15 ms·
Let's Encrypt root certificate trusted by Mozilla
- UnoriginalGuy 10y agoJust to be clear, this is important because eventually Let's Encrypt wants to no longer have to cross-sign their certificates for them to be considered valid. For that to happen they have to be added as a trusted CA in most major platforms (and Firefox which has their own CA store for some reason).
- jayflux 10y agoWould be interesting if there's a roadmap for who's left to be added to. I saw their aim is to migrate to their own root CA by the end of the year, but i wonder if this is realistically feasible?
- jaas 10y agoWe have also applied to the Apple, Microsoft, Oracle, Blackberry, and Google root programs. We had hoped to be accepted by all major programs by the end of this year, but even if that happened it would still take years for our root to sufficiently propagate.
- gkanai 10y ago> (and Firefox which has their own CA store for some reason). Firefox has it's own CA store because it's built for all 3 major (desktop) platforms. OSX and Windows have their own but Linux does not and uses Mozilla's.
- mynameisvlad 10y agoDoes Chrome provide its own CA store on Linux? It's also built for all 3 major desktop platforms but uses the OSX and Windows stores.
- gsnedders 10y agoThey just use a copy of Mozilla's one on Linux. Of course, distro packages of Chromium if they use the system NSS library may well use some system CA store.
- mook 10y agoMore importantly, Firefox has its own CA store because it was derived from Netscape, which had its own because they came up with the whole SSL thing way back. The subsystem is still called NSS, even though there's been backbend changes (pkix).
- kibwen 10y agoHave any other browsers announced their intent to do the same?
- heinrich5991 10y agoOther browsers do not have their own certificate stores but use those provided by the OS. Next interesting things are whether Windows and Mac OS X add the root certificate. I think Linux distributions tend to follow Mozilla's trust.
- mtgx 10y agoSince Google is already cracking down on OEMs modifying the Android root store for various countries, I think it would make sense for Chrome to have the same root store as Android does (especially in light of Lenovo's Superfish, Dell's eDellroot and so on).
- gsnedders 10y agoChrome already has the ability to de-trust things trusted at the OS level, FWIW. It might be interesting to see how many OEM-installed certificates they could de-trust by default.
- aroch 10y agoRE: Linux, You're correct, they're provided through the ca-certificates package[0]: "It includes, among others, certificate authorities used by the Debian infrastructure and those shipped with Mozilla's browsers. " RE: OSX - If you can get into Mozilla's trust stores, it's the same steps (and pro forma, more or less) [1] [0] https://packages.debian.org/wheezy/ca-certificates https://packages.debian.org/wheezy/ca-certificates [1] https://www.apple.com/certificateauthority/ca_program.html https://www.apple.com/certificateauthority/ca_program.html
- gsnedders 10y agoHistorically ca-certificates has included certificates that Mozilla had declined to include due to lack of audits, most notably the SPI root certificate because some Debian infrastructure relied upon it. They've also generally been relatively slow in removing root certificates after they're removed upstream, despite any removal essentially being a security issue. I think as of six months or so ago there is no longer any difference between ca-certificates and upstream.
- cmdrfred 10y agoI think comodo had the idea to be added as a root CA by Mozilla first. They should sue.
- 0xmohit 10y agoComodo also had the idea to trademark "Let's Encrypt" [0]. [0] https://letsencrypt.org/2016/06/23/defending-our-brand.html https://letsencrypt.org/2016/06/23/defending-our-brand.html
- polysaturate 10y agoThat's the joke
- wlesieutre 10y agoMore specifically, Comodo's defense of that included "We did a 30 day free SSL certificate first! Let's Encrypt is copying our business model!" The free certificate they were referring to was a time-limited free trial that you could use once and then start paying for.
- __jal 10y agoThey are truly amazing. Every time I think they've scraped the bottom of either the incompetence or the sleaze barrels, Comodo manages to get even worse. One of their sales droids hassled me a while back with some deeply slimy tactics, so I started grilling him about this and the various hacks they've had. Flat out lied about ever having had unauthorized certs made, and claimed he'd never heard of LE, but he just knew they'd never do that, and I must have bad information. (The first part of the second part I can believe.) Who knows, maybe Comodo could come back after some strategic executive-ectomies. Microsoft seems to be trying hard to rejoin the ranks of the not-outstandingly-terrible. But as of now, I have serious doubts I'd ever choose their services over someone more trustworthy, like, say, Bernie Madoff.
