On Wed, 16 Feb 2011 21:07:50 +0100 Klaus Layer <klaus.layer@xxxxxx> wrote: > Lora Roa <hechicera@xxxxxxxxx> wrote on 16.02.2011: > > 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. > > > > Getting < 1 connection a day now, bandwidth meter reports 7:1 ration of > > incoming to outgoing. Which made me recheck and retest all my port forward > > settings days ago when I saw it start, but I can't figure it out from my > > end. > I assume this has something to do with the donated bandwidth. If the donated > bandwidth is to low, the relay will (maybe only currently) not being used. This does not apply to bridges. > I have 5 relays running. One at me home DSL connection (3Mb/s download, > 512Kb/s upload) with 168 Kb/s donated bandwidth . Average bandwidth usage > 15Kb/s. It is almost unused. Another node runs at a cable modem connection > (32Mb/s down, 2,5 Mb/s up) and 1,4Mb/s donated bandwidth. Average usage 840 > Kb/s. Then I have a root server with two instances. One is under the top 40 > TOR nodes. Heavy usage around the clock always 1500-2000 concurrent > connections. Average 20 Mb/s. > > My impression is that from point of view of bandwidth it is useless to have > these small relays. Although they help to increase the number of servers and > make it more difficult for an attacker to monitor the TOR network. You're right -- low-bandwidth relays are nearly useless. For security reasons, Tor clients must choose the relays in their circuits at random. Tor attempts to load-balance circuits across the published relays by weighting the probability with which clients choose each relay, but this load-balancing approach cannot make good use of low-bandwidth relays. If you want to contribute your bandwidth on a slow (50 kB/s or less) connection to Tor, consider running a bridge to help users bypass Internet censorship instead of running a published relay. Robert Ransom
Attachment:
signature.asc
Description: PGP signature