The TLS 1.2 Deadline is Looming, Do You Have Your Act Together?:
In the pantheon of security configuration duties for organizations running internet assets, maintaining the latest TLS encryption protocols to keep the cryptographic apparatus at full strength is one of the most fundamental. TLS provides cover for the most sensitive personal and financial information that moves across the internet. As experts in measuring and monitoring third-party risk, RiskRecon and the data scientists from Cyentia Institute recently published a new report that leveraged unique scan data from millions of web servers around the world, via the RiskRecon platform, to see where the rollout of TLS 1.2[*] is going smoothly and where it is meeting resistance.
Together with its precursor SSL, TLS has long been in the crosshairs of both attackers and security researchers who understand that a weak or non-existent deployment of the protocol makes it trivial enough to carry out man-in-the-middle and other attacks against the vulnerable target.
[...] Sectors such as Education (47%), Energy (40%), and Public Administration (37%) have struggled to implement TLS 1.2 protocols. This revelation led us to ask another question – “Are these hosts collecting and transmitting important information using vulnerable protocols?” The RiskRecon portal also determines web host value by examining whether a website collects and transmits important PII or credential information. If we restrict our view to just these high-value hosts, we can zero in on where the lack of TLS 1.2 represents a substantial risk: 1 in 10 organizations transmit private information over flawed protocols.
While our study found that this fundamental protocol lacks attention from some IT Security teams, it does not need any further introduction to those who would look to exploit any vulnerability in web communications. The clock is ticking to properly secure your lines of internet communications, standard bodies and web browsers have put out their warnings, and there is no time like to present to get up to speed.
[*] The latest version of TLS (Transport Layer Security) is 1.3; see RFC 8446.
(Score: 5, Insightful) by SomeGuy on Thursday July 16 2020, @08:44PM (13 children)
The moment everyone gets caught up to 1.3 it will be time for 1.4 then 2.0 then 2.1 gold edition, and if you don't keep up, no fucking internet access for you because your browser is more than five minutes old! To top things off these nazis will bitch at you until you do things their way. All this new stuff is sooo perfect and flawless and secure until it magically gets "cracked" then all of a sudden it's worthless, when it actually always was.
Just make it stop already.
(Score: 0) by Anonymous Coward on Thursday July 16 2020, @09:22PM (2 children)
That's when you call in an expart and let them decide whit's impooooortant. May I suggest Jack Daniel's or Jim Beam. If it's catastropic you'll need Tommy Chong.
(Score: 0) by Anonymous Coward on Friday July 17 2020, @01:11PM (1 child)
Hold up. Tommy's on-deck every day down here, we call him "Tommy the Troubleshooter", and we have a 40-minute Zoom call every day starting at 4:20.
Secondly, it's "catastrophic" not "catastropic".
(Score: 2) by Muad'Dave on Friday July 17 2020, @01:23PM
Nor is it cat-ass-trophy, but it's mighty close.
(Score: 2, Interesting) by Acabatag on Thursday July 16 2020, @10:13PM (2 children)
The important thing is that only endpoints should have access to anything about your online behavior. And that Big Data should rule over the endpoints. So only use the Chrome browser, and only navigate to Google approved sites.
(Score: 4, Informative) by mth on Friday July 17 2020, @12:16AM (1 child)
Let's Encrypt makes it easier than ever to run TLS on a small site using certbot. You can rent a VPS for less than $5 a month. Running a small independent site is not the issue. The problem with web centralization is real, but it's a problem of users not valuing their privacy and independence enough when they decide which sites and services they use.
(Score: 2) by ilsa on Friday July 17 2020, @07:59PM
While true, this is separate from the whole TLS thing, and I can sympathize with the GP. This never ending treadmill of incremental improvements is _exhausting_ to keep up with, and it feels like it's getting worse instead of better.
(Score: 2) by mth on Friday July 17 2020, @12:06AM (2 children)
The TLS 1.2 RFC is from August 2008. Security experts have been telling people to upgrade for years, but apparently a significant number of web server admins aren't listening. So I think the browser makers have been more than patient enough before pulling the plug on outdated protocols.
(Score: 2) by driverless on Friday July 17 2020, @03:10AM (1 child)
Because TLS 1.2 is so five minutes ago, the current trend is 1.3, which despite its name is a completely new protocol with almost nothing in common with the previous TLS 1.x versions past the client hello. Except that by the time we all move to 1.3, the current fashion will be 1.7, which is not really any better (or worse) than the five previous versions but will be using the latest trendy technology, which means you need to start again from scratch if you want to move to it.
