Why so little bandwidth used?

[cmeclax-sazri]

On Wednesday 16 February 2011 12:21:43 Steve Snyder wrote:
> Tor has 30 HTTPS connections open as I type this.  I don't know what port
> "management port " refers to.
>
> According to my Tor log files I (ORPort =9001, DirPort=9030) am reachable
> and pass the bandwidth test.

I think the "management port" is 9051. telnet localhost 9051 and type the 
following:
authenticate "cunsivyvla"
setevents bw
Assuming that "cunsivyvla" matches HashedControlPassword in torrc, you'll get 
a line every second that says how many bytes went in and out. I max out for 
several minutes a few times a day, from what I've seen. (I haven't tried to 
graph it, but I will.) My max is 65536, with some larger bursts:
650 BW 49446 50739
650 BW 61042 64400
650 BW 53038 51298
650 BW 53672 57269
650 BW 29416 32638
650 BW 45324 45006
650 BW 55074 56720
650 BW 45762 45598
650 BW 58888 61624
650 BW 50106 45666
650 BW 53920 56350
650 BW 39128 39960
650 BW 38486 40842

There are various kinds of traffic that could be passing through your relay. 
There could be a hundred pairs of people talking on TorChat, and you'll see a 
586 (512 plus overhead) go by a few times a minute. There could be someone 
browsing a website, and you'll get maxed out some seconds, and other times 
random numbers up to your max. Or someone could be downloading an ISO or 
sending another TorChatter a big file, and your link will saturate for many 
minutes if it's the slow one.

On Wednesday 16 February 2011 14:22:28 Lora Roa wrote:
> Same thing, but with a bridge.  Can confirm ports forwarded, hybernate not
> messing with it.  I though it was just me or some pattern in how my bridge
> was handed out.

According to statistics I took before Egypt shut down the Internet, there are 
500 bridges and 600 entry guards, and there are 12,000 bridge clients and 
200,000 entry guard clients. So a bridge gets only 7% as much client traffic 
as an entry guard.

cmeclax