Details here[0].

There is, as you can see from Wikipedia[0], an encrypted version(esni), but that only sort of solves the problem. See [1] for more details on those.

The high level overview is, perfect secrecy of who you are talking to is a very hard problem on the Internet, and while some of these new features might help, there are a LOT of leaks to plug, so if someone is able to watch your traffic go by, chances are they can tell who you are talking to, but they maybe can't figure out what you are saying. Which may or may not matter, depending on your security threat(s).

0: https://en.wikipedia.org/wiki/Server_Name_Indication

1: https://tools.ietf.org/html/draft-ietf-tls-esni-05#section-7

> is it the certificate handshake where host header is leaked?

Yes, because of SNI [0].

In short, the ClientHello message sent by your browser as the first step of TLS negotiation (After the TCP connection is made, obviously) includes the hostname of the server you are trying to connect to unencrypted so that the server knows which certificate to present in the case of multiple sites being served on one IP/port combo.

[0] https://en.wikipedia.org/wiki/Server_Name_Indication

SNI requires the host header to be sent unencrypted, as the server needs to know which certificate to offer for the TLS session.

No. The SNI is transmitted in plain text as part of the ClientHello but TLS does not care about application implementation details. The HTTP Host header is encrypted along with the rest of the request.

I don't think ISP are a bigger threat to ad blocking than google, even if US ISPs have a very bad reputation.

s/EDNS0/& Client Subnet

Cool, sounds great! I went to NextDNS website only to find out they run Google Analytics...

I love how in this thread about, ostensibly, why tracking is bad and trackers should be blocked people are arguing that using GA is ok because doing anything else is "too hard". Why can't this same argument be applied recursively to any of the trackers in the NYT tracking sphere? I'm sure it can, and I'm sure that's why loading a NYT page loads dozens of various third-party tracking scripts: because doing anything else would have been "too hard" for the NYT marketing & technology departments.

To be slightly reasonable on this, GA on their site is nothing in comparison to the larger surface they provide. It’s not all-or-nothing.

Can you qualify that? In what sense is GA "nothing in comparison"?

Because the other companies are less well-known?

Because it's just one among 200 others? Even though it's owned by the largest adtech corp in the world?

Or is it because you told yourself that Google will probably do "less bad things" with that same data, after you give your users data away, completely out of your control.

> they run Google Analytics...

That's seriously disappointing. While I block GA and it's commonly used, my opinion of a site/service falls a fair amount when I see they're using it.

I'm not into the analytics business, but how does it come that, somehow, of all the major tech companies, it just happens to be Google to provide the one irreplaceable analytics service?

In particular, how that apparently never worried anyone before they became the "seriously nothing else can scratch this itch" quality analytics??

Because I saw it happen and it worried me. Some people must remember, about a decade (!!) ago, that half-joking nervous realisation that there was a single corporation whose server-controlled javascript ran on 90% of all webpages.

Or the part where you share all of your site's analytics with a third party you had no choice in? That wasn't even a thing before GA came around.

What is Google Analytics doing that it can't be replaced by anything that's not quite as ruthless with your visitor's data? (I really want to ask "is it that hard?" but I'm gonna assume there's something hard about it that I'm not thinking of)

Personally I'm afraid the reasons are dumb and shameful. I suppose Google Analytics is providing some additional details and data that it just happens to be unable to provide unless it tracks the everloving shit out of your visitors and accumulates this data on, say, Google servers. And people don't want to give that up, because weeeeell if it's spying on everybody and combining and keeping data anyway, they might as well get a slice of that pie, right? Flawed reasoning that work very well in unscrupulous people's heads.

And then you get someone complaining that the UX of the alternative isn't top notch. Which really tells you everything you need to know someone is willing to even begin thinking about sticking out a limb for.

I get this sinking feeling that in large parts of this industry there's less than 5% of people who actually think about and critically look at the ethics of what THEY are building, and they're probably listened to even less. It's probably even less, I've been talking to people that I consider very responsible engineers whose principles just wither as soon as you ask where the analytics data goes ... usually pointing at the client's choice. Except they're working on it and it's built into the infrastructure of the company and they provide it.