(Score: 0) by Anonymous Coward on Friday July 17 2020, @02:44PM
So, are you arguing we need to give up on security or what else do you want to do about it?
(Score: 2) by mr_bad_influence on Friday July 17 2020, @01:18AM
My employer, a public university, depended on web payments for business. I was responsible for all web payments and for us to be in compliance with PCI standards TLS 1.2 was required. I've been retired for a couple years now, and implementing TLS 1.2 was one of the last things I did there before I left. Anyone in a similar position should have already upgraded.
The thing is, we always have to stay one step ahead of the competition by fixing any security issues that are always present or there won't be any confidence from folks using the web for payments.
(Score: 2) by driverless on Friday July 17 2020, @03:06AM
It was a weird sales pitch actually, "panic, panic, you need to take action now, it's almost too late, you're going to miss out" but then no link to their consulting services at the end to provide relief from the panic they've just stirred up. So why write it in the first place?
(Score: 2) by Opportunist on Friday July 17 2020, @10:21AM
Lemme guess, you're the first guy lamenting and crying how webpages are insecure when they are a victim to identity theft, right?
Security is an arms race between those that try to break into your systems and those trying to protect them. It is most likely never going to end.
But if you prefer to not join the race, that's fine by me. It actually increases the security of my system. Because crime is a business and businesses tend to go with the lowest expenses necessary to achieve the target profit. As long as there are systems that are less secure than mine and easier targets, criminals will target those systems instead of mine.
So... thank you, I guess?
(Score: 1, Flamebait) by darkfeline on Friday July 17 2020, @10:14PM
>Just make it stop already.
Might I suggest committing suicide? Maintaining existence is an endless battle. First world countries have been momentarily spoiled by their recent level of living quality, but having to fight for survival is the norm, not the other way around (COVID-19 is a friendly reminder from reality). Your immune system is fighting and learning every single moment; the day it starts faltering is the day you start succumbing to age-related immune deficiencies and diseases.
Just as attacks are being developed constantly, security and software is constantly being improved. System admins should be constantly upgrading their systems. If not, you're failing at your job and you will be figuratively killed by natural selection. If that seems like too much work, maybe you should take a look at some of the container hype and understand what problems it's trying to solve.
Join the SDF Public Access UNIX System today!
(Score: 2, Interesting) by Zinnia Zirconium on Thursday July 16 2020, @11:28PM (20 children)
I don't even run a web server that includes SSL.
(Score: 2) by takyon on Thursday July 16 2020, @11:31PM (7 children)
The GOOG will punish your site in the search rankings. Or maybe that is the point.
[SIG] 10/28/2017: Soylent Upgrade v14 [soylentnews.org]
(Score: 1) by Zinnia Zirconium on Thursday July 16 2020, @11:46PM
Nope. Google finds me just as well. I know because every time I search for anyone who scrapes Jango Radio like I do, all I ever find is me.
(Score: 3, Funny) by driverless on Friday July 17 2020, @03:16AM (5 children)
Same here, just type in "skinhead dwarfs having sex with goats that are on fire while singing Verdi translated into French" and there I am, top search result, like magic.
(Score: 1) by Zinnia Zirconium on Friday July 17 2020, @04:03AM (3 children)
Well that was a disappointing search result. I turned off safe search and all I found was an actual farm.
(Score: 2) by driverless on Friday July 17 2020, @04:07AM (1 child)
Dammit, I knew I should have enabled TLS 1.2, now Google has shitcanned my site.
Which is about house renovation on a budget BTW, but how are you going to be the first Google hit for that?
(Score: 1) by Zinnia Zirconium on Friday July 17 2020, @04:14AM
Oh here's a good one.
Haha! A goat period.
(Score: 2) by Opportunist on Friday July 17 2020, @10:09AM
Yes, yes, we have to have a front so the normies don't get anxious. Keep clicking, you'll find what you're looking for.
(Score: 2) by ilsa on Friday July 17 2020, @08:05PM
Well that is a remarkablely eclectic mix of results, including a google translation of "skinhead dwarfs having sex with goats that are on fire while singing Verdi " into french, some random questionable porn sites, and a Springer article titled "Decadence in Literature and Intellectual Debate since 1945"
Well done, you broke Google.
