Re: [GNUnet-developers] [OpenWrt-Devel] [LEDE-DEV] [OpenWrt] Project proposal: The GNUnet of autonomous Things

[Daniel Golle]

Hi!

On Sat, Dec 03, 2016 at 09:49:52AM +0000, address@hidden wrote:
> > as Kathy mentioned as first answer to Daniel, a proper setup running
> GNUnet on our boards could give us a boost in the IoT world.
> 
> It won't boost anything because GNUnet is broken and unusable. Please stop

How did you get to this conclusion?
To me, after having looked at and used all available P2P overlay
networks I could find, GNUnet is still the most advanced (in terms of
security and privacy) and most reliable thing I could find. (and I've
tried cjdns, tinc, n2n and of course tor)
I'm currently using gnunet-vpn to traverse NAT/firewalls for remote
access and management without the need for any central service and thus
single point of failure. Using GNS identities to manage access control
is quite straight forward and I'm not aware of any alternative approach
which could potentially meet the same goals.


> following buzzwords and fads, fix GNUnet and make it usable before bringing
> more bugs to the already bug-ridden IoT.

Having actual users of long-existing low-level features (overlay vpn
with cryptographic identifiers) brings in a lot of testing without
adding new requirements. Making low-bandwidth applications which do
not require any database backend work first also seems like a very
good approach to me -- once that's well-tested in diverse environments
(think: different home networks with their NAT routers, mobile ISPs
with awkward traffic policies, ...) we can proceed more easily and
have more complex and demanding applications on top (such as
file-sharing or real-time video conferencing).


Cheers


Daniel