4 ms·

Couldn't this data be coming directly from these IM services' servers? I doubt any of them are E2EE.

Laforet 8y ago
It's possible but I really doubt it, considering the sheer volume of message being generated every moment. It would be much easier and less likely to have hiccups like these if they just give them server access on demand, rather than having a live data feed replicated to 18 separate locations. Moreover the tweets actually show names and addresses of various internet cafes in one of the pictures. This information should not be available if the data is coming from the service provider's backend. Contrary to popular opinion on HN and other forums in the Anglosphere, surveillance in China is a nuanced problem and tech companies are far from completely passive. Local police are often denied when they request information because technically they lack jurisdiction on companies located in a different town or province. As a result they resort to catch-all interception using devices like these makeshift DPI systems(there is no clear legal requirement, but internet cafe owners are easily coerced into installing them or risk having their business hut down over various infractions) and Stingray-like fake cell sites (a constant source of conflict with major telcos because they are often poorly installed and would interfere with normal cell sites). These ad-hoc schemes are usually put together by the lowest bidder, so they tend to be horribly inefficient and insecure like what's been shown in this case. But to local LE it's still preferable to going dark, not to mention there is usually little accountability when their system breaks.
- 6nf 8y ago
  It's rare to see comments like yours from people who know about the technical details of how the Chinese government manage the internet. Could you share some please? How often do people actually get into trouble for their internet usage?
  - Laforet 8y ago
    >the technical details of how the Chinese government manage the internet This is a very broad topic that require several books to cover, but I will try to outline the more prominent features in a short few paragraphs. Before I start, it's important to point out that in a authoritarian state like China, laws are often deliberately made very strict but only loosely enforced. In time, everyone is likely to have already committed some crime as a matter of necessity and thus nobody is safe from the law; the constant threat of law enforcement make people fear the authorities. Every website serving users in China needs to apply and hold an ICP license[0]. Without a license, no hosting company will do business with you and your domain name run the risk of getting blacklisted if you decide to host it on a foreign server instead. If you have an established online presence, it is in your best interest to avoid anything transgressions that might result in the loss of your license, including but not limited to: - Hosting illegal content - Not promptly deleting content at the request of authorities - Not complying with any of the cyber security laws Additional permits are required if the website hosts music/video, publish games or offer money handling services. These permits are often intentionally kept scarce to make the industries more restrictive and compliant. As for internet users, every service they use require an account tied to their real ID. This is usually done by registering a mobile phone number which is technically always tied to a real identity. Public internet access also require phone verification to access. Therefore anything posted to the internet can theoretically be traced back to its author. >How often do people actually get into trouble for their internet usage? This is a really tricky question. Because of all the issues I mentioned above, service providers are strongly incentivized to practice self censorship. Thus a lot of contentious stuff people say will disappear before the authorities gets involved, especially if they are on the open web. Over at the more closed platforms such as WeChat the standards are somewhat relaxed, but from time to time people do get in trouble for what they post. In any case it's hard to tell because the standards being used to judge your words are in a constant state of flux. What happens after that really depends on where you live. In large cities, the police have more real problems to deal with and are subject to more oversight, so they tend to turn a blind eye except for the most blatantly seditious messages. And even in the latter case most people are let go after a formal warning without much consequence. Small town cops, on the other hand, are a lot more eager to seek prosecution, especially if they are corrupt or have a personal vendetta against you. Conflict zones like Xinjiang are the worst because they are under effective martial law, so anything remotely out of the line will get a follow-up. [0]:https://en.wikipedia.org/wiki/ICP_license https://en.wikipedia.org/wiki/ICP_license
- peteretep 8y ago
  > I really doubt it, considering the sheer volume I feel like most people felt that way about the NSA's capabilities pre-Snowden
  - Laforet 8y ago
    Or did they? The global SIGNIT capacity of the NSA has been pretty well known before Snowden. The more revealing part of the Snowden leaks was that the NSA has been spying on US soil without clear authorisation. Personally speaking, Snowdon actually made me a lot less paranoid about the NSA: They are, like every other nation state backed blackhat, relys on having physical access and 0day exploits to do their bidding. No magical backdoor or quantum computers involved.
    - dTal 8y ago
      NSA spying on US soil was long known[1] (and of course "clear authorization" was never granted if you interpret that phrase to mean "public" and "accountable") I'm not sure where you got the notion that they need physical access to anything either. Apart from intercepting comms on the wire, they deliberately subvert crypto standards in such a way that only they have access [2]. That is as close to a "magical backdoor" as it's possible to get. [1] https://en.wikipedia.org/wiki/Room_641A https://en.wikipedia.org/wiki/Room_641A [2] https://www.theguardian.com/world/2013/sep/05/nsa-gchq-encryption-codes-security https://www.theguardian.com/world/2013/sep/05/nsa-gchq-encry...
      - Laforet 8y ago
        Well, Room 641A is a form of physical access. They are not yet capable of breaking into any commpany's network from outside. The slides in the guardian article is pretty vague. The only proven case of NSA inserting a backdoor would have been the DUAL_EC_DRBG algo, and people have been alleging that from the very beginning. IIRC, Bernstein went on further and suggested that the NIST ECC curves may be compromised, but that's far from proven. Most real life attacks we have seen such as logjam and beast arises from longstanding vulnerabilities that are in no way exclusive to the NSA, but they are probably in a better position to analyse past interceptions once an.exploit has been found.