(Score: 0) by Anonymous Coward on Friday July 17 2020, @12:12AM
Not but 30 mins ago I had a power outage. Everything came right back.
Have to reboot ye ol cable modem. TLS 1.2 warning. Sigh.... IoT is crap.
You probably run a few TLS 1.0 1.1 items that you have no idea about.
(Score: 2) by Opportunist on Friday July 17 2020, @10:23AM (8 children)
Then how do I verify that I'm actually talking with your server and not some imposter?
(Score: 2, Touché) by Zinnia Zirconium on Friday July 17 2020, @09:38PM (7 children)
How do you know I'm not some imposter? You don't and it doesn't matter because I deal second hand in publicly available data nobody cares about. Maybe if the data were private or secret or sensitive or important in any way then I might bother with encryption.
.......and yeah I use SSH for a private channel into my own server but also I see my SSH log and honestly I gain more security from port knocking. I got no break in attempts in my log. Nobody so much as touches my SSH because nobody guesses my knock sequence.
(Score: 1, Redundant) by darkfeline on Friday July 17 2020, @10:18PM
Nobody cares about you server so it doesn't need encryption? Great, good for you. Unfortunately, most servers do handle important data so they need TLS.
Join the SDF Public Access UNIX System today!
(Score: 2) by Opportunist on Saturday July 18 2020, @08:54AM (5 children)
If I cannot verify that I'm talking to your server and not an impostor, how could I be certain that the data I get from you is actually accurate? Even if I trusted you that you provide genuine data, I could not verify that I actually got that data from you and not someone trying to forge it, either to plant false data or to slander you and make your visitors think you provide false data.
(Score: 1) by Zinnia Zirconium on Saturday July 18 2020, @09:07PM (4 children)
I serve data second hand and I can't be sure I get accurate data from my primary source who could suddenly decide they fukken hate me and start providing only me specifically with junk data which I have no way of verifying for accuracy and then my visitors would get junk despite my honest efforts. Encryption doesn't solve the garbage-in-garbage-out problem.
(Score: 2) by Opportunist on Sunday July 19 2020, @08:49AM (3 children)
Thanks for the warning, I'll find a different source then. If you cannot even verify your own source, why would I want to use you as one?
(Score: 1) by Zinnia Zirconium on Sunday July 19 2020, @05:57PM (2 children)
OK bye. Don't search my Jango index and don't use my YouTube proxy. Go directly to Jango and YouTube instead. I don't care.
Better yet use Pandora. Everybody uses Pandora because Pandora is Pandora and everybody uses Pandora. I still don't care.
If I get Let's Encrypt certificates then someone will tell me Let's Encrypt isn't trusty enough because Let's Encrypt only verifies domains and domains can be hijacked.
A couple of whois searches would turn up the fact that I'm using Freenom and No-IP for my DNS which means I don't own my domains. And do you remember that time in 2014 when No-IP was hijacked by Microsoft. I remember.
A couple of web searches would turn up the fact that I do all my development work at Wikidot which doesn't do HTTPS either. Oh no. How will I verify my source code at Wikidot is untampered before I deploy it to production.
Wikidot also happens to host the SCP Foundation. SCP Foundation doesn't encrypt. SCP Foundation isn't trusty enough. Oh no. How will you verify the data about every SCP is accurate and not intercepted by an imposter. It's the SCP Foundation. How much do you actually care.
SoylentNews does TLS 1.3. There's a silly lock icon on the browser and everything. That's good isn't it. Nope. SoylentNews uses Let's Encrypt. Let's Encrypt didn't verify ownership of SoylentNews. SoylentNews can't be trusted. SoylentNews could be hijacked by Microsoft right now. Don't ever trust SoylentNews.
Where's the massive troll spam campaign to convince SoylentNews to buy real certificates from a real certifying authority instead of cheap Let's Encrypt which could be so very hijacked by Microsoft right now. Seriously where is the troll spam.
Trolls gonna troll me no matter what I do. Isn't that right troll.
