India's National Informatics Centre Forged Google SSL Certificates 107
NotInHere (3654617) writes As Google writes on its Online Security Blog, the National Informatics Centre of India (NIC) used its intermediate CA certificate, issued by Indian CCA, to issue several unauthorized certificates for Google domains, allowing it to do Man in the middle attacks. Possible impact however is limited, as, according to Google, the root certificates for the CA were only installed on Windows, which Firefox doesn't use — and for the Chrom{e,ium} browser, the CA for important Google domains is pinned to the Google CA. According to its website, the NIC CA has suspended certificate issuance, and according to Google, its root certificates were revoked by Indian CCA.
Repercussions? (Score:3, Interesting)
Will there be any repercussions for this?
The National Informatics Centre of India did abuse something.
Will the National Informatics Centre of India be able to continue with such abuses and do this again in the future?
Or will they lose this ability?
What will happen now?
They have shown that they can not be trusted. They must lose the power to do this.
Pull someones certificates or kill some CA. Someone needs to suffer because of this.
Re: (Score:1)
They must lose the power to do this.
No one can be trusted. The system/infrastructure must be designed to take into account untrustworthiness of all parties involved. WoT [wikipedia.org].
Re: (Score:2)
If you think that might work, then keep learning. The botnets' "vote" only gets counted if someone decides to trust all of them. And if you can arrange that, then you don't need a botnet, you just need one node.
All that matters is how your fake node (or web of fake nodes) is connected to the victim.
Re: (Score:2)
Oops, didn't realize we were talking about something like that.
That plugin is a kind of neat idea (I approve) but it's very poorly named and doesn't seem to have anything in common with a real "web of trust." I'd probably be madder about the atrocious name if I didn't happen to like the plugin.
That gives me an idea: I should make a program for X11 users, where the five hundredth and ninth time someone opens a new window, it generates a PDF containing an extravagant statement of the accomplishment. Then I
Re:Repercussions? (Score:5, Insightful)
This yet again highlights that the three-party trust system is broken.
There are ways around it, but there is no great solution - only workarounds.
Re:Repercussions? (Score:4, Interesting)
Re: (Score:2)
Re: (Score:2)
Never do business with anyone who outsources customer service at all. The 'representative' only has the power to read from the flip chart. They absolutely do not have the authority to fix the problem and they do not know who does or how to contact them.
Re: (Score:2)
Re: (Score:2)
They do know who the boss is and who his boss is. They know who signs their paychecks. They may not tell, but they know.
With internal CS, there is at least a chance that it is supposed to be more than an impenetrable barrier between the customer and someone with authority.
Re: (Score:2)
Outsourced customer service is generally paid by the call. This means that the ideal call is a short one where the customer is satisfied enough not to raise a ruckus, but will run into problems and have to call back. Even if there is any way to pass feedback to the software vendor, it isn't in the customer service's interest to provide it, as clarifying a confusing thing in the software could lead to serious loss of revenue. That also means that the software vendor doesn't have the information to either
Re: (Score:3)
What happens now is that there's an investigation. Depending on the outcome the CA may be revoked for good, or merely forced to reissue lots of certificates. The deciding factor is the reason for the screwup - for instance they may have got hacked, rather than been actively corrupt. In that case Microsoft will have to decide if they have patche
Re: (Score:3)
Expecting CA's to be able to reliably fight off professional hackers from dozens of governments and never ever fail is likely an impossible standard to ever meet.
Yet that is exactly what they are supposed to do. Its not even really that hard.
Every CA hack to date has been preventable as was the fault of the CA simply not putting the required effort into doing their job or being flat out malicious. Stop trying to make it out like its an uber hard job, its not.
Re: (Score:2)
Who do they think they are? (Score:5, Funny)
Re: (Score:1)
Who in the world do you think gathers intelligence?
Only the NSA?
Need a bridge? I have one for sale.
Re: (Score:2)
Re: (Score:1)
Strawman? Not so much.
So does the NIC have a legal mission to gather intelligence? Does forging certificates constitute legitimate intelligence collection?
Who can say? Do you have any thoughts on the matter?
Re: (Score:2)
Name one intelligence agency that doesn't use other government agencies to assist its endeavours.
Re: (Score:2)
Re: (Score:2)
All countries keep an eye on their neighbours, just like all people keep a general awareness of their surroundings. All countries don't tap the phones of their neighbours's leaders, or install malware on equipment sold to them, or even spies over. Morals aside, taking hostile action tends to backfire, as the US is learning. Reputation is a resource, and
Re: (Score:2)
Typical (Score:1)
Good old Indian "ethics".
Re: (Score:3, Insightful)
Re: (Score:1)
All about trust (Score:2)
Re: (Score:3)
Re:All about trust (Score:5, Insightful)
And, really, if the US is saying it's their right to tap into anything they want to ... how is it different when India does it?
India already forced BlackBerry to allow them to access BBM and the like.
Uncle Sam is causing as much disruption to US businesses abroad as anything, because people are realizing that American companies are effectively just extensions of the US spy apparatus -- because the PATRIOT act means they can demand whatever data they have, and you more or less have to assume they're doing it and being prevented from telling you.