So easy to get the top thread in this comment section arguing fervently against any and all forms of tracking ... and then you get this massive back peddling when someone dares to suggest not using GA.

Is it blocked if you use their DNS? ;)

Yes

What should they use? And don't say matomo until it supports drilling down byond one level.

In this particular case I find it very dishonest: they talk about net neutrality and privacy protection but use Google Analytics. It does not matter if GA or Matomo is the best solution, they made the decision to use GA and thus don't seem to value their customers data that much.

What is it with this all-or-nothing attitude? GA provides objectively superior data and they probably get a lot of value from it, otherwise they wouldn't be using it - value which allows them to grow their business which ultimately benefits their users.

(at least for me) Considering the value of their service, GA on their marketing page seems like a very small compromise. If you don't like it you are free to block it - hell, they literally provide a service to do so.

> GA provides objectively superior data and they probably get a lot of value from it

Which is why everyone uses it and we just sit back and accept the consequences. You can't be expected to be taken seriously if you simultaneously argue that other's shouldn't do something while yourself use it for precisely the same reasons everyone else does.

This is not about all or nothing. This is about being dishonest in my opinion. It's like a doctor who recommends you to stop smoking because it's bad, but smokes on its own.

Consider a doctor that smokes during consultations subjecting you the patient to the ill effects of their secondhand smoke whilst simultaneously recommending that you give up. Perhaps a more accurate analogy. Would you trust the doctor in this scenario?

There is no dishonesty there. Merely acting inconsistent to one's advice. It's not a serious deficiency.

BTW, in a number of countries a large percentage of doctors smoke. It's a cultural thing: They pick up the habit during the stress of medical school. Would you suggest that a large percentage of doctors in those countries not inform patients that smoking is bad for them?

As a politician recently complained about: Purity tests are usually a bad idea.

It's fair to criticize. It's silly to reject their word/work because of it.

Except there's no doctor-patient discrepancy, most people in this thread are in the business and actually do have to make these choices based on their own expertise in the field.

It's a lot more like if all doctors were telling each other they should wash their hands before surgery, but many of them don't really because the tap water is cold and kinda too far away and everybody is doing it and what's the harm really and at least I'm not actively sneezing into the wound, you know?

The doctor smoking does not invalidate their assessment that smoking is bad for you, nor does it make his other "doctoring" services less quality.

For more reasons why (some/small) compromises should be allowed and not considered completely against the core value, see religious extremism.

Sure thing, but the gigantic Google Analytics machine is hardly a small compromise versus all the other ubiquitous tracking.

I don't even consider it a compromise, because literally no one arguing to use GA anyway seems to be able to present both sides of the scale considering the compromise.

Unless the doctor is smoking right next to you and making you breathe the smoke, he's not really hurting anyone but himself.

In comparison here this company is giving out all their user's browsing data to a huge advertising company without the user's explicit consent.

It's not dishonest, whether something is good or bad is never absolute. Something that is bad for someone is not necessarily bad for others.

I think first off that is hypocrisy not dishonesty and in the second place it's probably actually a regrettable and regretted addiction.

Nah it’s like a rehab clinic that requires you to smoke while you read their brochures on why you should quit :)

My dentist smokes.

Doesn't mean she can't clean my teeth and drill holes in my skull or whatever with high proficiency.

What is it with this all-or-nothing attitude? Matomo is good enough.

Depends on what you're looking for. Demographic data can be really important for a business and Matomo couldn't possibly provide that.

If you’re collecting and relying on demographic data than privacy probably isn’t very important to you.

Where are you getting this demographic data?

You don't get that from server logs.

Sound like you're not arguing against Matomo but arguing that you really in fact do want to share data with a third party to extract information that your users very much did not consent to.

It's terrifying what comes out of the wood works sometimes.

What is demographic data and how is it collected?

The age, gender, etc. distributions of the people who visit your website. I don't know anything about GA, but I'm guessing they'd know this from the currently signed in Google account.

What if people don't want to volunteer this information?

