Re: Full bandwidth is not used.

[Scott Bennett <bennett@xxxxxxxxxx>]

     On Sat, 13 Mar 2010 09:05:03 +0100 Paul Menzel
<paulepanter@xxxxxxxxxxxxxxxxxxxxx> wrote:
     My apologies for letting this sit unanswered for so long.  I was tied
up in reconfiguring several disk drives where free work space was cramped
and inconveniently located for some of the moves.  Now I'm just starting to
go through nearly two weeks' accumulation of email.  Agghhh...

>Am Freitag, den 12.03.2010, 19:31 -0600 schrieb Scott Bennett:
>>      Well, as I've pointed out in the past, the values in cached-consensu=
>s=20
>> do *not* accurately reflect either the traffic load that your relay has
>> carried or the traffic capacity of your relay.  They are bogus a priori a=
>nd
>> should be ignored in attempting to ascertain your relay's actual loads.
>> The sad thing is that recent versions of tor clients now use the consensu=
>s
>> values for designing routes for circuits they will build, so the bogus
>> values produce load distortions throughout the tor network.  However, tha=
>t
>> fact has no bearing upon the numbers you're looking for.
>>      If you want to know the loads that your relay has carried, you shoul=
>d
>> look at the byte counts for reads and writes in the extrainfo documents o=
>r,
>> alternatively, the state file.  (The difficulty with using the state file
>> is that it gets updated everytime construction of a new circuit succeeds,
>> so the values for the most recent time periods change frequently and at
>> rather unpredictable intervals.  If you always ignore the most recent tim=
>e
>> period for read and for writes, then the state file becomes more usable
>> for this purpose.)  If, OTOH, you want to know the peak "10-second burst"
>> rate, then the value to trust is the one in your relay's descriptors that
>> appear in the cached-descriptors{,.new} files.
>
>Thank you for your response. I kept that in mind and compared it to the
>values in `/var/lib/tor/state` and they are around the same and maybe
>even lower. I also use tools like `nload` to verify the network load.
>
>You can also see bandwidth graphs at [2].
>
>I am a little confused why you are responding nitpicking at the values I
>give although I think it was confirmed in the whole thread that the full
>bandwidth is not used at all.
>
     I'm sorry that my point wasn't made clearly enough.  IIRC, you were
wondering why the consensus values didn't match what you were seeing your
router do.  (You've deleted the pertinent portion of the message, so I'm
just going by memory here.)  The point I was attempting to make is that
there is no good reason to expect to see any close relationship between
the consensus value and what your router does.
     Your router may very well have also had some real problem, but the
consensus is not a useful tool for diagnosing throughput capacity problems.


                                  Scott Bennett, Comm. ASMELG, CFIAG
**********************************************************************
* Internet:       bennett at cs.niu.edu                              *
*--------------------------------------------------------------------*
* "A well regulated and disciplined militia, is at all times a good  *
* objection to the introduction of that bane of all free governments *
* -- a standing army."                                               *
*    -- Gov. John Hancock, New York Journal, 28 January 1790         *
**********************************************************************
***********************************************************************
To unsubscribe, send an e-mail to majordomo@xxxxxxxxxxxxxx with
unsubscribe or-talk    in the body. http://archives.seul.org/or/talk/