Arturo:
We're working on specification tasks.
We completed backend spec; including specifying collector policy for both inputs and net-tests.
Stephen reviewed the spec: https://github.com/TheTorProject/ooni-spec/blob/master/oonib.md
Two points of discussion:
S 2.5 how we publish reports - Ooni suggested scp, but MLab may be different. We should leave that out of the backend spec, because it depends on the backend infrastructure.
How will the test_helper learn effective probe IP address? Stephen and Arturo realize they misunderstood the issue, and this should already be handled by ooni (see: https://github.com/TheTorProject/ooni-probe/issues/90). Daira discussed alternatives with Arturo.
Why is the collector exposed by https and tor hidden service is answered in the spec.
Added a feature which exposes inputs / decks from the backend.
Added a "bouncer" in the spec which aids in discovery for test helpers and collectors. This is similar to mlab-ns. Some aspects might be implemented by mlab-ns. There's a need for ooni to provide requirements for mlab-ns. Stephen will review that spec and send questions about "bouncer" to understand how similar it is to mlab-ns.
We improved some charts: backend dataflow diagram. Stephen will do something similar for MLab. See: https://github.com/TheTorProject/ooni-spec/blob/master/diagrams/ooni-backend...
Nathan asks: Should the ooni-spec and ooni-probe wiki be merged? Arturo suggests we (Arturo & Aaron) should consolidate ooni-spec, should migrate to the wiki.
Nathan gives a Threat Model presentation. That's me! -so I didn't take notes on that section. It was an overview of this wiki: https://github.com/TheTorProject/ooni-probe/wiki/Threat-Model
As is apparently now our custom: "Toodles!"
-- Nathan Wilcox Least Authoritarian
GPG id: CBA48BDD GPG fingerprint: 9309 EB72 87D4 192F 0D7E 0F16 1191 AF31 CBA4 8BDD