> What is it with this all-or-nothing attitude?

It's indicative of a lack of respect for their users.

> GA provides objectively superior data and they probably get a lot of value from it

That's certainly true. Also, that's orthogonal to the point. Being very valuable to website operators doesn't make its use any more acceptable to others.

> Considering the value of their service, GA on their marketing page seems like a very small compromise.

It's not a compromise until you consider the value of both sides of the equation. Please do elaborate on that, with some details on the negative externalities of blindly sharing your tracking data with Google.

Otherwise your argument just became "superior data has got some value to me, which is more than none, so yeah I got mine".

It's not less of an all-or-nothing attitude if you fail to seriously consider the other side of the supposed compromise.

What do you mean drilling down beyond one level? I use Matomo and find it just as useful as Google Analytics. Actually, I like their interface better than Google's.

Nothing? why do they need analytics?

To see what content is popular, to see what countries users are coming from, to how long a given page is retaining a visitor... all of this lets a business know what they need to change/offer/stop to better serve their customers and attract more customers.

I had a very niche ecommerce site for a couple of years, for ease I was only shipping to U.S. customers. I noticed that something like 20% of my traffic was coming from Canada so I decided to enable shipping to Canada but added a 5$ premium on top of the actual shipping cost. I had a sale to Canada the first hour of enabling it in my cart and Canada ended up being roughly 10% of my orders, even after I raised the premium/handling fee to 10$.

Looking at my GA page once caused more money to be going into my pocket and filled a need for Canadian customers. I didn't need to learn a programming language or spend tens of hours trying to figure out, through articles, how to copy paste a bunch of code together to recreate something comparable.

My question then would be, why do you need to use especially Google Analytics for that? Stats like "visitors per country" are a core feature of all major analytics suites. I set up analytics for many site on my job and I never experienced a use case where GA offered that specific feature no other software had.

GA is free and takes seconds to set up. Not everyone knows how to code, not everyone is a professional web dev or CS ninja, not everyone has piles of VC money to throw at such things. GA works, is well organized, and is often integrated with a lot of cart/web builder services where you simply have to plug in a short string and pop over to the GA website.

I needed a vegetable knife recently. I could have (and I do have the skills, and have made many knives over the past 25 years) purchased some appropriate steel and made one myself in 3-5 hours and had it incredibly sharp but I opted to buy a Mercer stainless steel one on Amazon for 20$~ because it required much less effort.

I also couldn't care less about Google or the NSA or Lectroids from another dimension tracking me so I don't rush to go "I'd better cater to the small percentage of my potential customers that want to leave zero trace on the internet, there are countless tools out there they can use to minimize that trail of breadcrumbs. This data is usable to me and Google makes it easy" and I imagine, even the company in question was like "some people want privacy but those that are concerned can easily block this with a browser extension, in their hosts file, and/or at a hardware level so we'll go ahead and use it and save a bunch of time".

> GA is free and takes seconds to set up. Not everyone knows how to code, not everyone is a professional web dev or CS ninja, not everyone has piles of VC money to throw at such things.

Imagine if restaurants had this attitude about hygiene.

I mean not everybody is a professional cleaner ninja or has money to hire them. Just wiping the counter a little is all everybody sees anyway, and everybody can do it.

If you're simply not skilled to do it properly, does that mean you get to earn the profit of doing it over your customer's backs anyway?

Your attitude really reminds me of, say, street food carts in some places. Because some of people just didn't receive food hygiene training, and they only have to get out there with a cart and something that looks edible.

You're essentially saying that violating your users' privacy is OK because not everyone has the skill or money to do it the right way?

In this case would you also say it's OK to be stealing in stores because it's cheaper than getting things the correct way by paying for them, and that stores concerned about theft should just do a better job at preventing you from stealing by using the real-life-equivalent of a hosts file to prevent you from entering the store?

I think his point is that he doesn't consider it a violation of privacy and that the people who do usually are better at controlling the data that is leaving their computer anyways.

Exactly. The people making all the fuss in this thread, are the people that are probably already blocking via one or more solutions.

