[tor-dev] New documentation for Tor Metrics website

Karsten Loesing karsten at torproject.org
Sat Nov 22 09:32:33 UTC 2014

On 21/11/14 23:05, Lei Yang wrote:
> Dear Dr. Loesing,
> I have a concern about the consensus_weight. The definition of it seems
> that the consensus_weight is a completely measured value by the
> authorities. However, from the Bandwidth Scanner Specification, if I
> understand correctly, it looks like a combination of observed bandwidth and
> the measured bandwidth. This is the description about how to calculate the
> consensus weight in the Bandwidth Scanner Specification, "we compute an
> average of the filt_bw fields over all nodes we have measured. These
> averages are used to produce ratios for each node by dividing the measured
> value for that node by the network average. These ratios are then
> multiplied by the most recent observed descriptor bandwidth we have
> available for each node, to produce a new value for the network status
> consensus process."
> If I am wrong, could you please let me know? Thank you very much!

Hi Lei,

thanks for your feedback.  You're correct that the consensus weight is
based on observed and measured bandwidth.  The definition I wrote was an
attempt to put the consensus weight, which is to some extent verified by
third parties, in contrast to advertised/observed bandwidth, which is
entirely self-reported by relays.  Let me try to make the definition
slightly more correct without making it unnecessarily complex:

consensus weight: value assigned to a [relay] that is based on bandwidth
observed by the relay and bandwidth measured by the [directory
authorities], included in the hourly published consensus document, and
used by [clients] to select relays for their [circuits].

Does this definition sound okay?  Anything else I can tweak?

Obviously, this definition is not as precise as the definitions given in
the Bandwidth Scanner Specification.  The idea is that whoever reads
specification documents doesn't need to read the shorter definitions on
the Tor Metrics website.  But at the same time people shouldn't get
confused by oversimplified definitions.

Again, thanks for your input!

All the best,

More information about the tor-dev mailing list