[tor-bugs] #22453 [Core Tor/Tor]: Relays should regularly do a larger bandwidth self-test

[Tor Bug Tracker & Wiki]

#22453: Relays should regularly do a larger bandwidth self-test
-------------------------------------------------+-------------------------
 Reporter:  arma                                 |          Owner:  juga
     Type:  enhancement                          |         Status:
                                                 |  needs_review
 Priority:  Medium                               |      Milestone:  Tor:
                                                 |  0.4.0.x-final
Component:  Core Tor/Tor                         |        Version:
 Severity:  Normal                               |     Resolution:
 Keywords:  034-triage-20180328,                 |  Actual Points:
  034-removed-20180328, tor-bwauth               |
Parent ID:  #25925                               |         Points:
 Reviewer:                                       |        Sponsor:
-------------------------------------------------+-------------------------

Comment (by teor):

 Replying to [comment:20 teor]:
 > Replying to [comment:18 arma]:
 > > I still think there's a strong argument for simplifying our design and
 code and ripping out the bandwidth self-tests. So long as there are
 bwauths of any sort running, they will produce load on the relays, which
 will have the same effect as the self-test. And if there aren't bwauths
 running, then either the existing client traffic is enough to make relays
 realize they are fast or it isn't, but the tiny amount of traffic from the
 self-test isn't going to help that much.
 >
 > I still think there's a strong argument for doing a larger bandwidth
 self-test.
 >
 > On the public network, torflow's partitions give some relays a
 ridiculously small amount of bandwidth. Until the relay picks up traffic
 from elsewhere, it is stuck in the small partition. I have literally had
 to use EntryNodes to run fake client traffic through relays to get them
 unstuck. Removing the small test is risky, because it makes the problem
 worse.
 >
 > On test networks, there may not be any bandwidth scanners. So the
 bandwidth self-test should be on in test networks (and it should be
 larger).
 >
 > I don't think sbws suffers from similar issues, so once all the torflow
 scanners are gone, we can disable the bandwidth self-test code. Until
 then, it should be on and larger.

 I am not sure if sbws suffers from similar issues, but it *does* need
 accurate self-reported observed relay bandwidths.

 And bridges need self-measured bandwidths, because they don't have a
 bandwidth scanner:

 > This is especially important for bridges, since they might go long
 periods without much use.

 https://github.com/torproject/tor/blob/12175987fc744f5ca6559a821867631457911451/src/core/mainloop/mainloop.c#L2271

--
Ticket URL: <https://trac.torproject.org/projects/tor/ticket/22453#comment:28>
Tor Bug Tracker & Wiki <https://trac.torproject.org/>
The Tor Project: anonymity online