"Don't track me!"

I can't, you are blocking it

"You shouldn't be tracking me!"

>You're essentially saying that violating your users' privacy is OK

I'm not holding a gun to their head and telling them to visit my site. If I walk into a business, or someone's house, I assume I'm being monitored. Websites are the same thing.

> I'm not holding a gun to their head and telling them to visit my site.

And I'm not threatening to skin you alive if you don't stop it.

How are threats of violence okay again?

If I walk into a non-Google business or someone's house, I don't assume that Google is the one monitoring me.

> I didn't need to learn a programming language or spend tens of hours trying to figure out, through articles, how to copy paste a bunch of code together to recreate something comparable.

Tens of hours, people. THIS is what your privacy is worth to some.

You forget that, while doing the unethical thing might seem a lot less hard, simply NOT doing the thing is also not hard at all! So that's not an excuse.

You seem to think you have a right to the earnings of the difference between those two, over the backs of your visitors' privacy.

THAT is the argument you have to explain. Not the part of how you managed to get fancy analytics for free by selling out your visitors.

This is an honest, detailed answer from an end user. Basically, he values simple demographic information about his visitors and doesn’t care about privacy. This is what countless other GA users have said here on HN and elsewhere.

If you want more privacy then build a better GA (from the typical end user perspective, recall the famous comment here that no one will use Dropbox because they can instead just run some Unix commands).

All of that can be done with local server logs,complete with pretty graphs and reports.

I am out of that game. Once upon a time, I labored over WebTrends for about a hundred different sites. It was not a great time.

I am not sure what would be a good replacement for Google Analytics these days but I had an absolutely terrible time with WebTrends and server-side logs. I was pretty doubtful about the results produced, as well.

And then crawlers completely mess up your data. And it’s another thing to maintain and process.

Or with a couple of mouse clicks it can be done with GA. Most people don't want to stare at lines of code, write cod, and fiddle around in logs for hours and hours. I have literally no idea how to extract that data in a comparable fashion and it would likely take me many hours of reading and trying to copy paste random bits of code I find in articles together to attempt to do the same. Obviously the company in question should have someone capable of doing that but most smaller companies and small business website owners don't, hiring someone to create that would cost money.

Stuff like GA you can copy paste an identifier or line of code, make it live and you're done. You now have a bookmark you can go to and see lots of actionable data in a nice visual form.

Most people don't care about a website tracking their OS/location/time spent on change etc so why would I, or someone else, put a bunch of extra work into fashioning something from scratch when there's a perfectly usable product that takes seconds to deploy and is easily blocked by those that don't want you to easily access that information about them?

If you need a knife most people would buy a knife. They would not buy a length of steel, cut it to rough shape, file or grind it down to the final shape, then spend an hour or more putting an edge on it with wet stones. Besides, the people that don't want tracked by Google are probably already blocking google via software and/or hardware solutions.

> To see what content is popular, to see what countries users are coming from, to how long a given page is retaining a visitor...

All of those things can be determined by using a web server log analyzer.

They should not use anything.

Or server logs, if they really must.

It seem like most people have forgotten that before Google Analytics, we all just looked at our server logs.

Sure, you can't get the exact same information, but you can get enough to do capacity planning and some basic stats.

I've started to wonder why people care about Google Analytics, what does it tell you that you actually need to know. Again capacity planning is useful, but other than that, isn't sort of pointless?

>exact same information

you could actually get more from plain old log parsing in some cases. First of all with GA you can't access the raw data. With logs you can create new type of stats/metrics/charts and apply them on past data. Also, IIRC GA uses random sampling - only a % of connections are recorded and the data being shown is extrapolated from those samples. While I'm not arguing that given the vast amount of data at disposal, huge capacity and great engineers they can make those extrapolations very accurate, I'm not entirely convinced it is precise for (very) small traffic websites. And since a small traffic logging requires both small storage space and small processing capacity the resources needed for keeping and processing your own logs are insignificant while the results might be useful.

Some of the larger website builders/hosting , e.g. Wix don’t even give you access to server logs.

