Relay bandwidth needed to pay back Hidden Service usage

[Scott Bennett]

     On Sun, 10 Jan 2010 01:00:57 -0500 grarpamp <grarpamp at gmail.com>
wrote:
>Say a hidden service makes available a 100,000 byte file. Now
>another user downloads that file. There was obviously some 'cost'
>in bytes to the six hop relay system for doing that.

     Six?
>
>Say the user who downloaded that file feels obligated to repay his
>usage back to the system by running his own non-exit relay. How
>many bytes should he offer back to the network?
>
>It may be more proper to think of it as bandwith. Server serves a
>stream 24x7x365 at 100,000bps. User consumes it 24x7x365. How much

     How so?  How could someone download a single 100,000 B file non-stop
for a year?  Hidden services are already a pain to use, thanks to their
a) extra long circuit construction times, b) slow transport over circuits
with so many hops, and c) higher rate of circuit breakage.  A person who
really needs to keep her accesses concealed may be more willing to put up
with these problems than someone who doesn't really need the anonymity.

>inbound and outbound pipe should user provision to replace his usage?
>
     Why should such a user be expected to "pay back" anything?  Keep in
mind that the people who most desperately need to use tor's hidden
services facilities are likely to be among those least able to provide
relay services themselves.
>
>When compared to the internet, the network is a closed system with
>finite resources, so the cost is certainly not equal to his usage
>[ie: 1x]. His usage causes a depression of overall available resources
>during the transfer.

     Yes, both the Internet and the tor net can be modeled in this way.
Your point is?
>
>I thought it might simply be 6x his usage due to 6 hops. But maybe

     I thought there were only 5 as diagrammed below, assuming the
standard of routelen=3.  If both sides use the same routelen, then the
number of relays has to be an odd number because the final relay from
each side is the same.

>it involves adding 6 increasing fractions? And what about the middle
>relays that both receive and send the supply stream?

                       Rendezvous
     C----CSR1----CSR2---- X ----HSSR2----HSSR1----HS
                         Relay

where
      C    = client
      CSR  = client-side relay
      HS   = hidden server
      HSSR = hidden server-side relay
>
>Tor overhead and packet retransmissions can be ignored. Tor's
>configuration limits and the availablity of any given pipe size can
>also be ignored as this is only an excercise.
>
>The return tcp ack stream would not be negligible but could be left
>out by reference... only if it was simply additive on top of the
>supply stream using the same calculations.

     I guess I must have missed something because I don't see what you're
driving at.  Sorry if I'm just being dense tonight. :-]


                                  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 at torproject.org with
unsubscribe or-talk    in the body. http://archives.seul.org/or/talk/