(Score: 2) by Opportunist on Sunday July 19 2020, @06:54PM (1 child)
Let's Encrypt does not verify ownership because that's not what a certificate is supposed to verify. A certificate does exactly what Let's Encrypt allows you to do: To verify whether the server you are connecting to is the server you may expect to reach at this address. Who that server belongs to is beyond the scope of a certificate, as is whether that server belongs to who you think it belongs to. If you connect to www.bankofmurrica.com and expect to do safe online banking because you see the lock symbol, you misunderstand the purpose of certificates.
A certificate makes no statement about the ownership of a system. Only that the system you are connecting to is the system that claims to belong to that domain name.
(Score: 2, Interesting) by Zinnia Zirconium on Monday July 20 2020, @03:13AM
Uh. No. Let's Encrypt doesn't verify ownership because Let's Encrypt is cheap. Ain't nobody at Let's Encrypt wanna gawk at a notarized photo of my government issued photo ID to prove I am who I say I am. That would take time and effort and somebody would want to get paid to do the work and it would raise the price of the certificate above free.
Let's Encrypt does only so much work as can be easily automated for free: challenge the HTTP server at a DNS domain name which I specify to produce a fukken stupid response. And by fukken stupid I mean "respond to this HTTP request by copying the request into the response."
I got Let's Encrypt to issue a certificate for my YouTube proxy which was the most challenging of my servers because my YouTube proxy is an HTTP server in a bash script. So wow I had to write two lines of code to pass the fukken stupid challenge that Let's Encrypt claims is proof enough that I'm me. But what if I'm not me. What if I'm some DNS hijacker who hijacked my domain. I am using No-IP.
Did I mention No-IP got DNS hijacked by Microsoft a few years ago. So now every troll says everybody should use Let's Encrypt and everybody knows every troll says everybody should use Let's Encrypt including every DNS hijacker. So now this year when somebody like Microsoft wants to hijack everybody at someplace like No-IP all they gotta do is take the extra step of renewing all the Let's Encrypt certificates for all the hijacked domains which the hijackers legitimately control according to Let's Encrypt fukken stupid challenge response shht.
Bam. Every browser shows the fukken lock icon and everybody trusts they connected to the server they expected and nobody notices Let's Encrypt is even more dangerous than not encrypting at all. At least when not encrypting everybody knows not to do stupid shht like type passwords and credit card numbers into a song search form.
So when is SoylentNews gonna get DNS hijacked and someone collects a nice collection of reusable passwords. Or is SoylentNews already hijacked. SoylentNews does use Let's Encrypt which just screams unnoticeable hijack.
See I don't need encryption. I don't accept passwords and I don't accept credit cards. I'm not a business and I'm not a bank.
I might actually go ahead and finish setting up socat with Let's Encrypt certificates and put socat in front of my HTTP servers for that warm fuzzy HTTPS feeling. But I would do it just for the technical challenge. It's all fukken pointless.
(Score: 0) by Anonymous Coward on Friday July 17 2020, @02:46PM (1 child)
ISPs and government agencies would like to thank you for all of that browsing data and a free 0-day delivery vector.
(Score: 1) by Zinnia Zirconium on Saturday July 18 2020, @04:47AM
Welllllll actually this might be a compelling reason to consider encryption just to stop shhtty ISPs from unhelpfully rewriting my shht.
Maybe I'll think about thinking about Lets Encrust. First I gotta convince Lets Encrust shirtbot to spit out just a certificate without trying to unhelpfully reconfigure my servers. Second I gotta prove I control my servers but I insist on running nonstandard web servers and one of my HTTP servers is a bash script I wrote myself. It doesn't serve files at all but I guess I could add a few lines to my bash server to generate a response proving to Lets Encrust that I control it. And then third my actual HTTPS server would be socat or something that I can configure entirely with command line options.
Yeahhhhhh maybe I should. As soon as I start caring. Sometime between now or never.
(Score: 0) by Anonymous Coward on Friday July 17 2020, @01:30AM (3 children)
Test your site from https://www.ssllabs.com/ssltest/ [ssllabs.com]
It will let you know if your TLS is up to snuff and lots of other info.
While TLS 1.2 and above are likely supported by most web servers. But not everyone has disabled older versions (TLS 1.0/1.1).
Check it out.
(Score: 0, Disagree) by Anonymous Coward on Friday July 17 2020, @07:18AM (2 children)
For practically all websites disabling TLS 1.0 and/or 1.1 on the server side is a pretty silly thing to do. The protocol has version negotiation (and the browser people have fixed their silly problems in this regard) so if you and the client support higher versions that will be used. Disabling the older protocols just means that your website will work with less existing software for no real reason.