Which means Indians are already being spied on by (at least) their own government AND the USA.
Do you expect there to be sympathy for an American company when a foreign government taps into them? Because I hear an awful lot of people saying they think it's perfectly OK when the US does it to foreigners.
Re: (Score:2)
It's only fair that you either get to protest when every and any country pulls something like this, or not at all.
Re: (Score:2)
I don't disagree with you, but the hypocrisy of "but that's the job of the NSA" that I hear when someone points this out is maddening.
And one which was doing business in their country. Like it or not, Google in India is subject to India's laws.
How many corporations and people in fore
Re: (Score:2)
Still, Google may have a presence in India but it's not an Indian company,
Re: (Score:1)
The only thing that will come out of this is lack of trust for som
Re:All about trust (Score:5, Insightful)
As a US resident, I'd be perfectly content to see the heads of various rights-invading federal agencies put away in prison.
So no, it's not ok. Not for the US, not for India.
Re: (Score:2)
Agreed. They might or might not put the bodies in prison with the heads, I'm good with it either way. :-)
Re: (Score:1)
No one is going accept a cert from them again.
Yeah. Just like no one trusts Comodo CA. Oh wait.
Re: (Score:2)
Re: (Score:1)
Remember DigiNotar ?
They went bankrupt because nobody trusted them anymore.
Re: (Score:2)
Re: (Score:2)
The whole point of issuing certs is to be a trusted third party. No one is going accept a cert from them again.
Sounds like what we need is a cert-issuing protocol based on Bitcoin security. Everyone (plus or minus epsilon) trusts that Bitcoins can't be forged.
Re: (Score:2)
Pseudonyms exist to protect people from the rabid - like yourself.
Think about the stupidity of comparing the establishment of a pseudonym to posting your SSN? LOL.
Re: (Score:1)
Funny, I looked up "Assmasher" in the White Pages and various international name lookup services and didn't get a single hit. It's almost as if you're hiding your identity no differently than the very ACs that you proclaim to want to be abolished. Man up and give us all your personal details or STFU.
Re: (Score:1)
Wow, I guess the guys who built /. who thought AC should stand for "Anonymous Coward" didn't know that "Desler" knows best and that and AC and a registered user are exactly the same thing. Wonder why they bothered with creating the AC system? Idiots. Really. I mean, they should have just asked you obviously.
Ignoring the rest of the stupidity of what you posted, maybe you could come to realize that the difference between AC and a registered user is that registered users can develop a reputation for their
Re: (Score:1)
How does having a registered account mean anything? You can register one with a throwaway email account. Plus many registered people do use AC from time to time.
Re: (Score:2)
Because it's a pain to do so. It helps cut down on the DB anonymous posting. You can quickly discern if they're schills, flametards, et cetera.
I agree, I post on occasion as AC when I'm on another device, and like I said, I never had any problem with people posting AC until the past few years when people seem to be using it to simply spam /. with total garbage, or hatred, et cetera.
Re: (Score:1)
Because it's a pain to do so
Yeah clicking a button and typing a couple dozen characters is sooo hard. Registration takes less than 5 minutes in total.
Re: (Score:2)
Doesn't it require a valid e-mail address and confirmation first? It certainly used to.
Re: (Score:2)
5 minutes is a lot of time for the people who go around spouting hatred and ugliness all over internet forums. This is why the don't register, because it's not worth the effort - especially when they get banned - especially if that ban is by IP.
Re: (Score:1)
No it's not.
Old-media contact info (Score:2)
Re: (Score:2)
I was gonna say set your preferences to -5 AC posts, but I can't find the setting at the moment - did they get rid of it for beta? Somebody probably can post the link to the scoring prefs.
Re: (Score:2)
Somebody probably can post the link to the scoring prefs.
https://slashdot.org/users.pl?op=editcomm [slashdot.org]
Or you can click on one of the "edit" links in the score details window.
Re: (Score:2)
The difference between India and some other countries is that India is 2nd-rated enough to be caught immediately when they do something like this. That makes them more stupid, but less of a threat than, say, the US.
So SSL is nothing more than an honor system? (Score:4, Insightful)
Re:So SSL is nothing more than an honor system? (Score:5, Insightful)
Re: (Score:2)
That's a cop-out, though. Yes, there is always an element of trust in whatever you do. That's unavoidable, though it's smart to minimize the amount of trust you must put in others. Taken to the extreme it's ludicrous, as you've pointed out. But, that doesn't mean that there's no merit in limiting the amount of trust you put in third parties. Just because you can't completely trust your OS or compiler, doesn't mean that you should throw the entire concept of limiting trust out the window. It's dishonest to s
Re: (Score:2)
You're just figuring this out? Have you been living under a rock for the past ~20 years or are you just incredibly naive?
Re: (Score:3)
This is nothing new.
And, let's face it, I bet the NSA et al have demanded more private keys be handed over to them than you'll ever know about. Where's your outrage over that?
The five eyes all use each other to spy on their own (and others) citizens, and share the information among themselves. Where's your outrage over that?
I see this as a symptom of a greater problem, but no different from what a bunch of other countries are already doing.
