Re: Gnunet and TUDelft ipv8

[Lluís Batlle i Rossell]

Hello,

IPv8 provides "overlays" over internet, which are isolated networks with
its own addressing based public key identities and not rented IP
addresses. They also accomodate for some anonymity with onion routing.

Multiple overlays can coexist in an application to achieve a service,
where multiple identities serve different purposes.

The trustchain is an interesting blockchain to decentralize trust, over
which multiple applications can be built. Financial, reputation, etc.
These are used for the onion routing too. Articles about trustchain are
easy to find.

The TU Delft people put on practice since many years these concepts in
their Tribler VideoStreaming program.

I also don't know what is not reasonable about IPv4 and UDP; ipv8 is the
transport that hides any particular IP details, be it IPv4, UDP or
anything else.

In any case I'm interested in opinions of those who did not just stop
reading at the front page of the python docs. :)

On Mon, Apr 13, 2020 at 12:21:10PM +0200, Schanzenbach, Martin wrote:
> Hi,
> 
> 
> > On 13. Apr 2020, at 12:02, Lluís Batlle i Rossell <address@hidden> wrote:
> > 
> > Hello,
> > 
> > I think that GNUNet and TUDelft's ipv8 (and whole ecosystem of trustchain,
> > onion routing, etc.) have a big overlap.
> > 
> > Is there any shared work? or declaration opposite work? Or evaluation of
> > both by one or the other side?
> 
> I do not think there is any detailed review or comparison. I have never heard 
> of it until now.
> Reading the "Feature" section here [1] I do not think this is well thought 
> through concept.
> It cites two publications, one of them only dealing with NAT and at the same 
> time claiming it is "academically pure" (whatever that means, I assume they 
> mean "sound"? The papers are 7-9 years old).
> It relies on UDP via IPv4 as transport which does not seem like a reasonable 
> choice given the options on the table today.
> Flexibility with respect to transports is something we have identified as a 
> core issue. Especially since eliminating metadata-exposing 
> transports/addressing on the physical layer is what we need in the long run.
> 
> To be honest I also stopped reading at "ledger-based storage of reputation 
> data".
> Without trying to bash, looks to me like yet another blockchain-based future 
> internet tech.
> The website at least fails to provide the key value offerings. Are there 
> other resources?
> If they have any insights into better NAT handling, that would a great 
> resource, but the paper is already 9 years old so I would assume people read 
> it already.
> 
> BR
> Martin
> 
> [1] https://py-ipv8.readthedocs.io/en/latest/
> 
> > 
> > Thank you,
> > Lluís.
> > 
> > --
> > (Escriu-me xifrat si saps PGP / Write ciphered if you know PGP)
> > PGP key 7CBD1DA5 - https://emailselfdefense.fsf.org/
> > 
> 



-- 
(Escriu-me xifrat si saps PGP / Write ciphered if you know PGP)
PGP key 7CBD1DA5 - https://emailselfdefense.fsf.org/