- gruez 10y ago
- chillydawg 10y agoExcellent news. The more trust the better. It's still no good using LE for API endpoints as many client libs (java, etc) don't trust it or it's cross-signer.
- ge0rg 10y agoIt is a good practice to use key / certificate pinning for your API endpoints anyway, and there you can provide any trusted KeyStore you want...
- pfg 10y agoThe most recent Java update added the DST (IdenTrust) root certificate. Of course it's going to take a while until that version is widely deployed, but this gives vendors the option to tell API consumers to just update Java (as opposed to manually modifying the key store), so that should help with adoption.
- teach 10y agoOh, that's excellent news!
- niij 10y agoJava 8u101 added trust for their cross signer, IdenTrust. https://bugs.openjdk.java.net/browse/JDK-8154757 https://bugs.openjdk.java.net/browse/JDK-8154757
- 0xmohit 10y agoIt's about time that HN switches to Let's Encrypt.
- chc 10y agoWhy? Is there something wrong with HN's current cert?
- 0xmohit 10y agoSuffices to say that it has been issued by Comodo.
- Nullabillity 10y ago1. Funds Comodo's shady behaviour, and arguably increases their brand recognition 2. Ridiculous expiry time
- 0xmohit 10y ago> Ridiculous expiry time Looks like a cert is similar to a President. Should be around for 5 years.
- stephenr 10y agoYou know that Comodo won't issue a refund if they decide not to use them any more right? They've paid the money. I can understand moving things to show support for better alternatives but let's not kid ourselves - moving now or moving when the cert expires, doesn't hurt Comodo any differently.
- bhhaskin 10y agoLet's Encrypt is pretty great, but if you have the money get a paid SSL. Not all SSL certs are created equal.
- koolba 10y ago> Let's Encrypt is pretty great, but if you have the money get a paid SSL. Not all SSL certs are created equal. Say what? Besides the faux security of the green bar for an EV cert, what's the difference between a LetsEncrypt and a paid one? (non-EV)
- xutopia 10y agoThe one thing stopping adoption for a lot of people is wilcard support. https://community.letsencrypt.org/t/please-support-wildcard-certificates/258 https://community.letsencrypt.org/t/please-support-wildcard-...
- dimino 10y agoI feel like this is one of those "faster horse" moments. You don't need a wildcard cert! Just get certs for each one of your subdomains, even internally.
- xutopia 10y agoI get where you're coming from but I depend on a subdomain for each one of my user signups and we're talking thousands. LE limits us to 100 subdomains on a single cert and would require a restart for each signup. I'll definitely be using LE for other projects but many use cases require wildcards.
- tracker1 10y agoAnd most of those wildcard use cases are business sites/applications that can/should just buy a wildcard cert. That said, I wouldn't mind seeing better support for the freedns/dyndns subdomains.
- eeeeeeeeeeeee 10y agoSome people are using dynamic dns records, so the list of subdomains is not static. This is where wildcard ssl is perfect. And what if you have hundreds or thousands of subdomains? It unnecessarily bulks up your configuration as you have to do a cert per site instead of a single one.
- devy 10y agoIt's been discussed in details here the reason why they don't support wildcard: "doing domain validation for wildcard certificates is not currently in the ACME spec because it's a hard problem."[1] LetsEncrypt CA allows Subject Alternative Names (SAN), the true need for an unlimited sub-domains TLS cert vs. a SAN TLS cert is minimum, given Certbot's automation capability. [1]: https://github.com/certbot/certbot/issues/66#issuecomment-160487331 https://github.com/certbot/certbot/issues/66#issuecomment-16...
- wiradikusuma 10y agoQuestion: any possible case of bad apples that make let's encrypt suddenly lose their trust? Eg bcoz it's free, it's used by "bad guys" just like .info tld.
- hughes 10y agowhat's wrong with the .info tld?
- chc 10y agoPeople distrust it because it's 99% spam?
- duskwuff 10y ago.info is really not all that bad, in the grand scheme of things. .biz is much worse (and has been since its launch in 2001!), and some of the new TLDs like .top and .xyz have been abused pretty heavily as well.