Until someone creates
Re: (Score:2)
Until someone creates a new encryption system which isn't susceptible to MITM attacks
Uh, some of the earliest encryption algorithms ever created are immune to MITM.
The core of the MITM issue is that anything sent over it could be intercepted or spoofed.
So ALL your communication must be encrypted.
All you need a pre-shared key to initiate the connection. Whether that's a password or a certificate or something else makes no difference. What matters is the pre-sharing. You have to fucking know and trust the source of that key. If you're just using a list of certs issued by people you don't
Re: (Score:2)
Yes, and they were built for communications between two parties, who knew they'd be communicating, and could exchange keys in advance.
Now, tell me one which is applicable to the problem of a large number of potential users, all unknown up front, and coming from random devices.
The problem with modern public key encryption (and its strength as well) is that you don't need to pre-exchange keys. But this opens you up to MITM attacks
Re: (Score:3)
Anybody that looked into the SSL certificate system has known that for a very long time. Quite a few people used to use self-signed certificates, as as least there somebody that bothered to find out could be sure it was secure.
I think the fundamental brokeness of the SSL certificate system is because of deep naivety with regard to the trustworthiness of governments and because of active sabotage of by said governments way back. I hope at least that issue is fixed after Snowden. Governments are even more evi
Re: (Score:3)
SSL goes beyond the naivety of government trust. It also suffers from what amounts to a global namespace/trust/etc issue.
Any CA can issue a certificate for any domain, a domain generally can only have one certificate, and the trusted CA list is managed by the browser, not the user.
So, if you trust your government (naievely), and distrust everybody else, it won't work. Your browser will constantly be wanting to add CAs you don't trust, and might not include ones you trust. Then, if you drop a bunch of CAs
Re: (Score:2)
Indeed. That is why I wrote "governments" as in the sum of all of them. One corrupt one is enough to break things.
Re: (Score:2)
There are two TLS extensions that fix these problems - one is including your certificate fingerprint in DNS and the other is multiple signatures. Both have good standards and the industry is painfully slow to adopt them.
Re: (Score:2)
It's a shame that browsers have such freakouts over self signed certs, because there is really little difference between them and officially signed certs. IMHO SSH did a better job of this by simply having you inspect the certs the first time you log on to a site and storing the result, only freaking out if the cert changes. It eliminates the complex chain of trust that in the end comes down to just trust
Re: (Score:2)
It's a shame that browsers have such freakouts over self signed certs, because there is really little difference between them and officially signed certs
Exactly. Especially since you can get a "real" cert from one of many, many, free cert signing services. What is the point?
Re: (Score:2)
Re: (Score:2)
That is an existing capability within the SSL process. NIC will be restricted to issuing certificates only for a set of domains that are specific to India. Just be careful if you want to have financial transactions over the Web with institutions based in India.
Re: (Score:2)
There are any number of proposals out there to replace or augment CA certificates for SSL purposes (the EFF has Sovereign Keys, there is the DANE proposal to store certificates in DNS with DNSSEC security and there are other proposals out there designed to make it much harder for these kinds of "bogus certificate" type attacks)
Why aren't any of these proposals actually gaining any traction?
Re: (Score:2)
It sounds like we need the ability to limit the scope of certificate authorities to signing for only certain domains.
http://tools.ietf.org/html/rfc... [ietf.org]
Internet Explorer IS vulnerable though (Score:2)
This is a big deal. If you use a browser on Windows that does NOT counter this, such as Internet Explorer, then you ARE vulnerable. I imagine Microsoft will come out with a special-purpose patch, but still, this is a pretty nasty issue.
Untrustworthy CAs have been a problem for a long time; we need mechanisms to address them. The terrible cert revocation system makes it even worse; you can't be sure that the certs are checked in many cases. Chrome's CRLSets are not the answer; they are not even the be
Isn't it time we apply name constraints? (Score:2)
I think intermediate CA certificates issued to certificate vendors, ISPs, governments, should all have name constraints so that they can be used to sign only certificates for an appropriate part of the namespace.
http://tools.ietf.org/html/rfc... [ietf.org]
Not a Problem with Mozilla-Based Applications (Score:5, Informative)
This is not a problem with Firefox, SeaMonkey, or other Mozilla-based applications. They use a certificate database separate from Microsoft's, a database that does not contain the certificate used in the forgery.
The certification authority at fault (NIC) has an open request to have its root certificate added to Mozilla's database. However, NIC has failed to respond to requests for further information, requested over a year ago by the Mozilla person who is in charge of the process of approving certificates. Furthermore, Mozilla persons -- both staff and users -- are aware of NIC's problem; some have suggested that NIC's request be rejected and NIC be permanently banned from the database.
To see the discussion, see https://bugzilla.mozilla.org/s... [mozilla.org].
Some certification authorities and some of their subscribers complain that Mozilla takes too long to approve root certificates and then to add those certificates to Mozilla's database. At least in this case, delay served to protect users. The delays are significantly caused by Mozilla's requirement for independent audit reports and for a period of public review and comment on each request. Hooray for Mozilla!!
Corruption is cultural (Score:2)