4 ms·

Ask HN: I need your help. Trying to develop secure and anonymous email

4 points by simonszki 13y ago

http://securmail.launchrock.com/ Hi guys, I have a problem and I need your help: me and a couple of my friends decided to create a web service for free anonymous and secure mail after hearing about the NSA debacle(Lavabit closing his doors). We created a launch page to get email for people to try out the service as beta tester. Just go check it out and if your are interested, signup. If you have suggestions, don't hesitate. We are working hard to build something great and useful for everyone. If you could share the service to help us out too, it would be awesome. Thanks Really, it all comes down to how you want to protect your data. Now that we know Google's been working with the NSA and other email services. There aren't many options left for secure email. Ever since news of PRISM broke, there's been a lot of confusion and denial about exactly how the NSA is getting the data from the companies that have been collecting it. Now Google's fessed up to the details, and it's unsurprisingly simple: by FTP or even by hand. With what we are building, it would be basically impossible to do. Built on AES-256 Encryption and Host Proof Hosting, ensuring that your sensitive data is never accessible to anyone but you. We never have your un-encrypted data. Even if our servers were hacked your emails would be safe and when you delete an email, it is not recoverable(US DOD, 7 passes). We do not log any traffic or session data. For this reason, we’re unable to identify any user of our service. Here is the launch page: http://securmail.launchrock.com/

malandrew 13y ago
This is essential reading/watching here: http://ritter.vg/blog-deanonymizing_amm.html http://ritter.vg/blog-deanonymizing_amm.html https://www.youtube.com/watch?v=_Tj6c2Ikq_E https://www.youtube.com/watch?v=_Tj6c2Ikq_E
cyphunk 13y ago
I do not know if it is possible but you should tell people a bit more about your service on your webpage. I could find no details about how any of your claims are implemented.
- simonszki 13y ago
  We are thinking about implementing a blog so that we can share part of the coding process so that anyone can have a general idea of how our servers are secured and emails encrypted, etc.
cjbprime 13y ago
With respect, I think it's likely that you don't understand the threat model here and are making dangerous promises of security that you can't keep. If the NSA shows up and demands that you help them access a user's data and not tell anyone about it, they won't merely be asking for a copy of your hard disk. They could ask that you modify your javascript (since you're a webapp, presumably you're doing client-side cryptography in JS, which is a terrible idea but let's ignore that for now) to send the user's key to you, or otherwise transfer cleartext back to you. I don't mean to be harsh, but I think it would be better not to release anything at all than to release a product that can't live up to its privacy promises. If you really want to innovate in this field, please do it with open source code and peer review from expert cryptographers.
- simonszki 13y ago
  It think you have a great idea about peer review of the code. We already have contacts with coders in the online password management business. The only thing that might be a problem is the involvement of the government. Like with Lavabit, they force them to reveal there SSL master key so that they could intercept every communications. We are currently trying to figure out if we could host our server in switzerland or something.(it fixed the problem for several security company we know.) I think our product present a huge leap in privacy for email compared to google and the likes who allow free access to data. Keep that in mind. http://gizmodo.com/how-google-gives-your-information-to-the-nsa-512840958/all http://gizmodo.com/how-google-gives-your-information-to-the-...
  - cjbprime 13y ago
    I would not trust your service more than Gmail. With Gmail, I can be fairly sure that the use of encryption is competent and that they've invested in entire teams of people dedicated to monitoring the service for hacking attempts -- I might have to worry about them handing data over to the NSA, but I don't have to worry about them handing data over to a random black hat who decides to modify the JS you deploy. I also trust that Google's legal team is capable of some level of argument with the NSA, rather than just folding to them the first time actual money would be required to pay a legal team. I don't see how hosting a server in Switzerland helps you when the NSA is at your door, telling them that you must help them and you must not tell anyone about it. The weak point is you, not your servers.
professorTuring 13y ago
A question: have you hired a security consultant to review your procedures and processes? or, are you yourself a security expert? I mean, do you know that a secure mail is much more than encryption and public-private key? You really need to have your service security reviewed.
SamReidHughes 13y ago
If you found a proof that P != NP, you'd want to communicate it with the rest of the world. In doing so, you'd want to portray yourself in such a way that people don't think you're a crank. You'd speak in the standard style of prose that related CS papers use, you'd avoid the signs of a wrong proof as described at http://www.scottaaronson.com/blog/?p=458 http://www.scottaaronson.com/blog/?p=458 or explain why they don't apply here. You'd portray yourself as somebody that knows what they're doing. When it comes to the topic of secure and anonymous email, you haven't done that. You've done the opposite. There's about a 1 in 1000 chance that somebody writing the post you've made here is in any position to be writing this software. In particular, you don't predict questions people would have and address them beforehand. Your post leaves so much room for doubt, doubt that you should be aware of and have resolved yourself. You've also failed to address cjbprime's question in a way that suggests you have a realistic view of how the world works -- you didn't mention having a team member in Switzerland who's the only one that can access the server. That would be an essential fact, but you've omitted it. Whether it's true or false, you didn't know it was essential, and now it seems like you don't know what you're doing.

