Because those ephemeral LE certificates are such a great idea...
replies(6):
Then, when you have only two or three big SSL providers, it's way easier to shut someone off by denying them a certificate, and see their site vanish in mere weeks.
- We went from the vast majority of traffic being unencrypted, allowing any passive attacker (from nation state to script kiddie sitting in the coffee shop) to snoop and any active attacker to trivially tamper with it, to all but a vanishing minority of connections being strongly encrypted. The scare tactics used to sell VPNs in YouTube ads used to all be true, and no longer are, due to this.
- We went from TLS certificates being unaffordable to hobbyists to TLS certificates being not only free, but trivial to automatically obtain.
- We went from a CA ecosystem where only commercial alternatives exist to one where the main CA is a nonprofit run by a foundation consisting mostly of strong proponents of Internet freedom.
- Even if you count ZeroSSL and Let's Encrypt as US-controlled, there is at least one free non-US alternative using the same protocol, i.e. suitable as a drop-in replacement (https://www.actalis.com/subscription).
- Plenty of other paid but affordable alternatives exist from countless countries, and the ecosystem seems to be getting better, not worse.
- While many other paths have been used to attempt to censor web sites, I haven't seen the certificate system used for this frequently (I'm sure there are individual court orders somewhere).
- If the US wanted to put its full weight behind getting a site off the Internet, it would have other levers that would be equally or more effective.
- Most Internet freedom advocates recognize that the migration to HTTPS was a really, really good thing.
- We now provide a completely free certs for a malicious web-sites
- Degraded encryption value so much it's not even indicated anymore (remember the green bar for EV?)
- Pavlov-trained everyone to dumb-click through 'this page is not secure' warnings
- SNI exists and even without it anything not on CDN is blocked very easily
This is fundamentally a naive understanding of both security and certificates. Paying money absolutely makes people trustworthy because it's prohibitive to do it at scale. You might have one paid malicious certificate but you can have thousands of free ones. The one malicious domain gets banned, the thousands are whack-a-mole forever.
Further, certificates used to indicate identity in more than a "the domain you are connected to" sense. There was a big PR campaign to wreck EV certs but EV certs generally were extremely secure. And even Google, who most loudly complained about EV, has reintroduced the Verified Mark Certificate (VMC) to replace it and use for new things like BIMI.
Do we have any statistics for how many people are actually doing this? Such warnings are so rare in my experience that, by default, I don't trust a site that has no SSL/expired or invalid certs and won't click through if I see that warning.
I don't much care about BIMI. People keep trying to resuscitate that particular dead dog (email security), maybe one day they will succeed but I don't expect to be involved.
I still don't understand why this is so terrible.
Public wifi networks were certainly a real problem, but that's not where the majority of internet usage happens, and they could have been fixed on a different layer.
If you're on a traditional home internet connection, who exactly can tamper with your traffic? Your ISP can, and that's not great, but it doesn't strike me as blaring siren levels of terrible, either. Even with HTTPS, the companies behind my OS and web browser can still see everything I do, so in exchange for all this work we've removed maybe 1 out of 3 parties from the equation. And, personally, I trust the OS and browser vendors less than I trust my ISP!
Some progress is better than none, and it's still nice that my ISP can't tamper with my connection any more. Unfortunate, TLS also took away my ability to inspect my own traffic! This makes it more difficult for me to monitor what my OS and browser vendor are doing, and as I've said previously, I trust these parties comparatively less than my ISP.
> - We went from TLS certificates being unaffordable to hobbyists to TLS certificates being not only free, but trivial to automatically obtain.
Sure, but it's also trivial to just throw up a website on Github Pages, or forgo the website completely and use Instagram. TLS is "trivial" if you rely the infrastructure of a specific external party.
Please help me understand what I'm missing because I find this really frustrating!
It might be more correct to say that Certificate Pinning made it so you can't inspect your own traffic - for sites with TLS but without certificate pinning, you can just as easily create your own root certificate and force the browser and OS to trust the cert by installing it at the OS level. This is (part of, atleast) how tools like Fiddler and Charles Proxy allow you to inspect HTTPS traffic, the other part being a mitm proxy that replaces the server's actual cert with one the mitm proxy generates [0]
[0]: https://www.charlesproxy.com/documentation/proxying/ssl-prox...
