Opened 8 years ago

Closed 8 years ago

#5908 closed project (fixed)

Get all the knowledge related to OONI into the wiki

Reported by: hellais Owned by: hellais
Priority: Medium Milestone:
Component: Archived/Ooni Version:
Severity: Keywords: SponsorH201206
Cc: ioerror, isis, phobos Actual Points:
Parent ID: Points:
Reviewer: Sponsor:


There is now a wiki on trac that will be used as a knowledge base for OONI:

We should start collecting all OONI related information there.
This includes:

#5866 - Research on test to write (

#5867 - Devise methodology taxonomy (

#5865 - Evaluate other tools and solutions (TODO)

#5864 - General Research Methodology (

#5906 - Evaluate the OONI Architecture design (

#5907 - Design the OONI Backend (

Child Tickets

Change History (6)

comment:1 Changed 8 years ago by hellais

Milestone: Sponsor F: July 1, 2012Sponsor H: June 2012

comment:2 Changed 8 years ago by isis

Status: newneeds_review

There are only two things left to be done:

  1. Add the revisions to the FOCI '12 paper on OONI to the Architecture wiki page.
  2. Revise the Backend wiki page to not say "HTTP", but rather "HTTPS/HTTP over Tor HS", and finish the discussion on possible usage of rsync or other protocols for report submission.
    • The currently proposed method for keeping state is to first upload a YAMLOONI file containing pointers to all other YAMLOONIs containing results.
    • The client should also be able to request information from the backend through this process, such as "Hey, Backend! Give me a TCP traceroute to!" and the Backend can send that back to the client.

Unless more things come up, I believe we can close this when those two wiki pages are revised.

comment:3 Changed 8 years ago by karsten

Keywords: SponsorH201206 added
Milestone: Sponsor H: June 2012

Switching from using milestones to keywords for sponsor deliverables. See #6365 for details.

comment:4 Changed 8 years ago by isis

Right. The only thing left for this deliverable now is to finish the design of the Backend, which we hadn't entirely agreed upon. During our meeting in Firenze, the following concerns were raised:

  • HTTP is not secure, and the wiki page does not make it clear that 'HTTP' here refers to the protocol, which will, of course, offer SSL as a security mechanism.
  • HTTP is 'dirty', ie, is it not efficient, it's compression is not the greatest, it does not support deltas or diffs, and it is generally not what one uses to move files around from one computer to another.
  • Should we instead use rsync (or duplicati or some other variant) and only use HTTPS/O as as a fallback? Rsync seems to solve pretty much all the problems we've considered, including resume support, verifiability, security, efficiency, compression, and portability.

ioerror, hellais: Can we find time soon to discuss this soon, either here, email, or IRC, so that we can resolve this and be finished?

comment:5 Changed 8 years ago by hellais

I think we should continue the discussion of this on the tor-dev mailing list. I threw in a few points here:

comment:6 Changed 8 years ago by hellais

Resolution: fixed
Status: needs_reviewclosed

Closing since #5907 is closed.

Note: See TracTickets for help on using tickets.