- tracker1 10y agoI actually like the .info tld... perfect domain for "informational" websites, wikis etc. Though there may be better ones with the recent goldrush of new tlds available. Though I also liked .io, and felt the higher pricing and harder registration kept a lot of the squatters away.
- yacn 10y agoDo you mean .tk? AFAIK .info costs money.
- gtk40 10y agoYes, but it's usually cheaper.
- jerf 10y agoThe purpose of Let's Encrypt, and the SSL certificate infrastructure in general, isn't to prevent "bad guys" from getting certificates. It's to ensure that if you own the box, the certificate verifies that a web client is speaking directly to that box with nothing in between. (Or more generally, directly to an authorized end point by the owner of that DNS entry. Authority can be delegated.) In other words, it keeps bad guys out of the middle, not the end point. That's all SSL can do, even if it works perfectly. Bad guys will still own end points, in both the conventional sense of the word own and the pwning sense of the word own. SSL can not (directly) do much about that. If you speak SSL to a bad actor, well, there aren't any other actors between you and the bad actor, but you're still speaking on an encrypted, authenticated channel to a bad actor. This is in contrast to the DNS infrastructure in which it is sensible for a TLD owner to attempt to prevent "people they don't want on their TLD" (more generally than "bad guys" since a lot of the restrictions enforced are far beyond that).
- lllorddino 10y agoHacker News should switch from Comodo to Let's Encrypt. Scumbags attempted to trademark Let's Encrypt. https://letsencrypt.org/2016/06/23/defending-our-brand.html https://letsencrypt.org/2016/06/23/defending-our-brand.html
- theandrewbailey 10y agoHN uses ycombinator's wildcard certificate, and it's not up until August 2019. It's likely that they don't want to go through the trouble until it's really needed.
- Walkman 10y agoWith Let's Encrypt, the trouble became "Whoaaa I just ran a command and everything works like magic!"
- dyladan 10y agoSometimes magic isn't a good thing, especially when you're operating a service used by as many people as hn daily. Magic means things happened that I didn't explicitly instruct.
- mholt 10y agoThat's called automation and it's a good thing.
- erikpukinskis 10y agoAutomation is not a universal good, it can be useful and it can be detrimental. Like all tools it should be used with care.
- ChristianBundy 10y agoExactly. Value is a vector: automation increases the magnitude, but the direction depends on exactly what is being automated (and how reliable it is).
- mholt 10y agoLet's Encrypt's blog post: https://letsencrypt.org/2016/08/05/le-root-to-be-trusted-by-mozilla.html https://letsencrypt.org/2016/08/05/le-root-to-be-trusted-by-...
- discreditable 10y agoI find it a little hilarious that the cert for the Test URL, https://helloworld.letsencrypt.org https://helloworld.letsencrypt.org is a 90-day certificate that is expired as of a long time ago. https://i.imgur.com/1bQLHuF.png https://i.imgur.com/1bQLHuF.png
- joecool1029 10y agoIf you read through the linked bug inside the bug, I think they had to do that as part of the review process. They had to wait 90 days after the initial cert was issued to have it reviewed. Since LE only issues 90 day certs this means they'd have to review an expired one. The whole Mozilla CA review process is a little crazy and the thread talks about ways they could reform it in the future. (overview: https://wiki.mozilla.org/CA https://wiki.mozilla.org/CA )
- discreditable 10y agoIt's funny because LE's stance has been that 90-day certs are cool because it is easy to automate renewal. However, no one bothered to set up renewal on their example server.
- dijit 10y agoif you read the posters comment he explains why, they had to leave the original cert with the hash in place while the CA was being validated, despite it being expired.
- pfg 10y agoYou're missing the point here. helloworld.letsencrypt.org was used as the Test URL in the root inclusion ticket[1]. The certificate was renewed multiple times[2], but the expired certificate had to be restored at one point in order to satisfy some of the requirements of the root inclusion process. [1]: https://bugzilla.mozilla.org/show_bug.cgi?id=1204656 https://bugzilla.mozilla.org/show_bug.cgi?id=1204656 [2]: https://crt.sh/?q=helloworld.letsencrypt.org https://crt.sh/?q=helloworld.letsencrypt.org
- 10y ago