Edit: To be clear, I'm not even suggesting the software would be doing this maliciously! Apps do all sorts of weird things when you try to proxy them, I know this because I do run most of my traffic through a proxy (for non-privacy reasons). Just for example, QUIC gets disabled.
While I don't really have a scary threat model, I don't love the idea that my ISP could have been watching my traffic. Maybe there's a world where my government has ordered ISPs to log specifics about traffic in order to trap dissidents doing things they don't like. But sure, I live in the US, which isn't (yet) an authoritarian nightmare (yet!). But maybe I live in Texas, and I'm searching for information about getting an abortion (illegal to have one there in most cases). Maybe I'm a schoolteacher in Florida, and I'm searching information on critical race theory (a topic banned from instruction in Florida schools). I want that traffic to be private.
> Even with HTTPS, the companies behind my OS and web browser can still see everything I do, so in exchange for all this work we've removed maybe 1 out of 3 parties from the equation
I mean, that's on you for using a proprietary OS owned by a for-profit corporation. I get that desktop Linux or a de-Googled Android phone isn't for everyone, but those are options you have, if you're really worried.
And there are quite a few major browsers that are open source, so even if you can't inspect their traffic at runtime, if you really are truly serious about this, you can audit their source code and do your own builds. Yes, I would consider that unnecessarily paranoid, but the option is there for you, and you can even run these browsers on proprietary OSes. And honestly, I assume you use Chrome anyway; if that's the case then you clearly are not serious about this if you're using a web browser made by an advertising company. (If you're using something else: awesome, and apologies for the bad assumption.)
> Unfortunate, TLS also took away my ability to inspect my own traffic! This makes it more difficult for me to monitor what my OS and browser vendor are doing
You can still do this, but it does require more work setting up your own CA and installing it as trusted in your own devices, and them MitM'ing your traffic at the router in order to present a cert from your CA before forwarding the connection on to the real site.
Yes, this is out of reach for the average home internet user, but if you are the kind of person who is thinking about doing traffic monitoring on your home network, then you have the skills to do this. Meanwhile, the other 99% of us get better privacy online; I think that's a perfectly fine trade off.
> and as I've said previously, I trust [my OS and browser vendor] comparatively less than my ISP.
My ISP is Comcast; even if my OS and browser vendor was Microsoft or Apple, I think I'd probably still trust Comcast less. Fortunately my OS and browser vendors are not Microsoft or Apple, so I don't have to worry about that, but still.
> Sure, but it's also trivial to just throw up a website on Github Pages, or forgo the website completely and use Instagram. TLS is "trivial" if you rely the infrastructure of a specific external party.
Running a website, even from your home internet connection, still means relying on the infrastructure of a third party. There's no way to get away from that.
And you still can run one without TLS. Browsers will still display unencrypted pages, though I'll admit that I'd be unsurprised if some future versions of major browsers stopped allowing that, or made it look scary to your average user.
> Please help me understand what I'm missing because I find this really frustrating!
I think what you are missing is that people actually do value connection encryption, for real reasons, not paranoid, tin-foil-hat reasons. And while you do present some valid downsides, we believe those downsides are overblown, or at the very least worth it in the trade off. It's fine for you to not agree with that trade off, which is a shame, but... that's life.
Malicious websites never had a problem buying certs before. Sure, the bar is lower now, but I don't think it was a particularly meaningful bar before. Besides, the most common ways to get malicious websites shut down are to get their webhost to cut them off, or get a court order to seize their domain name. Getting their TLS cert revoked isn't common, and doesn't really do the job anyway.
> Degraded encryption value so much it's not even indicated anymore (remember the green bar for EV?)
No, we've degraded the identity verification afforded by EV and those former browser features. Remember that the promise of SSL/TLS was two things: 1) your traffic is private, 2) it verifies that the server you thought you were contacting is actually the one you reached.
I think (2) was always going to be difficult: either you make it hard and expensive to acquire TLS certificates, and (2) has value, or you don't, and it doesn't. I think pervasive encryption is way more important than site owner identity validation. And I don't think the value of an EV cert was even all that high back when browsers called them out in their UI. There are lots of examples of people trivially managing to get an EV cert from somewhere, with their locally-registered "Stripe, LLC" or whatever in the "validated" company name field of their cert.
