▲ 44 points
back
22 comments
"RandomDNS aims to improve the security, privacy and anonymity of DNSCrypt."
"It can randomize the server choice at runtime and can rotate it frequently."
Re privacy, randomizing the server choice reduces privacy. It means they all end up with your DNS queries instead of just one of them. This is exactly why Tor has entry guards.
I can't see any evidence of this application increasing security or "anonymity" either.
What about a system that uses a different server for each separate lookup?
Still less preferable to just using one server
Yeah, it should just be DNSSEC-over-Tor.
Why not just forward DNSCrypt queries over Tor?
What's the point? DNSCrypt is already encrypted and Tor offers zero privacy but it does offer the FBI helpfully monitoring to de-anonymize Tor clients and servers. Using a VPN and DNSCrypt together would far better. Plus, Tor adds unnecessary latency (unusable for daily use) and suspicion, rendering it's benefit nearly nullified apart from paedophiles or people trading drugs.
i2p and others seem to be headed in the right direction as to future of distributed, overlay networks.
> DNSCrypt is already encrypted and Tor offers zero privacy but it does offer the FBI helpfully monitoring to de-anonymize Tor clients and servers.
This is pure FUD.
> Using a VPN and DNSCrypt together would far better.
https://gist.github.com/joepie91/5a9909939e6ce7d09e29
Don't use VPN services.
We'll, I trust my VPN equally or more then my ISP. In Germany, there has been a lot of debate about the so called data retention law. While I assume that my VPN does the same, the data is a lot less valuable because I'm well over 4,000 miles.
> the data is a lot less valuable because I'm well over 4,000 miles.
That is a very dangerous assumption to be making.
Can't see the benefit when I have load-balancing and HA using dnsmasq fronting dnscrypt-proxy to 6 separate hosts.
My assumption was that this service would provide random responses to look ups.
Now that would be secure.
If only people stopped writing this stuff in javascript.
At least it's not written in Coffeescript and then further insulting you by calling it a JavaScript project despite being written in Coffeescript.
I've been foolish enough to have been quoted as saying "you can use JavaScript anywhere!" in the past (admittedly at the time, I was working on an embedded platform for TVs where apps were written in JS). Sadly it seems not everyone knows, just because you can doesn't mean you should.
I agree. It seems like a maintainability issue. I can't imagine anyone familiar enough with DNS to contribute has experience working with Node.
Maybe I'm wrong...
While I agree with the "why!? Node" sentiment of parent, you'd be wrong with this extension. Myself as well as a few friends (who have extensive history with DNS at both a sysadmin and developer level) disprove the theory that there's no overlap in the two skill sets.
I would never contribute to a JS project, but it is the most used language in the world. Which language would you expect to have more overlap than the most used language in the world?
It is NOT the most used: http://www.tiobe.com/tiobe_index
Java is number one, still. Like six years running at least. JavaScript is somewhere between 4-8 depending on the survey you look at.
yes. please stop the endless javascripting of the world..
like it or not, js is the programming language that runs on the most devices in the world.
While, I know this is flamebate:
This statement is factually incorrect. C is the most common higher level language across all platforms, especially when you include embedded devices. (I'm not considering Assembly as a higher level language)
mm good point. i stand corrected.