10 ms·
The HTTPS-Only Standard
- some_furry 12y agohttps://istlsfastyet.com/ https://istlsfastyet.com/ This is a step in the right direction. Worldwide HTTPS adoption makes the Internet a safer place for everyone.
- Someone1234 12y agoGreat. When that is done, then do email with is frankly a far more retractable problem in general (and a field where there is almost zero innovation or improvement, thanks to Microsoft (Exchange/Outlook/Outlook.com), Apple (Mail), and Google (Gmail)).
- bitJericho 12y agoWhat's wrong with email? It's a 40 year old standard that still works great.
- Someone1234 12y agoIt is insecure in so many ways...
- PaulHoule 12y agoOne thing that amazes me is that I get a huge volume of spam emails that claim to be from financial institutions. I use gmail for two reasons: (i) deliverability to mailing lists I need to be on and (2) other mail programs don't filter that junk out. In 2015 it should be impossible to send a fake email from chase.com
- bitJericho 12y agoIf chase.com uses dkim and spf records and if your mail server is properly configured then it is indeed impossible.
- me1010 12y agoand, of course, DMARC -- so rejection policies can be set. PGP FTW! Now if only STEED would be implemented... http://g10code.com/docs/steed-usable-e2ee.pdf http://g10code.com/docs/steed-usable-e2ee.pdf But, unfortunately, even mail from a properly configured mail server on properly protected domain will still end up in gmail users' spam boxes by default. Domain and server rep systems are a bear to work with.
- bitJericho 12y agoIn what ways? Mail servers can be configured to require only encrypted connections. You can also use GPG to encrypt your messages.
- jcrites 12y agoI believe he's referring to the fact that a lot of email communication is still plaintext: https://www.google.com/transparencyreport/saferemail/ https://www.google.com/transparencyreport/saferemail/ Of the communication that's nominally protected by TLS, a lot of it can probably be trivially broken by an active MITM attack that either (1) prevents a connection from being upgraded to TLS, in which case it typically remains plaintext (see STARTTLS), or (2) establishes the connection under a self-signed certificate, which will suffice since many email systems do not perform certificate path validation.
- mauricemir 12y agoMm you mean RFC 822 - well thats the down side of internet standards x.400 had a lot of nice security features but the cheap and cheerful SMTP email standards won that war a long time ago. I used to do X.400 and x.500 for a Large Telco back in the Day. It's a trade off between ease of use and cost - and is a tech example of Gresham's Law
- kstrauser 12y agoOutstanding! This is wonderful news. I've heard people ask whether a given site or protocol really needs to be secure but I hold the opposite opinion: everything should be encrypted unless there's a specific and compelling reason otherwise. I'm thrilled that major organizations are coming to the same conclusion.
- JoshTriplett 12y agoThe most important line in this standard: > All browsing activity should be considered private and sensitive.
- skrowl 12y agoClearly this government office isn't talking to some of the other ones, right?
- higherpurpose 12y agoI think this is happening mainly because of some newly hired employees that are pushing for strong security for government websites.
- sliverstorm 12y agoThe Fed employs 2.7M people. I don't really understand the people who imagine it's one great big well-oiled machine with everybody in agreement, on the same page of the same playbook.
- shit_parade 12y agothat's 2.7M people willing to work for and assist an entity that openly tortures and murders without consequence, so yeah seems like they are on the same page to me.
- rhino369 12y agoI bet you give that very organization thousand of dollars a year.
- shit_parade 12y agoYeah, they would put me in prison otherwise right, so by threat of force I must materially support a group which kills and tortures, wonderful.
- 12y ago
- bcg1 12y agoHey gov, how about promoting a transparent decentralized system for certificate signing that doesn't require paying a vig every year to a corporation that can easily be leaned on by not-so-well-meaning authorities? Oh wait, nevermind... I just realized something.
- onezeno 12y agoWhat did you realize?
- fweespeech 12y agoI believe he meant to imply that the government wants the ability to lean on the SSL cert companies.
- ori_b 12y agoWhat's the alternative? A web of trust where thousands of individuals can be leaned on by not-so-well-meaning authorities? Trusting the individuals or groups who validate identity at some level seems to be unavoidable.
- michaelt 12y agoThe best alternative I've heard so far is actually a patch: We keep the CAs, but they publish signed, append-only lists containing every cert they've issued. No cert is trusted if it isn't in such a list, and the lists are mirrored (and cross-signed) by a bunch of trustworthy authorities, hashes written into the blockchain, and so on. Then any CA who issues a cert for www.google.com will also have to publish irrevocable proof that they done fucked up; and when their CA status is revoked it's possible to grandfather in certs they've already issued to avoid breaking lots of sites. Of course, it won't solve the problem of CAs unjustifiably charging $$$ for certain types of cert - or the problem of the dubious authentication done for affordable domain-validated certs.
- throwaway7767 12y ago> Then any CA who issues a cert for www.google.com will also have to publish irrevocable proof that they done fucked up; and when their CA status is revoked it's possible to grandfather in certs they've already issued to avoid breaking lots of sites. I wish. Even in cases today where we have conclusive proof of CAs willfully issuing fraudulent certificates, they get a pass. TrustWave still has a valid CA cert even after having been caught with their hand in the cookie jar. Oddly, DigiNotar got the death sentence. The only lesson I can see in this is that incompetence is inexcusable, but willfully subverting the CA system is A-OK.