> Pavlov-trained everyone to dumb-click through 'this page is not secure' warnings
Not sure what that has to do with this. That was more of a problem back when we didn't have Let's Encrypt, so lots of people were using self-signed certs, or let their certs expire and didn't fix it, or whatever. These days I expect certificate warnings are fairly rare, and so users might actually start paying attention to them again.
> SNI exists and even without it anything not on CDN is blocked very easily
ESNI also exists, and while not being available everywhere, it'll get there. But this is a bizarre complaint, as it's entirely trivial to block traffic when there's no TLS at all.
Browser vendors removed the extra UI around EV certs not because certs in general are easier to get, but because the identity "guarantee" afforded to EV certs was fairly easy to spoof. EV certs still exist, and you can foolishly pay for one if you want. Free ACME-provided certs has nothing to do with this.
This characterization in on the same level of sophistication as "the Internet is just a series of pipes". Every transit station has the opportunity to read or even tamper with the bytes on an unencrypted http connection. That's not just your ISP, it also includes the ISP's backbone provider, the backbone peering provider, your country's Internet Exchange, the Internet Exchange in the country of the website, the website's peering partner, and the website's hosting partner.
Some of those parties may be the same, and some parties I have not mentioned for brevity. To take just one example: there is only one direct link between Europe and South America. Most traffic between those continents goes via Amsterdam (NL) and New Jersey (US) to Barranquilla (CO), or via Sines (PT) to Fortaleza (BR). Or if the packets are feeling adventurous today, they might go through Italy, Singapore, California and Chile, with optional transit layovers in Saudi Arabia, Pakistan, Thailand or China.
Main point being: as a user, you have no control over the routing of your Internet traffic. The traffic also doesn't follow geographic rules, they follow peering cost. You can't even be sure that traffic between you and a website in your country stays inside that country.
And already has! ISPs used to inject ads into unencrypted connections: https://www.infoworld.com/article/2241797/code-injection-new...
Cybercriminals work at scale. The opinion you shared here is why Google, Microsoft, and Amazon are so easy to use for cybercrime. It's incredibly easy to hide bad behavior in cheap, disposable attempts on free accounts.
Cost virtually eliminates abuse. Bad actors are fronting effort and ideally small amounts of money to effectively bet on a high return. You make the cost to attempt high, it isn't worth it. Apart from some high profile blogs demonstrating the risk, EV certs have to my knowledge never been used maliciously, and hiding them from the browser bar just makes useful, high quality data about the trustworthiness of a site buried behind hidden menus.
However, in a security context "takes some effort" is far better than "takes no effort".
If CAA records (with DNSSEC) were used to reject certificates from the wrong issuer, we might even be able to get to "though very imperfect, takes a considerable amount of effort".
DANE is supposed to be the solution to this problem but it's absolutely awful to use and will lead to even more fragile infrastructure than we currently have with TLS certs (and also ultimately depends on DNSSEC). HPKP was the non-DNS solution but it was removed because it suffered from an even worse form of fragility that could lock out domains for years.
More than one thing can be true at the same time.
Yes, vastly increasing the traffic that is encrypted is a great thing for many reasons.
Simultaneously, it is also true that if a public CA-issued (as opposed to unknown me self signing) certificate is effectively required, that sure provides a handy hammer to shut something down. History tells us that whenever such handy hammers get built, they inevitably get abused.
In practice this means you have to consider the possibility that anyone on the entire internet can inspect your traffic. Traffic from your home in Seattle to Google's west coast data center? For all you know it could be going via Moscow.
Wrong. Company names are not guaranteed to be unique per-country.
The main issue you are missing is that putting undeserved trust in things like DV / EV flags greatly increase the value of such attacks. If users are trained to blindly trust that shiny green bar, the odd attacker will be able to walk away with an absolute fortune. Nobody will be suspicious about that page, because Green Bar. Why is the "bank" asking odd questions? Who cares, it had a Green Bar, so it must be legitimate.
Why bother with hundreds of small attacks when one big one will make you rich?
Like, you need to realize most phishing scammers are perfectly happy to make a bankofamericaaa.glitch.me page, because it's free, and without any good indicators for the legitimate bank to use like EV, really doesn't look that much different to the nontechnical customer than bankofamerica.com.
Maybe a more relatable scenario for you - it was only a few years ago that you could turn cable modems into promiscuous mode to see ALL PLAIN TEXT TRAFFIC of the people living in your street!
So, if you you still think encryption isn't needed for the average person - what's your gmail username and password?