[tor-reports] Karsten's July 2014

[Karsten Loesing]

Attended the Paris dev meeting, talked about Sponsor O and a bit about
metrics.

Added a new parameter to Onionoo's API that accepts a fingerprint and
returns documents from all relays in the family (#12521).  This
parameter can be useful for websites showing aggregate data from all
relays run by the same person/organization.

Added warnings to Onionoo that the cronjob includes in its mails.

Finished re-processing descriptor archives for adding advertised
bandwidth and consensus weight graph data to Onionoo (#11388).  What
remains is that Atlas, Globe, or another Onionoo client plots graphs
using these data.

Started looking into better algorithms to detect Sybil attacks on the
Tor network.  Current thinking is that we should define relay similarity
metrics like common IP address prefix length or time between first seen
in a consensus, go throw the consensus archives, and see which relays
look similar but are not defined to be in the same family.

Related to Sybil attack detection, briefly looked into parsing votes to
notice when certain relays were first bad-exited, invalidated, or even
rejected by the directory authorities.

Talked about Sponsor O deliverables with Sponsor O folks.

Did some refactorings to Onionoo's codebase to facilitate testing.
Looked into Google's dependency injection framework Guice, but didn't
apply any changes for that yet.

Spent one week doing private stuff.