Sure you can setup and run your own site and CMS easily enough, but running even hourly bulk log ingestion is usually not as straight forward and the information you can derive is very limited comparative to js based tracking.

Matamo ( formerly Piwik) is decent but still takes some time to setup and get right.

The main thing that the analytics tells you if you are promoting anything, is which of those promotions is actually working and driving visitors to your site.

>Some of the larger website builders/hosting , e.g. Wix don’t even give you access to server logs.

I don't think it's unreasonable to require them to provide you some sort in insight into your traffic data in that case.

You can to promotion tracking with just log parsing, depending on how your system is built. There's a large number of sites that handle that by simply having unique URL for each promotional partner.

> what does it tell you that you actually need to know?

Have you actually used Google Analytics? Doesn't sound like it.

First, being able to get stats on real traffic and not bots/crawlers is very important and GA does an excellent job of this.

Google Analytics also allows you to see how people actually use your site. Stats like how long, what their visit path looks like, and when they leave your site. It also lets you see demographic info, like age group and gender.

I've only touched the surface of what GA does. Yes, it is equally frightening and amazing how it tracks users.

Good luck doing these with your server logs.

> It seem like most people have forgotten that before Google Analytics, we all just looked at our server logs.

Yep, and lots of us still do. I wouldn't feel right throwing my users under the bus by subjecting them to GA.

Simple analytics? What would they even use the data for...

There are plenty of alternative solutions:

Piwik, GoatCounter, GoAccess, Open Web Analytics, clicky, Snowplow, Gauges, etc.

The correct solution is to self-host all analytics involving data that can be used to identify, track, or analyze individuals.

And none of these offer the ease of setup, ease of analysis, or ease of recognizing what actions to take that GA offers for $0 license and no hosting fees. These on the list are all admirable (esp. Snowplow) but all assume savvy users to be able to get things that are out-of-box with GA.

> don't say matomo until it supports drilling down byond one level.

everybody, this is what your privacy is worth to some people.

I’d be using GoAccess if it meant not giving up my users to Google. Matomo is a dream.

I use nextdns, it's excellent.

However beware that it's not a "set and forget" solution.

Example: This morning I did my occasional sweep of what I've blocked where, and to see if there's a new allowed domain in top N that should've been blocked. What I found is that ocsp.int-x3.letsencrypt.org.edgesuite.net is blocked by "kowabit.de - bl*cklist of death". I've added that to my whitelist now, I want certificate revocation to not be blocked.

Why would anyone block all of Akamai? This is the problem with user-generated-lists; sometimes the user clearly has no idea what they're doing.

Strange, cannot find any entry related to LE in the list: https://blocklist.kowabit.de/kwbtlist.txt

ocsp.int-x3.letsencrypt.org.edgesuite.net

matches

edgesuite.net

I whitelisted only the FQDN and left edgesuite.net largely blacklisted.

FYI: *.edgesuite.net is the domain that Akamai (CDN) allocates to nearly all of their customers.

That all sounds great, but what is their business model? How do I pay them for their service?

I checked.

It seems to be free for the first 300,000 queries per month, then switches to a regular DNS (no blocking). Unlimited queries are $1.99 per month.

https://nextdns.io/pricing

Ok, but by paying you basically hand them your personal information which they can now tie to your internet access patterns. So now, besides your ISP, there's one more party with this info.

EDIT: they accept cryptocurrencies, so the problem is slightly less critical.

In case you use their DOH service (or TLS), it's only them, not your ISP.

You ISP still sees which IP addresses you eventually connect to.

Which are mainly going to be owned by AWS, GCP, Azure, DO, or CloudFlare

SNI (Server Name Indication) leak is still present and your ISP may know what website you're asking.

It's $2/month for unlimited requests, apparently: https://nextdns.io/pricing

That's surprisingly very reasonably priced.

Pricing is surely very reasonable, but I'm not comfortable with the idea of paying for not getting something. The whole system is flawed and building businesses on a flaw is a step in the wrong direction if we want to fuel motivation to solve that flaw.

