4 ms·

Kazakhstan's National HTTPS Certificate Is Cancelled

85 points by wyldfire 7y ago

slenk 7y ago
Was posted earlier today by a very new 'tsarka' account: https://news.ycombinator.com/item?id=20632373 https://news.ycombinator.com/item?id=20632373 Have you heard of Tsarka before? They don't have any press releases from when it started? As others have echoed - could this just be a state sponsored actor?
- steve19 7y ago
  As I wrote in a comment on that post: "I can't help but think if the government were wanting to spread positive propoganda about these tests, your post is exactly what they would write: "everyone wins and everyone is happy with our results of MTIMing the entire country". Who knows how much information the state hoovered up during this "pilot".
  - slenk 7y ago
    I should have quoted you directly - I just didn't look deep enough.
  - cm2187 7y ago
    Starting by all passwords...
- askar_yu 7y ago
  By looking at their clients, if not sponsored, at least clearly affiliated. What junk load of nonsense press release. My view of what happened is some clowns at the top wanted a proper, full interception of traffic. Besides the attention in international media it received, people had real issues accessing the internet properly (those who installed). So they simply could not execute their MITM properly. Putting aside privacy, etc. I expected simple operational ineptitude, which is what happened (gladly) and they dropped saying "it was a mere test". PS, in 2011 the government tried something similar with Google's traffic and after Google started redirecting the KZ traffic they backed, saying some similar B.S. like "it was a test".
marksomnian 7y ago
Discussion in mozilla.dev.security.policy (mailing list to discuss TLS implementation in Firefox/NSS): https://groups.google.com/forum/#!topic/mozilla.dev.security.policy/wnuKAhACo3E https://groups.google.com/forum/#!topic/mozilla.dev.security...
- walrus01 7y ago
  https://bugzilla.mozilla.org/show_bug.cgi?id=1567114 https://bugzilla.mozilla.org/show_bug.cgi?id=1567114
ilaksh 7y ago
How do we know that other super CAs aren't being used for MITM by other countries?
- southerntofu 7y ago
  Some are. The French and Tunisian governments have been caught emitting fake TLS certificates through their national CA in the past. There's probably many more who don't get caught because their attacks are more targeted. The whole Public Key Infrastructure is fully insecure because of this property that any root CA can emit a internet-wide valid certificate for any domain. That's why people work on DANE for DNS and Tor onion services (among others) to reduce the attack surface to the name resolution level, which in the case of onion services equals breaking an ed25519 keypair because the onion domain name is tied to the key. We need to ditch PKI as soon as possible!
  - hedora 7y ago
    PKI weaknesses extend far beyond state sponsored attacks. Things like Let’s Encrypt allow anyone that can intercept 100% of the traffic to your site to get a signed cert for your domain with no further validation. Unless your site is directly peered to multiple backbones, you are trusting at least one service provider not to man in the middle your traffic. Cert pinning sort of protects against this, but it is widely considered impractical. The certificate transparency efforts also sort of protect against this, though I think there are still CA’s that don’t participate. As long as browsers accept certs from those CA’s, certificate transparency doesn’t solve the problem either. (Also, what percentage of sites actually monitor to see if there are malicious certificates for their domains?)
    - southerntofu 7y ago
      > Things like Let’s Encrypt Any CA will do that. Paying a hundred bucks for a certificate will has never stopped a malicious actor, therefore Let's Encrypt does not introduce any new concern. > Cert pinning sort of protects against this DANE was designed for this. Since DNS is always a SPOF on the web, registering TLS public keys in DNS reduces the attack surface (instead of introducing a new channel for transport security). > The certificate transparency efforts also sort of protect against this That's bullshit. Nothing prevents a CA to emit certificates without declaring them. We should not rely on those big corporations for our security: our interests are FAR from aligned.
andrerm 7y ago
> the tests were completed, all the tasks set during the pilot were successfully solved So, it was a pilot and it succeeded. Now what?
mikece 7y ago
Don’t believe it — Uzbek propaganda!!!
southerntofu 7y ago
Am i the only one reading this article as a harsh criticism veiled to pass censorship? I'm not sure, though, but this sentence sounds absurdly false: « Everyone got their own: we got the free Internet, the government got an instrument for fighting digital weapons. »

