Re: Tor 0.2.0.30 is released

[Scott Bennett <bennett@xxxxxxxxxx>]

     On Thu, 21 Aug 2008 15:44:21 -0400 Roger Dingledine <arma@xxxxxxx>
wrote on or-announce:
     [many wonderful things skipped  --SB]
>  o Performance improvements (bandwidth use):
>    - Don't try to launch new descriptor downloads quite so often when we
>      already have enough directory information to build circuits.
>    - Version 1 directories are no longer generated in full. Instead,
>      authorities generate and serve "stub" v1 directories that list
>      no servers. This will stop Tor versions 0.1.0.x and earlier from
>      working, but (for security reasons) nobody should be running those
>      versions anyway.

     I gather the second item explains the disappearance of ~47% of the
servers between 21:30 GMT yesterday and 05:00 GMT today, as well as most
of the traffic going through my server (0.2.1.4-alpha).  The numbers have
come back by maybe a hundred servers or so since I got home at about 05:00
GMT, but it looks like a huge loss--hopefully very temporary!--of connectivity
and maybe bandwidth.
     If most other servers are experiencing big drops in traffic levels,
then the fact that tor updates its "observed" bandwidth downward as well as
upward when it publishes descriptor updates will also produce a huge illusory
loss of *apparent* (not real) bandwidth in addition to the real losses caused
by the cessation of v1 directory distribution by the authorities.  I assume
that this divergence of estimated bandwidth from actual bandwidth will
eventually be automatically corrected by tor, but I fail to see why it was
necessary in the first place.
     Publication of "observed" bandwidth updates ought to serve as successive
approximations of the actual bandwidth capacity of each server.  In most
cases, because a server's physical capability and that of its network
connection(s) tend not to change very often, allowing tor to reduce its
estimate below its previous estimate (since initialization, of course) seems
counterproductive.  If tor only changed the "observed" value when it had
grown since the last report with no intervening restart, then actual capacity
information about servers would be preserved intact through major events that
take a big bite out of tor traffic for a while, and that preservation of the
capacity information ought to speed recovery of the tor network as a whole by
aiding proper statistical allocation of circuit routes over the servers that
remain on-line during and after the events.
     Anyway, I'm only a little over half way through reading Roger's
announcement, and it looks just great!  I'll have to save the rest until
after I get some sleep, though. :-)  Thanks much to Roger, Nick, and all
the others who worked on the features and fixes in the 0.2.0.30 release!


                                  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         *
**********************************************************************