The city of flint has a lot of water theyd like to sell you, it comes with extra so it follows your model. You should also love basic Hulu, they sell you a version that comes WITH added ads, such benefit.

In reality there are plenty of places you are paying for refined or filtered products. Organic or pesticide free foods, gas (you or your supply chain use it), higher end CPUs (less defects), it goes on and on.

You're also paying for the Internet access. Go build one and we'll pay you.

Is it? You can run a DNS server on a watch.

Not on my watch

It is free during beta. Then essentially a freemium model: https://nextdns.io/pricing.

I'm sorry for my ignorance, but how does NextDNS compare to 1.1.1.1 by Cloudflare?

I've been using 1.1.1.1 for some time. For me it has worked fairly well in the USA where my connection speed is ~200mbps. While I was in India, 1.1.1.1 caused substantial slowdowns, in some cases made a few websites unusable. Typically I had it disabled. I'm unsure why this was happening because the website says it makes web browsing faster--I do not have the technical chops to understand this, so would anybody be kind enough to explain in a layman's language? Would NextDNS and 1.1.1.1 differ in terms of speed?

https://1.1.1.1

I dont think the increase in speed is a result of you getting responses to DNS queries any faster, it's simply that if you only need to load the content you want and none of the content you dont (ads and whatnot), then the stuff you do want to see has that much more bandwidth available.

But at the current time I feel CloudFlare has more reputational skin in the game, that they're merely selling off spare capacity as opposed to making a separate business model work, and they don't offer logging as a service although they can certainly break their promises at a reputational risk.

In time NextDNS can build up their reputation perhaps to a point even exceeding CloudFlare, but right now I feel those who are joining early are "paving the way" via risk to their privacy.

To my knowledge 1.1.1.1 does not offer as blocking? Or are you referring to some other cloudflare offer?

They're most likely referring to the provider reselling your data (basically you're trading scattered tracking on many providers to NextDNS getting all your DNS lookups, which can be resold without your knowledge)

You are correct. Cloudflare’s DNS does not assist you with blocking ads and trackers.

I do something similar by running CoreDNS on a pi with a bunch of blocks on domains. Works very nicely with caching as well, and the data stays where I expect it.

Wonder how this compares to AdGuard[0]

[0]https://adguard.com/en/adguard-dns/overview.html

EDIT: They have a lot more products than just DNS (if you poke around the site)

Maybe a silly question, but how do they identify you? Simply based on your IP address?

You'll have a unique 6 character id, which you'll use as a subdomain or path in dns-over-https

For example, if your id is abc123

DNS-over-HTTPS https://dns.nextdns.io/abc123

DNS-over-TLS abc123.dns.nextdns.io

You also have ipv6 hostnames, which has the id in it

Logs of requests can be viewed at:

https://my.nextdns.io/configuration/abc123/logs

Can anyone who knows your id also view your logs?

As a fallback for standard UDP DNS over IPv4, they do use your IP address in a rather clever way: they have multiple IPv4s so you can have different rule sets for different devices under the same IPv4 and they identify which ones by linking the two addresses together (NET1->DNS1 gets RULES1, NET1->DNS2 gets RULES2, etc.).

What's the advantage in using this over something like say, Blokada?

1. You can a share your config across all your devices. Also works on desktops (AFAIK Blokada is just for mobile)

2. You can create multiple configs and easily switch between those

3. Works everywhere (Pi-hole is very cool but only works within the local network you set it up, AFAIK)

4. On my smartphone Blokada regularly stopped, probably because I use an energy saving profile. I could never keep it running for longer than a few days, no matter what I tried. NextDNS seems to work fine so far, had it running for some weeks without a single crash

Regarding #3: you could run PiVPN on your PiHole device and remote into the PiHole.

Mobile browsing on the go without ads!

>On my smartphone Blokada regularly stopped, probably because I use an energy saving profile.

I had an issue with it stopping every couple of minutes, recently I figured out it was Google Fi's VPN causing it to stop after noticing it was only doing it on my phone with a Fi sim and not on my device with a Sprint sim. At some point Fi updated it so that the VPN is always on instead of only turning on when it finds an unsecured wireless network and automatically connected you (which is extremely rare where I live).

