[tor-relays] What's the timeline for a Tor relay to start routing a large amount of traffic?

[Nathaniel Suchy (Lunorian)]

Thank you Matt for your rapid response. My Scaleway hosted exits have a
200Mb/s port and I want to provide as much bandwidth and relay as much
traffic as possible out of the port speed.

I've been told it's better to run eight small relays than one large
relay (diversity of the network) and I'm currently running three small
exits. I'd be happy if even each relay could reach just 50Mb/s worth of
transfer or close to. I want each relay to reach its maximum potential.

Next up, my provider Scaleway offers x86, ARM, and ARM64 CPUs, which
type works the best for high traffic Tor Relays?

Cheers,
Nathaniel Suchy

> On Apr 27, 2018, at 3:00 PM, Matt Traudt <pastly at torproject.org> wrote:
> 
>> On 4/27/18 14:34, Nathaniel Suchy (Lunorian) wrote:
>> I started to run a few Tor Exit Relays (My exit policy only allows ports
>> 80 and 443 only to minimize the amount of abuse) a few days ago.
>> Currently Tor Metrics shows the relays are only advertising 700Kb/s
>> despite the fact the minimum port speed on the VPSes I host on is
>> 100Mb/s. I've been told by friends of mine it's because it's a new relay
>> - at what point does the relay start to route a large amount of traffic
>> equivalent to the port speed.
>> 
>> I'm hoping to contribute to the Tor Project in a variety of ways over
>> the next few months and the summer and starting relays is a first step.
>> 
>> Cheers,
>> Nathaniel
>> 
> 
> Thank you for running an exit relay!
> 
> A relays advertised bandwidth is usually the maximum amount of traffic
> it has seen over a recent 10 second period. Tor does not try to
> determine and then advertise your NIC (or port) speed.
> 
> You might still be ramping up because you're a new relay, yes. But
> because of current network conditions, exits tend to reach their full
> potential much quicker than non-exits. The "lifecycle of a new relay"
> blog post doesn't really apply to exits.
> 
> Depending on a wide variety of factors -- many of which are vague and
> hard to reason about -- you may never see your relay push 100 Mbps.
> These factors include things like what your machine has for CPU and
> memory (easy to understand); but also things like how close you are to
> the bandwidth authorities that are measuring you, how close you are to
> the web servers they are using to help measure you, where you are in
> relation to other relays in the network, how many people are using Tor,
> how many instances of the bandwidth scanning tool are currently not
> crashed, and whether the bandwidth authorities have switched to a new
> tool (maybe someday soon!).
> 
> Please do not be discouraged if you never reach 100 Mbps. I run many
> relays (some exits) and I don't think they tend to reach the maximum
> possible speeds the machine is capable of. They also vary wildly in how
> much they are used. Just look at this[0] relay's 1 year graph.
> 
> If after a few months you still are using a small part of your available
> NIC speed, you should consider running a second relay on the same
> machine. You can run two relays per IPv4 address.
> 
> Thanks again for running a relay.
> 
> Matt
> 
> [0]:
> https://metrics.torproject.org/rs.html#details/91516595837183D9ECD1318D00723A8676F4731C
> _______________________________________________
> tor-relays mailing list
> tor-relays at lists.torproject.org
> https://lists.torproject.org/cgi-bin/mailman/listinfo/tor-relays

-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 833 bytes
Desc: OpenPGP digital signature
URL: <http://lists.torproject.org/pipermail/tor-relays/attachments/20180427/03254fab/attachment-0001.sig>