tor replacement - was Re: Box for simple Tor node.

grarpamp grarpamp at
Mon May 11 01:45:43 PDT 2020

On 5/9/20, jim bell <jdb10987 at> wrote:
> At this point, I see the main impediment is finding somebody with the
> motivations and qualifications to write the software.  An additional
> complication is that whoever volunteers, he might not be trusted by

While coders can provide design meta, coders can also
be hired more or less to just meet some premade spec.
It will be opensource where trust as to code exploit
is somewhat reasonably determinable. Trust as to protocol spec
design itself holding up to adversaries in operation is a different
area of evaluation.

> What is to be done?

People might start surveying relevant existing networks and papers,
past and present, note and annotate all their design and features in
some big comparison tables, cut out their bad parts, invent new parts,
assemble all the then viable parts into some design specification.
Parade it around to see how badly it gets attacked and broken.
Then scrap or amend it, and code and deploy it.

Or skip all those traditional formalities and just start
hacking stuff together.

> The one situation that I consider intolerable is that TOR remains as a
> monopoly in the "anonymization marketplace".

Yes, there should be some solid competition in the
deployed overlay network space.

A good generic overlay transport network might be one that will be able
to carry, and thus cater to, many people's desires to otherwise go off and
create single purpose networks that would generally have the same anonymous
overlay feel but for different applications... such as one net for messaging,
one net for storage, one for cryptocurrency, voice, grid compute, etc, etc.
Doing ten different application nets seems a bit redundant effort and tech,
instead of ten different plugins into one net.

Of course if you restrict yourself to only same basic functions as Tor
(onionland + exits) under an alternative new Tor design + say chaff,
things become easier, at expense of being able to plug more
applications generically over it.

Defining what you want to be, and how, is work.
Coding is more trivial.

Tor does have a monopoly over automagic exit capability.
But networks like i2p and phantom do compete with it in
offering psuedo TCP network stack compatible hidden services.

There are probably at ten or so reasonably well papered overlay
networks that never got implemented and could be drawn from.

The internet just transports messages around a packet switch,
only the applications know whether they're storage, coin, voice,
messages, etc.

More information about the cypherpunks mailing list