4 points by simonszki 13y ago

malandrew 13y ago

This is essential reading/watching here: http://ritter.vg/blog-deanonymizing_amm.html http://ritter.vg/blog-deanonymizing_amm.html https://www.youtube.com/watch?v=_Tj6c2Ikq_E https://www.youtube.com/watch?v=_Tj6c2Ikq_E

cyphunk 13y ago

I do not know if it is possible but you should tell people a bit more about your service on your webpage. I could find no details about how any of your claims are implemented.

simonszki 13y ago
We are thinking about implementing a blog so that we can share part of the coding process so that anyone can have a general idea of how our servers are secured and emails encrypted, etc.

cjbprime 13y ago

With respect, I think it's likely that you don't understand the threat model here and are making dangerous promises of security that you can't keep. If the NSA shows up and demands that you help them access a user's data and not tell anyone about it, they won't merely be asking for a copy of your hard disk. They could ask that you modify your javascript (since you're a webapp, presumably you're doing client-side cryptography in JS, which is a terrible idea but let's ignore that for now) to send the user's key to you, or otherwise transfer cleartext back to you. I don't mean to be harsh, but I think it would be better not to release anything at all than to release a product that can't live up to its privacy promises. If you really want to innovate in this field, please do it with open source code and peer review from expert cryptographers.

simonszki 13y ago
It think you have a great idea about peer review of the code. We already have contacts with coders in the online password management business. The only thing that might be a problem is the involvement of the government. Like with Lavabit, they force them to reveal there SSL master key so that they could intercept every communications. We are currently trying to figure out if we could host our server in switzerland or something.(it fixed the problem for several security company we know.) I think our product present a huge leap in privacy for email compared to google and the likes who allow free access to data. Keep that in mind. http://gizmodo.com/how-google-gives-your-information-to-the-nsa-512840958/all http://gizmodo.com/how-google-gives-your-information-to-the-...
- cjbprime 13y ago
  I would not trust your service more than Gmail. With Gmail, I can be fairly sure that the use of encryption is competent and that they've invested in entire teams of people dedicated to monitoring the service for hacking attempts -- I might have to worry about them handing data over to the NSA, but I don't have to worry about them handing data over to a random black hat who decides to modify the JS you deploy. I also trust that Google's legal team is capable of some level of argument with the NSA, rather than just folding to them the first time actual money would be required to pay a legal team. I don't see how hosting a server in Switzerland helps you when the NSA is at your door, telling them that you must help them and you must not tell anyone about it. The weak point is you, not your servers.

professorTuring 13y ago

A question: have you hired a security consultant to review your procedures and processes? or, are you yourself a security expert? I mean, do you know that a secure mail is much more than encryption and public-private key? You really need to have your service security reviewed.

SamReidHughes 13y ago

If you found a proof that P != NP, you'd want to communicate it with the rest of the world. In doing so, you'd want to portray yourself in such a way that people don't think you're a crank. You'd speak in the standard style of prose that related CS papers use, you'd avoid the signs of a wrong proof as described at http://www.scottaaronson.com/blog/?p=458 http://www.scottaaronson.com/blog/?p=458 or explain why they don't apply here. You'd portray yourself as somebody that knows what they're doing. When it comes to the topic of secure and anonymous email, you haven't done that. You've done the opposite. There's about a 1 in 1000 chance that somebody writing the post you've made here is in any position to be writing this software. In particular, you don't predict questions people would have and address them beforehand. Your post leaves so much room for doubt, doubt that you should be aware of and have resolved yourself. You've also failed to address cjbprime's question in a way that suggests you have a realistic view of how the world works -- you didn't mention having a team member in Switzerland who's the only one that can access the server. That would be an essential fact, but you've omitted it. Whether it's true or false, you didn't know it was essential, and now it seems like you don't know what you're doing.