On 2013/07/30 (Jul), at 3:40 PM, Steve Dougherty wrote:

>> ... it is easier to block than Tor connections?
> 
> I'd say so - the list of opennet seed nodes is public,[1] so blocking
> access to those IPs would mean no opennet.

I'm sure Tor does something similar, it would be interesting to know how Tor 
nodes bootstrap.

> [0] https://blog.torproject.org/blog/hidden-services-need-some-love
> [1] https://downloads.freenetproject.org/alpha/opennet/seednodes.fref

I'm not trying to stir anything up, and maybe I'm getting a bit paranoid myself 
after watching the Security Now episode on SSL... but it's worth noting & 
thinking about this as an attack vector. Don't we have a list of attack vectors 
somewhere?

If, for example, we don't bundle some seed nodes with the distribution, then 
any a business-level attacker (i.e. that does "SSL Inspection") could just make 
that url return an empty file, or a big-isp-level attacker can make sure you 
get a list of only Sybil nodes by conjuring up a ssl certificate. Both would 
have a different private key fingerprint, as I understand it.

--
Robert Hailey

Reply via email to