85 points by wyldfire 7y ago

Was posted earlier today by a very new 'tsarka' account: https://news.ycombinator.com/item?id=20632373 https://news.ycombinator.com/item?id=20632373 Have you heard of Tsarka before? They don't have any press releases from when it started? As others have echoed - could this just be a state sponsored actor?

steve19 7y ago
As I wrote in a comment on that post: "I can't help but think if the government were wanting to spread positive propoganda about these tests, your post is exactly what they would write: "everyone wins and everyone is happy with our results of MTIMing the entire country". Who knows how much information the state hoovered up during this "pilot".
- slenk 7y ago
  I should have quoted you directly - I just didn't look deep enough.
- cm2187 7y ago
  Starting by all passwords...
askar_yu 7y ago
By looking at their clients, if not sponsored, at least clearly affiliated. What junk load of nonsense press release. My view of what happened is some clowns at the top wanted a proper, full interception of traffic. Besides the attention in international media it received, people had real issues accessing the internet properly (those who installed). So they simply could not execute their MITM properly. Putting aside privacy, etc. I expected simple operational ineptitude, which is what happened (gladly) and they dropped saying "it was a mere test". PS, in 2011 the government tried something similar with Google's traffic and after Google started redirecting the KZ traffic they backed, saying some similar B.S. like "it was a test".

marksomnian 7y ago

Discussion in mozilla.dev.security.policy (mailing list to discuss TLS implementation in Firefox/NSS): https://groups.google.com/forum/#!topic/mozilla.dev.security.policy/wnuKAhACo3E https://groups.google.com/forum/#!topic/mozilla.dev.security...

walrus01 7y ago
https://bugzilla.mozilla.org/show_bug.cgi?id=1567114 https://bugzilla.mozilla.org/show_bug.cgi?id=1567114

ilaksh 7y ago

How do we know that other super CAs aren't being used for MITM by other countries?

southerntofu 7y ago
Some are. The French and Tunisian governments have been caught emitting fake TLS certificates through their national CA in the past. There's probably many more who don't get caught because their attacks are more targeted. The whole Public Key Infrastructure is fully insecure because of this property that any root CA can emit a internet-wide valid certificate for any domain. That's why people work on DANE for DNS and Tor onion services (among others) to reduce the attack surface to the name resolution level, which in the case of onion services equals breaking an ed25519 keypair because the onion domain name is tied to the key. We need to ditch PKI as soon as possible!
- hedora 7y ago
  PKI weaknesses extend far beyond state sponsored attacks. Things like Let’s Encrypt allow anyone that can intercept 100% of the traffic to your site to get a signed cert for your domain with no further validation. Unless your site is directly peered to multiple backbones, you are trusting at least one service provider not to man in the middle your traffic. Cert pinning sort of protects against this, but it is widely considered impractical. The certificate transparency efforts also sort of protect against this, though I think there are still CA’s that don’t participate. As long as browsers accept certs from those CA’s, certificate transparency doesn’t solve the problem either. (Also, what percentage of sites actually monitor to see if there are malicious certificates for their domains?)
  - southerntofu 7y ago
    > Things like Let’s Encrypt Any CA will do that. Paying a hundred bucks for a certificate will has never stopped a malicious actor, therefore Let's Encrypt does not introduce any new concern. > Cert pinning sort of protects against this DANE was designed for this. Since DNS is always a SPOF on the web, registering TLS public keys in DNS reduces the attack surface (instead of introducing a new channel for transport security). > The certificate transparency efforts also sort of protect against this That's bullshit. Nothing prevents a CA to emit certificates without declaring them. We should not rely on those big corporations for our security: our interests are FAR from aligned.

andrerm 7y ago

> the tests were completed, all the tasks set during the pilot were successfully solved So, it was a pilot and it succeeded. Now what?

mikece 7y ago

Don’t believe it — Uzbek propaganda!!!

southerntofu 7y ago

Am i the only one reading this article as a harsh criticism veiled to pass censorship? I'm not sure, though, but this sentence sounds absurdly false: « Everyone got their own: we got the free Internet, the government got an instrument for fighting digital weapons. »