Skip to content

LIRA: Lightweight Incentivized Routing for Anonymity [pdf]

nrl.navy.mil
10 pointsowlmusic4 comments
On HN

Comments

Therefore, we assume that the bank faithfully executes the
protocol and only makes observations that are part of that
protocol. In particular, he does not act as an onion
router or as a client, only observes messages that are
sent to him, and does not collude with the network
adversary.

"Not collude" part worries me. Can anyone smart enough comment on this?

No, a better way is to do outreach to the people who have the resources to contribute. I can think of an important group, represented by HN: web site consultants and web companies that may have excess hosting resources at the end of the month, every month.

Essentially Tor could be that thing you run on your servers before you launch; and if you launched, then before you get big. Think of it as a kind of load test for your young, under-used infrastructure.

ADD: How to install tor server:

   https://www.torproject.org/docs/tor-relay-debian.html.en
I'm doing it right now...But I'm realizing that I really need better bandwidth constraints. I want to be able to tell Tor to limit itself by connection count, bandwidth, and memory consumption.

ADD: Holy cow, they really want you to lock down your Tor box:

   https://trac.torproject.org/projects/tor/wiki/doc/OperationalSecurity
Anyone know why they expect a box running a tor relay to be locked down so tightly?

Perhaps because a compromised box is useless for running Tor? The goal here is anonimity. A box that is compromised is one more chink in that anonymity.

AboutSource Built by g1lg1l

Hackerly is an independent reader for Hacker News, built on the public HN API. Not affiliated with Y Combinator.