Discuss mailing list

[Coyo <coyo AT darkdna.net>]

On 09/22/2013 10:33 PM, John David Galt wrote:
> On 2013-09-22 20:17, Coyo wrote:
>> Verisign is a very influential company.
>>
>> This company has abused that trust innumerable times.
> They're not the only ones. [...] SSL should be considered completely 
> compromised.
>
> What we need is either a new, web-of-trust protocol, or at the least a
> new, trustworthy SSL authority that can take the place of all the ones
> we have now.

There have been attempts at such a thing.

The problem with things like PGP-based server-to-server authentication
is that there is no ready-and-waiting module, library or process that
system and network administrators can use for things like DNS,
Federation, Email, XMPP, and day-to-day operations.

There are many browsers, applications, libraries, frameworks, servers,
etc. that rely on the CA system, and either do not natively support
DANE, or would be very time consuming to port. All of the
standardization authorities have been compromised even more than SSL/TLS
and DNS itself, which is shameful.

However, such a system is more than possible on paper and in theory.

Using an alternate blockchain with the namecoin client (headless)
provides a distributed TLD namespace, then use slightly enhanced BIND9,
PowerDNS, and other ordinary existing nameserver software packages, with
beefed up mandatory DANE and DNSsec, IPsec and whatnot.

What DANE does is replace the CA system with the nameserver hierarchy.
The DNS servers become the CAs.

You could distribute the root domain with (alt)namecoin, such as with a
long time-between-blocks-mined, a higher difficulty, and very high value
of domains as the TLD.

Then you could mount the ICANN namespace to a special .icann pseudo-TLD,
which tunnels the request elsewhere, possibly over a VPN.

www.google.com.icann. Just wait till you see the look on their faces at
W3C, IEEE, ISOC, ANSI and ISO. XD

Treating ICANN's namespace like a dustbin relic would really put them in
their place.

Instead of treating opennic, i2p, onion or namecoin namespaces as the
odd names out, why not turn the tables and simply begin treating ICANN
like the nostalgic curios they are.

Heck, if I had my way, I'd throw out the entire IPv4 address space and
reallocate all of it. Since I'm unlikely to get that, I'll just wait
till whatever comes after IPv6 begins rolling out.

As for me, personally, I'm not waiting for anyone. I'm currently working
out the kinks for Internet Protocol version 14 (IPv14).

It involves end-to-end circuit-switching, end-to-end encryption, secure
information retrieval, zero knowledge proof, secure caching, distributed
computation, inter-cloud exchanges.

It also involves convergence of on-chip networks, network fabric
extension into the home and across submarine cables, seamless
unobstructed dedicated bandwidth from device to device,
fiber-to-the-component broadband as a taken-for-granted free service,
photonic crossbar integrated circuits, and billions of cores in every chip.

Silicon will be a thing of the past. Diamond quantum cryptography,
diamond photonic computation and storage, diamond holographic storage,
flattening and distribution of the network end-to-end, from individual
thread, process and internal component to every other thread, process
and component, inside any other computer, inside many computers, or in
many virtual computers.

This is also made possible by wireless power transmission,
communications aeroship-based freespace optical communications, photonic
phased arrays, and carrier-assisted P2P and F2F content distribution.

The general direction is diamonds, diamond electronics, diamond
photonics, diamond quantum compute and storage, convergence of all
currencies, networks, utilities and infrastructure, inter-cloud
standardization and commoditization of compute, storage and bandwidth
resources, and the unification of infrastructure.

Suffice to say, IPv14 is going to be mindblowingly awesome.