God, the chaff fill concept is so dang simple - every time period T,
send as many packets to PO(outgoing_bandwidth) to link endoint, if
insufficient packets to send, send pseudo random chaff fill packets
to get up to PO.

That's the outgoing side.

The incoming side is of course similar: over T, shape down to PI,
dropping packets beyond that limit (+/- S 'slack' packets I guess).

SIE (incoming packet shortage event): If PI is deficient, lower
incoming rate PI (and notify endpoint of course).

Rate reliability of endpoints based on SIE counts per time period
RT (reliability analysis time period) - which might also be
multi-layered (term?) i.e. do SIE analysis over multiple relevant
time periods, e.g. 1s, 10s, 1m, 10m, 1hr.

Simple (i.e. asynchronous) event/message protocol to request
bandwidth, offer bandwidth in response to request, and offer
bandwidth pre-emptively, including time parameters of duration (offer
to guarantee rate PI/PO for T, or "best effort").

Best effort might be suitable for some requestors, insufficient for
others - make the requests you require, offer what you can, meet your
contracts when made, rate those endpoints who fail in their
contracts, possibly also include "allegations" (i.e. ratings as made
by peers about other peers they've had experience - obviously
allegations would carry less weight than personal/own-node
experiences, yet, as in the torrent world, may well be useful to
improve overall network characteristics.


The more fundamental issue with Tor is its currently TCP only design.

I2P is therefore probably a better starting point for chaff fill -
everything else can be layered on top of ethernet, IP and/ or UDP.

Good luck,



On Thu, Aug 02, 2018 at 04:43:20PM -0700, Steven Schear wrote:
> I've been waiting for Loopix, or a similar, anonymity overlay to go into
> widespread testing but it's seems to be stillborn.
>
> On Thu, Aug 2, 2018, 4:38 PM Zenaan Harkness <zen@freedbms.net> wrote:
>
> > On Thu, Aug 02, 2018 at 04:27:53PM -0700, Mirimir wrote:
> > > On 08/02/2018 04:14 PM, Steven Schear wrote:
> > > > "Augur’s creators claim they don’t have control over what its users
> > choose
> > > > to do with the protocol—or the ability to shut it down. This creates a
> > > > problem that is “endemic” to blockchain technology, says Wright, who
> > > > recently co-wrote a book on the subject: “If you do not have a very
> > > > concrete intermediary—i.e., a company or group of people that are
> > running
> > > > the marketplace—how do you apply laws and prevent that activity from
> > > > occurring?”
> > > >
> > > > This is, as they say in marketing, not a problem but a feature.
> > >
> > > Yep. I mean, that's the fucking point!
> > >
> > > But I gotta say, they need to work on the anonymity aspect. The argument
> > > that participation anonymity doesn't matter, as long as adversaries
> > > can't attribute stuff, is weak. Look at Freenet, for example.
> >
> > Free speech (free from censorship and arbitrary punishment against
> > 'speech the state does not like') depends on anonymity, which depends
> > on the transport - e.g. localised face to face conversation on the
> > beach (and no sand bugs in your ears), or some form of mesh
> > networking.
> >
> > At the very least, for digital anonymity, trusted entry nodes to a
> > network, some number (math math) of trusted peer nodes (if all your
> > entry node's peers are untrusted and say kill your link
> > simultaneously, or in symmetric semi-rapid succession, they will
> > likely be able to correlate your communication) and chaff fill links.
> >
> > 'Till then... pigs ain't flyin'.
> >
> >
> > > > On Thu, Aug 2, 2018, 3:55 PM jim bell <jdb10987@yahoo.com> wrote:
> > > >
> > > >>
> > > >>
> > https://www.technologyreview.com/s/611757/this-new-ethereum-based-assassination-market-platform-could-cause-napster-size-legal/
> > >
> > > <SNIP>
> > >
> >