It's possible but I really doubt it, considering the sheer volume of message being generated every moment. It would be much easier and less likely to have hiccups like these if they just give them server access on demand, rather than having a live data feed replicated to 18 separate locations. Moreover the tweets actually show names and addresses of various internet cafes in one of the pictures. This information should not be available if the data is coming from the service provider's backend. Contrary to popular opinion on HN and other forums in the Anglosphere, surveillance in China is a nuanced problem and tech companies are far from completely passive. Local police are often denied when they request information because technically they lack jurisdiction on companies located in a different town or province. As a result they resort to catch-all interception using devices like these makeshift DPI systems(there is no clear legal requirement, but internet cafe owners are easily coerced into installing them or risk having their business hut down over various infractions) and Stingray-like fake cell sites (a constant source of conflict with major telcos because they are often poorly installed and would interfere with normal cell sites). These ad-hoc schemes are usually put together by the lowest bidder, so they tend to be horribly inefficient and insecure like what's been shown in this case. But to local LE it's still preferable to going dark, not to mention there is usually little accountability when their system breaks.

6nf 8y ago
It's rare to see comments like yours from people who know about the technical details of how the Chinese government manage the internet. Could you share some please? How often do people actually get into trouble for their internet usage?
- Laforet 8y ago
  >the technical details of how the Chinese government manage the internet This is a very broad topic that require several books to cover, but I will try to outline the more prominent features in a short few paragraphs. Before I start, it's important to point out that in a authoritarian state like China, laws are often deliberately made very strict but only loosely enforced. In time, everyone is likely to have already committed some crime as a matter of necessity and thus nobody is safe from the law; the constant threat of law enforcement make people fear the authorities. Every website serving users in China needs to apply and hold an ICP license[0]. Without a license, no hosting company will do business with you and your domain name run the risk of getting blacklisted if you decide to host it on a foreign server instead. If you have an established online presence, it is in your best interest to avoid anything transgressions that might result in the loss of your license, including but not limited to: - Hosting illegal content - Not promptly deleting content at the request of authorities - Not complying with any of the cyber security laws Additional permits are required if the website hosts music/video, publish games or offer money handling services. These permits are often intentionally kept scarce to make the industries more restrictive and compliant. As for internet users, every service they use require an account tied to their real ID. This is usually done by registering a mobile phone number which is technically always tied to a real identity. Public internet access also require phone verification to access. Therefore anything posted to the internet can theoretically be traced back to its author. >How often do people actually get into trouble for their internet usage? This is a really tricky question. Because of all the issues I mentioned above, service providers are strongly incentivized to practice self censorship. Thus a lot of contentious stuff people say will disappear before the authorities gets involved, especially if they are on the open web. Over at the more closed platforms such as WeChat the standards are somewhat relaxed, but from time to time people do get in trouble for what they post. In any case it's hard to tell because the standards being used to judge your words are in a constant state of flux. What happens after that really depends on where you live. In large cities, the police have more real problems to deal with and are subject to more oversight, so they tend to turn a blind eye except for the most blatantly seditious messages. And even in the latter case most people are let go after a formal warning without much consequence. Small town cops, on the other hand, are a lot more eager to seek prosecution, especially if they are corrupt or have a personal vendetta against you. Conflict zones like Xinjiang are the worst because they are under effective martial law, so anything remotely out of the line will get a follow-up. [0]:https://en.wikipedia.org/wiki/ICP_license https://en.wikipedia.org/wiki/ICP_license
peteretep 8y ago
> I really doubt it, considering the sheer volume I feel like most people felt that way about the NSA's capabilities pre-Snowden
- Laforet 8y ago
  Or did they? The global SIGNIT capacity of the NSA has been pretty well known before Snowden. The more revealing part of the Snowden leaks was that the NSA has been spying on US soil without clear authorisation. Personally speaking, Snowdon actually made me a lot less paranoid about the NSA: They are, like every other nation state backed blackhat, relys on having physical access and 0day exploits to do their bidding. No magical backdoor or quantum computers involved.
  - dTal 8y ago
    NSA spying on US soil was long known[1] (and of course "clear authorization" was never granted if you interpret that phrase to mean "public" and "accountable") I'm not sure where you got the notion that they need physical access to anything either. Apart from intercepting comms on the wire, they deliberately subvert crypto standards in such a way that only they have access [2]. That is as close to a "magical backdoor" as it's possible to get. [1] https://en.wikipedia.org/wiki/Room_641A https://en.wikipedia.org/wiki/Room_641A [2] https://www.theguardian.com/world/2013/sep/05/nsa-gchq-encryption-codes-security https://www.theguardian.com/world/2013/sep/05/nsa-gchq-encry...
    - Laforet 8y ago
      Well, Room 641A is a form of physical access. They are not yet capable of breaking into any commpany's network from outside. The slides in the guardian article is pretty vague. The only proven case of NSA inserting a backdoor would have been the DUAL_EC_DRBG algo, and people have been alleging that from the very beginning. IIRC, Bernstein went on further and suggested that the NIST ECC curves may be compromised, but that's far from proven. Most real life attacks we have seen such as logjam and beast arises from longstanding vulnerabilities that are in no way exclusive to the NSA, but they are probably in a better position to analyse past interceptions once an.exploit has been found.