On 26.02.2016 11:54, Tim Wilson-Brown - teor wrote:
On 26 Feb 2016, at 11:52, Random Tor Node Operator <tor@unterderbruecke.de mailto:tor@unterderbruecke.de> wrote:
On 26.02.2016 05:15, torserver@datakanja.de mailto:torserver@datakanja.de wrote:
- Next, i noticed a frequent (daily) behavior of the Tor server dropping traffic to around zero. Inspecting this, let me to understand, my provider was disconnecting me and reassigning a new IP on a daily basis, which took some time to propagate. Even worse: It did not propagate on its own, i needed to restart the tor service to reinitialise...
Instead of a Tor Relay, you can operate a Tor Bridge, perhaps with obfs4. A regularly changing IP address is less of a problem for bridges. It may even be of advantage. Once its IP address gets blacklisted by adversarial actors, you already have a new one.
But how do users find that new address? (For some users, the bridge authority might tell them when provided with the bridge's fingerprint, but only if their other bridges work.)
Yes, that's the thing when being within a censored environment. A user needs to have at least one bridge or other way to find their way into the Tor network.
Imagine a world where all Tor bridges have a static IP. After a while, the adversarial actors would have a complete list of all bridges and successfully block all access to the Tor network. (Except when a brand new bridge pops up) Every bridge would only be useful until first detection by the adversary.
But with bridges with dynamic IP, the adversary has to play whack-a-mole with the bridges and chances are, the user within the censored environment will have *some* moles (bridges) which haven't been whacked yet, allowing them access to the Tor network.
So in terms of censorship resistance, bridges with occasionally changing IP are better for the Tor network than those with static IP.