- PaulHoule 12y agoI guess this means you'll visit a government web site and 2 times out of 10 gets a popup that says the certificate is out of date.
- Millennium 12y agoThe HTTPS-only folks mean well, and I support it as a stopgap solution, but it is useful only in that it can probably be implemented more quickly than IPSec-everywhere (or, if IPSec proves to be unsuitable, then some successor standard with the same goal of encrypting all traffic). The latter, however, should be preferred as a permanent solution. The Web is by no means the only part of the Internet that needs to be secured.
- konklone 12y agoIt's worth noting that HTTP(S) has broadened outside of the web, in the sense of web browsers. Most native mobile apps, and lots of APIs used by desktop apps, etc., all use HTTP to get their job done. Definitely doesn't cover everything, but I think it's fair to say that HTTP is basically the protocol of our lives right now.
- cesarb 12y agoIPSEC and HTTPS work at different levels. With IPSEC, your computer can be sure it's talking to the computer at 198.51.100.1 and not to any other computer. With HTTPS, your browser can be sure it's talking to www.example.gov and not to any other web server. Both work equally well against passive eavesdroppers, but they authenticate different things and so will work differently against active attackers.
- azernik 12y agoIn other words, IPSEC is useless without DNSSEC, and that isn't getting universal adoption anytime soon.
- cesarb 12y agoOh, no, it's useful for what it's designed for: to protect communication between two computers. If I have IPSEC protecting the connection between my desktop and my internal DNS server, and between my desktop and my database server, I know that connection to my database server is protected by IPSEC. It doesn't protect the mapping between a computer name and a IP address, but that's not its job.
- u23KDd23 12y agoAn HTTPS only standard is not going to mitigate MITM attacks which is invariably the biggest issue regardless of perpetrators although this is clearly a step in the right direction.
- rando3826 12y agoProper HTTPS is secure at MITM attacks. HTTPS can be MITMed if the certs or local systems have been breached. But the MITM is just a byproduct of a different failure. Absolutely any secure communication protocol has to be honored by both parties for it to be secure.
- u23KDd23 12y agoI think preventing MITMs when you don't have control of the local systems or network architecture is important.
- rando3826 12y agoBut it's impossible. If I'm logging your keystrokes, game over. Unless you add some firmware to do dsa encryption in your brain, and become very fast at typing cyphertext.
- kysol 12y agoThere's something about their favicon being the default green lock ( https://https.cio.gov/assets/favicon.ico https://https.cio.gov/assets/favicon.ico ) that unsettles me. It feels like a social engineering trick.
- LeoPanthera 12y agoSuddenly, Safari's behaviour of hiding the favicon unless you click the address, makes sense.
- kysol 12y agoFrom what I was reading, the removal of the favicon in Safari was more just a UI redesign decision to remove "clutter". Personally I don't find a 16x16 icon too intrusive, but hey, what ever floats their boat. I was hoping that it was as you said, and was to prevent maliciously designed favicons from tricking users on plaintext sites (where the protocol had been stripped by the UI) into thinking they were on a secure site. I don't use Safari, so I don't know how they render their address bar.
- konklone 12y agoThat's an interesting point. I'll be straight, it's lifted right from https://istlsfastyet.com https://istlsfastyet.com. And cio.gov is in the HSTS preload list [1] so (once the list makes it into stable channels) the chances of the domain being downgraded to plaintext are pretty low. But I had not thought of that angle. Hmm. [1] https://18f.gsa.gov/2015/02/09/the-first-gov-domains-hardcoded-into-your-browser-as-all-https/ https://18f.gsa.gov/2015/02/09/the-first-gov-domains-hardcod...
- kysol 12y ago..and people complain about transparency :) I wasn't calling it a social engineering trick, more that it just felt like one. To the average person they wouldn't second guess the icon. To those who believe in HTTPSAllTheThings, we question anything out of the ordinary.. and that little padlock shouldn't appear in the tab. As I said, it just felt weird, sort of the same feeling you get when you go to Apple or YouTube and there's a warning on the lock icon. You just want to hit the back button almost instantly fearing something dodgy is happening.
- jvehent 12y agoThey also use a pretty strong TLS conf https://gist.github.com/jvehent/98fdfef139b6237f7bf5 https://gist.github.com/jvehent/98fdfef139b6237f7bf5 What strikes me is that their conf is inaccessible to old client (no sslv3, no sha-1 cert). Honest question: does the government think it's ok to break website access from ancient clients (XPSP2 IE6)? Or will they be forced to enable legacy ciphers when the first citizens complain? I'm genuinely interested, because the topic of TLS modernity is a hard problem to solve. If that was easy, google.com wouldn't be using RC4-MD5 and a sha-1 cert...
- 001spartan 12y agoWith the number of attacks against SSLv3 and/or RC4, I think it's a good idea. Especially given that it's a government website. With a target that large, it's a safe bet that, at any given moment, someone is trying something against that site. And really, who's going to access cio.gov from an XP box? Heh.
- konklone 12y agoJust speaking for myself here, but what makes sense in terms of legacy client support for e.g. taxes or health care may be different than what you need for a set of technical resources.
- pcunite 12y agoI want to select who my browser should trust. Kinda like how I choose which NTP server to get my time sync from. Make it a popup the first time a browser starts.
- tem5050 12y agoNice! They even go in to detail about cipher and protocol choices
- zx2c4 12y ago> "the financial cost of procuring a certificate" Doesn't the government have a few of its own CAs installed in every browser? Can't they procure their own certificates for free?