[tor-talk] Massive Bandwidth Onion Services

[Alec Muffett]

On 23 Dec 2016 11:17 am, "Alec Muffett" <alec.muffett at gmail.com> wrote:

  "an n-squared mesh of daemons which have to communicate with and
authenticate to each other using an application-specific protocol, as well
as maintain some kind of consensus of which workers are alive, which are
temporarily or permanently dead. Also: shared consensus protocols suck."

In short: not being able to use the HSDirs as a "source of truth" makes the
whole architecture a lot more complex and synchronous.


I suppose a possible solution is for the nodes to check their current
introduction points into a shared 'git' repository, which they then
cross-sync with all the other nodes, along with a shared understanding of
the content to be served from any given one of them.

It'd work, but it's still a kludge, and will have a lot of frequent
node-to-node fingerprintable "chatter"; the old-style method using HSDirs
is a bit more like a dead-drop.

    -a