[tor-relays] new VPS bridge bandwidth under-reported

[isis]

starlight.2015q1 at binnacle.cx transcribed 0.5K bytes:
> Whoa wow. . .
> 

Bridge relays conduct both bandwidth and reachability self-tests.  (See §2.1.3
of torspec.git/path-spec.txt.)  The Bridge then includes its self-measured
bandwidth as the bandwidth-observed value on the "bandwidth" line of its
(bridge-)server-descriptor, which it submits to the BridgeAuthority.  (See
§2.1.1 of torspec.git/dir-spec.txt.)

The BridgeAuthority then (also) conducts a reachability test of the Bridge's
ORPort, afterwards the BridgeAuthority takes the lesser value of the
bandwidth-observed (as measured by the Bridge itself) and the bandwidth rate
limit (as specified by the Bridge's operator in the Bridge's torrc) (§3.4.2
torspec.git/dir-spec.txt) to be the value for the "w" line in the
bridge-networkstatus document (which is kind of like a mix between a
microdescriptor-vonsensus and a consensus, except that it's really its own
crazy document format, and it's only for Bridge relays).  (To see examples of
what these descriptors look like, see [0])

> It just popped to 700KB, presumably because I used it for to browse and then
> download the TBB bundle as a test.

I believe that client activity shouldn't effect your reported bandwidth (but
I'm not entirely sure, actually).

> So I guess that means the bandwidth measurement for a bridge is strictly
> passive?

Yes, currently there is only the Bridge's self-reported bandwidth.  There is
no infrastructure for authoritatively measuring any of the Bridges'
bandwidths, however, I need this infrastructure, and its creation is currently
being prioritised, so hopefully it will exist quite soon. :)

> Presumably that also means that it is not used as a criteria for
> dissemination?

No, it is not used.

Bridges may lie all they like about their bandwidths in their descriptors; in
fact, a Bridge may lie about many things in its descriptors.  However,
BridgeDB is programmed, loosely speaking, to take certain descriptor
information more or less seriously, depending on which information it is and
which of the Bridge's descriptors it originated from.

[0]: https://para.noid.cat/bridgedb/descriptors.html
-- 
 ♥Ⓐ isis agora lovecruft
_________________________________________________________
OpenPGP: 4096R/0A6A58A14B5946ABDE18E207A3ADB67A2CDB8B35
Current Keys: https://blog.patternsinthevoid.net/isis.txt
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 1154 bytes
Desc: Digital signature
URL: <http://lists.torproject.org/pipermail/tor-relays/attachments/20150108/2fa34900/attachment.sig>