> 3. Works everywhere (Pi-hole is very cool but only works within the local network you set it up, AFAIK)

I run a PiHole on an AWS EC2 instance, then VPN to it on my phone. The VPN is configured so that only DNS requests get sent to it and all other traffic just goes straight through the LTE connection so that I'm not paying for all the traffic through AWS.

Could you please elaborate on your setup. How do you achieve sending only DNS traffic over VPN? What do you do when your phone is connected to your home WiFi network?

> How do you achieve sending only DNS traffic over VPN?

I use OpenVPN on both my phone and the server with the "redirect-gateway def1 bypass-dhcp" option enabled on the server. See https://docs.pi-hole.net/guides/vpn/only-dns-via-vpn/

> What do you do when your phone is connected to your home WiFi network?

Nothing. The phone still uses the PiHole in AWS. I don't run a PiHole on my home network, as I use uBlock Origin to block ads on my desktop. I make my phone use a PiHole to prevents apps that aren't my web browser from getting ads, such as Google Now.

Thanks for the clarification.I'll give it a trial run to see how it pads out for me.

They some track record before I can start trusting them.

That's why I wish the Signal foundation started building 1. Password managers 2. DNS providers

NextDNS's anycast network seems to be built on top of M247 and Vultr. I wonder if it is stable enough.

There are a few others in there as well: https://bgp.he.net/AS34939

So if I am running both Nextdns and adblocker locally, is there any point ? Which one is blocking first ?

I do the same and together they work the best I find. Ad-blockers act at a different level than DNS blockers so can catch things that the DNS didn't (DNS only had domain, while ad-blockers have the full path)

But the order of evaluation is first local ad blocker then dns blocker I guess, right ?

From what I understand yes

NextDNS is great. However with these tools, problems appear when you specifically want to display your Firebase dashboard etc. directly going to a service from the "tracker" itself = analytics.google.com

Marking at domain level is too generic to wrestle with this problem.

You could have an alternative browser with unfiltered DOH for this.

Or maybe that's the right level... if tracker should be blocked then I guess it's better to not fuel this business.

Thank you for sharing this! I have never heard of them before, but I switched from Cloudflare to it after reading your comment and I’ve been a very happy user for the past 4 days :-)

> founded [..] in Delaware, USA by two French founders

This sets off my spidey sense. Delaware is known to be a hot bed of fly-by-night corporations and front companies used exclusively for shady dealings[1]. I'm not saying that's the case here, but proceed w/ caution.

[1] https://www.transparency.org/news/feature/delaware_the_us_co...

It's also an extremely common place to set up a legitimate business.

> More than 50% of all U.S. publicly traded companies and 63% of the Fortune 500 are incorporated in Delaware.

https://en.m.wikipedia.org/wiki/Delaware

Delaware is used as the base for a lot of corporations of all stripes. It's because they provide an enormous tax advantage.

I actually had an argument with my business attorney about this many years ago. He wanted me to incorporate in Delaware for the tax benefit. I wanted to incorporate in my own state because that's the right thing to do.

and what prey, does a global pi-hole grid give that an actual RaspPi and pi-hole can't? What does it add?

Given they are based in Delaware, USA my starting point is far below zero privacy trust: Nextdns are gathering maximum possible data from everyone's DNS queries and selling it, or proposing to use it for advertising soon(tm). Enough independent reviewing and auditing might eventually persuade me that's unfair, maybe. If they actually cared for privacy why not incorporate in the EU and proudly wear full GDPR compliance? Or CA's upcoming privacy legislation?

There privacy policy is succinct and to the point:

https://nextdns.io/privacy

> 1. We do not (and will never) sell, license, sub-license or share any of the data submitted directly or indirectly by our users with any person or entity.

Which is easily said. Whereas their pricing page tells a different story:

https://nextdns.io/pricing

> Completely free during the beta, then free up until about 300,000 DNS queries/month — $1.99/month for unlimited queries.

