Silk Road forums
Discussion => Silk Road discussion => Topic started by: TheGoodSon on November 10, 2012, 02:41 pm
-
The success of SilkRoad has likely afforded the founders many luxuries in life. I'm a little curious to know why there are no safeguards. Computer hardware is dirt cheap, you could put several "servers" online for a couple thousand dollars These could be configured to failover in the event of a server crash or other outage. Alternatively, these could be configured to load balance with database transactions being synced via ssh on the clearnet. There are probably a thousand ways to prevent an outage like this, why has nothing been planned for?
-
Load balancing over such a low bandwidth medium as Tor seems like it would have more negative drawbacks than positive.
Alternatively, these could be configured to load balance with database transactions being synced via ssh on the clearnet.
Good suggestion, although it's not ideal; if somebody was to analyse the traffic on a set of exit nodes a server that was in constant communication with the Tor network would look mighty suspicious; at best it would simply be identified as an augmenting service to a random clearnet/hidden service (although only one hidden service that I can think of would require such a thing...), at worst it would be pegged as something worth investigating.
-
The point I was trying to make in the first postwas:
You have three servers, each at a different physical location.
You have a heartbeat and database sync over the clearnet through a highly encrypted channel (maybe a VPN?)
In the event of the heartbeat stopping, another server takes over.
The main SR URL could be a server that randomly picks one of the three servers to send traffic to in theory.. but, the better way would be to have all servers communicate through heavy crypto, with heart beats and then in the event that server A crashes, the main director (silk road URL server) picks the next best server.. so on and so forth. This would virtually guarantee 100% uptime unless the SR URL redirector crashed, and even then, you just have to pass out the alternate URLs on the forum.