[Author Prev][Author Next][Thread Prev][Thread Next][Author Index][Thread Index]

Re: [tor-dev] Latency measurement from exits



Hello,

I haven't measured that, but maybe this is helpful: I have measured some
time ago was the latency in a hidden service (rendezvous) circuit:

Client - guard - middle - rendezvous -- middle - middle - guard - HS

[6 hops]

I did the test by installing an OpenVPN TCP server on the hidden service
that listened on 127.0.0.1 and allocated VPN IP's in subnet 10.8.0.0/24.

The client connected via Tor's socks5 to the .onion address where the
VPN server was listening and I was able to do ICMP and UDP between vpn
client (IP: 10.8.0.5) and vpn server (IP: 10.8.0.1).

On average the latency was about 600-700 ms. The result is heavily
dependent on all the relays in the path, so I am not sure how can we
test this to an exact value. You can try now and get lucky with only
fast relays in the path and have a 400 ms latency and you can try later
and be less lucky and have a 800 ms latency.


On 3/10/2016 4:42 PM, Zack Weinberg wrote:
> Before I go reinventing wheels, has anyone got code to measure network
> latency (either RTT or one-way packet travel time will do) from a
> chosen Tor exit to an arbitrary destination?  Latency from client to
> exit would also be useful.
> 
> Thanks,
> zw

Attachment: signature.asc
Description: OpenPGP digital signature

_______________________________________________
tor-dev mailing list
tor-dev@xxxxxxxxxxxxxxxxxxxx
https://lists.torproject.org/cgi-bin/mailman/listinfo/tor-dev