(Score: 2) by Opportunist on Friday July 17 2020, @10:25AM
This is true until I sit in between you and the server and tell both of you to use an insecure version because I pretend that the other side doesn't support better security.
Downgrade attacks [wikipedia.org] are a thing, ya know?
(Score: 0) by Anonymous Coward on Friday July 17 2020, @11:17AM
I hope your clients read this. Because they should fire you.
TLS 1.0/1.1 have multiple exploitable and exploited vulnerabilities:
https://www.acunetix.com/blog/articles/tls-vulnerabilities-attacks-final-part/ [acunetix.com]
Yes, web browsers will negotiate the protocol level, and that's an attack vector [wikipedia.org]. By negotiating protocol and/or cipher "downgrade," miscreants can exploit the known vulnerabilities.
Which is why it's not only important to require a sane (TLS 1.2 or greater) transport, it's also important to disable weak ciphers (many of which are enabled by default on many web servers).
So, no. TLS 1.0/1.1 need to terminated with extreme prejudice. I know, I know, TLS 1.2 is far too new to put into production. After all, the RFC was only just published twelve years ago.
You're talking nonsense. But go ahead and run whatever you want. But if you really believe the bullshit you spewed, you might want to educate yourself.
I won't hold my breath.
(Score: 3, Interesting) by hendrikboom on Friday July 17 2020, @03:18AM (16 children)
I just use http on my server. No certificate issues.
-- hendrik
(Score: 2) by Opportunist on Friday July 17 2020, @10:27AM (13 children)
How do I make sure I talk to your server and not some imposter?
(Score: 2) by coolgopher on Friday July 17 2020, @10:39AM (8 children)
How do you make sure you talk to the right https server and not some imposter injected by your ISP via DNS and route hijacking?
(Score: 0) by Anonymous Coward on Friday July 17 2020, @11:19AM (4 children)
RFC 1149 [ietf.org] FTW!
(Score: 2) by hendrikboom on Friday July 17 2020, @11:29AM (2 children)
I see. The pigeons know. And a pigeon hunter isn't likely to pick up enough packets to form a meaningful message?
(Score: 0) by Anonymous Coward on Saturday July 18 2020, @06:17AM (1 child)
That's why it's necessary to use strong encryption with large keys. ;)
(Score: 0) by Anonymous Coward on Monday July 20 2020, @01:42PM
Large keys?
We talking a good 20g here, 100g, or some half kilogram whopper from the medieval ages?
(Score: 3, Insightful) by coolgopher on Friday July 17 2020, @11:32AM
And here I was expecting a link to RFC3514 [ietf.org] instead.
(Score: 2) by hendrikboom on Friday July 17 2020, @11:27AM
Because my browser does do https.
(Score: 2) by Opportunist on Friday July 17 2020, @12:32PM (1 child)
By checking whether the certificate matches the page. As long as you didn't somehow manage to inject your certificates into my browser store, it's pretty trivial to verify whether the certificate presented belongs to the server.
(Score: 2) by coolgopher on Monday July 20 2020, @01:05AM
Yeah fair point, it does take quite a bit more effort to pervert the initial set of top level certificates. Then again, time and time again we find about CAs that have been handing out certs willy-nilly >.<
(Score: 2) by hendrikboom on Friday July 17 2020, @11:26AM (3 children)
You don't. Nor does my server care who you are.
(Score: 2) by Opportunist on Friday July 17 2020, @12:34PM (1 child)
Ok, then I hope I'll never have to talk to your server, because I'd really love to know whether the person (or server) I'm talking to is actually who they claim to be.
I mean, you (hopefully...) don't respond to calls like "hi, this is your IT department, we'd need your username and password to figure out a problem we have with it..."
(Score: 2) by hendrikboom on Friday July 17 2020, @04:38PM
No. I don't.
-- hendrik
(Score: 0) by Anonymous Coward on Friday July 17 2020, @05:10PM
lahu zahur!
(Score: 2) by leon_the_cat on Saturday July 18 2020, @05:44AM (1 child)
you have a real sexy server
(Score: 1) by Zinnia Zirconium on Saturday July 18 2020, @09:19PM
I prefer minimalist design so I like how it's plain HTML. (Technically it's MathML unnecessarily.)