... which makes for a very particular business model indeed.

There is no such thing as a free lunch. Which makes me wonder how they keep the lights on, and where the money comes from.

Im failing to see how the privacy and pricing pages stand in any kind of contrast to each other.

This business model is like literally every other SaaS platform that has a free tier. Would you trust them more if they had no free tier?

DNS-over-TLS is what they add.

Which is great, because then you can configure your Android to use nextdns as the private DNS provider and thus all adverts and tracking in your mobile apps and websites are covered too, even when you're away from your home network.

You can configure the same on your own PiHole, or VPN to the network it's on.

I think the value is convenience, all this is done for you, no maintenance.

My understanding is that pi-hole does not support DNS-over-TLS.

Are you saying that it does?

I've just looked and cannot find mention of it.

I know one can install stubby to achieve this, but now the convenience of nextdns is greater.

> My understanding is that pi-hole does not support DNS-over-TLS.

You have to run Unbound, as well.

Presumably it's just convenience for people that don't want to run their own.

It would work off of your network. If you are using their DNS it would work when you took your laptop to a coffee shop or when your phone switched to 4G/5G. That seems like a big benefit to me.

If you have your DNS server its trivial to connect to it from anywhere you are.

Do most people expose their pi-hole to the internet? I didn't think that was a common thing to do.

No. VPN tunnel in between.

Reliability, speed, accessibility, ease of use, more secure defaults... Enough?

The founders appear to live in the US so a US company is a lot more convenient for them, and Delaware has some very nice benefits for companies that AFAIK don't have any privacy drawbacks. Your assumption that they "are gathering data [...] and selling it" is beyond baseless.

EU users are protected by the GDPR anyways and despite what ProtonMail etc. would like you to believe, founding your company in Switzerland or whatever does not make you magically trustworthy.

From where I see it, they deserve no more and no less initial trust than any other company.

> Reliability, speed, accessibility, ease of use, more secure defaults... Enough?

Very glib, but not helpful nor accurate. Reliability of a web service? Speed faster than local network or local VPN tunnel? Convenience appears to be the top and bottom of it from other replies.

> assumption that they "are gathering data [...] and selling it" is beyond baseless

Not at all. US freedom of information goes far beyond everyone else's and leaves the expectation that personal data is collected, shared, sold, misused and abused, and that's been the case since long before the web. Personal data that most places consider private and inappropriate to share, and often have laws for, is frequently easily available in the US.

The web just brought it further into the gutter.

Europe has had freedom of information with constraints on personal data. Data protection has been around since the mid 90s, and prior to that there were other restrictions on certain types of data collection. The discrepancy between the two approaches has been there for probably seventy years, perhaps more, and it's widening not narrowing.

Those national norms set starting point of expectations, and what each nation tends to take as axiomatic. For a typical European, for the reasons mentioned, US privacy provision and expectation starts negative. Certain industries and categories get very antsy about data going, even briefly, to the US as a consequence of that. If it's any consolation I start with the presumption every company is untrustworthy.

> they deserve no more and no less initial trust than any other company

that is they deserve none. No company gets any trust from me until they prove that they deserve it.

I agree, but asserting they are evil with no evidence to support that is a lot further into the negative than "no trust".

Given numerous examples from the past this is what "no trust" equals to. I'm not advocating publicly accusing them of being evil without evidence, but assuming they are when deciding if and how to use their services is at least reasonable.

In other words in public my opinion on them is "neutral" until I have some information/data/signs to form/change my opinion, but internally I assume they are doing[0] something evil until proven otherwise.

[0] or at least assume they are capable of doing something evil and there is a non-zero chance that they will engage in evil-doing. (but again, not accusing them publicly until there are reasons to do so; I'm just describing internal thought-process)

Believing someone's claims out of thin air does not make it trustworthy either.

The existence of the gdpr is not enough. You'd also need to start reporting/suing those breaching the rules.

True, but my point was that incorporating in Delaware doesn't prevent the GDPR from applying to you (for EU customers). It's true that taking action against a non-EU company for GDPR violations is harder, but not impossible.