Business wants a new, profitable internet
Yep, the root of all evil is 'government', unrestrained economic trade will save us all...not http://slashdot.org/yro/01/07/26/1553257.shtml James Choate Product Certification - Operating Systems Staff Engineer 512-436-1062 jchoate@tivoli.com
http://slashdot.org/yro/01/07/26/1553257.shtml These are a bunch of people who want to make fundamental architectural changes to the internet, to make it so they can prevent people from getting services unless they are paid money. Oddly enough, this comes at a time when I've been thinking very seriously about some of the implications of "Plan D." Basically, we need to think very hard about the infrastructure, if we intend to build something that is truly censorship-proof (as opposed to merely content-neutral). Mojo nation and freenet are current appoximations, but they work over the Internet, and that may be their downfall. The problem with Plan D, if implemented over the current Internet, is that the low levels of the internet are a tree rather than a proper network. There are choke points and listening points at which all of a particular person's traffic can be guaranteed to be intercepted. Every packet that traverses the internet can be queried to see where it's going, where it came from, how many hops it has left to live, etc. Most are associated with particular applications, and easily identifiable by a port number contained in the packet headers. Encrypted traffic stands out. Mixes are complicated by the absence of true broadcast (radio, ethernet in promiscuous mode, anything...) anywhere in the infrastructure, with the result that while you may not be able to tell which of the Mixers are responsible for a particular packet, but you can damn sure tell who the Mixers are, and if the Mix becomes enough of a problem you can outlaw it and stomp anybody who traffics in its characteristic packets. Every "solution" to these problems requires identifiable nodes to traffic in detectable types of packets over an increasingly monitored and controlled infrastructure. And the business types, as well as the pols, who haven't been able to cope with the internet's chaotic nature, want the infrastructure *more* monitored and *more* controlled. Since these are the groups that have the money and the power, respectively, they *will* get their way. Most cypherpunkish "dream" applications don't stand a chance of actually surviving full-out censorship and the descendants of the DMCA, in this network or the network that these people want to build. So, while these guys want to make the Internet into some kind of centrally-controlled monopoly, I've been wanting to create something that goes completely the other direction -- a "chaos web" with its own routing and switching and content-migration algorithms, designed specifically to facilitate the desire of some or all nodes and operators to remain impenetrable, uncensorable, and content-unlinkable to any unauthorized listener or would-be spoofer, regardless of the resources (including government) the would-be attacker has available. Characteristics of a "chaos web" include mobile content -- an idea already espoused by mojo nation and freenet -- but that means you can't hook up the database servers on the other end of your website and monitor where the people go, so mainstream businesses will probably not use a chaos web. ultimately though, it comes down to some kind of alternate infrastructure. Bear
At 01:54 PM 7/26/01 -0700, Ray Dillinger wrote:
Oddly enough, this comes at a time when I've been thinking very seriously about some of the implications of "Plan D." Basically, we need to think very hard about the infrastructure, if we intend to build something that is truly censorship-proof (as opposed to merely content-neutral). Mojo nation and freenet are current appoximations, but they work over the Internet, and that may be their downfall.
All your comments are correct, but forget replacing the IPv6 protos for about 50 years.
On Thu, 26 Jul 2001, Ray Dillinger wrote:
Characteristics of a "chaos web" include mobile content -- an idea already espoused by mojo nation and freenet -- but that means you can't hook up the database servers on the other end of your website and monitor where the people go, so mainstream businesses will probably not use a chaos web. ultimately though, it comes down to some kind of alternate infrastructure.
That won't happen, of course. Far too few care enough to pay for it. What you'd probably want is a VPN sorta configuration layered on top of the existing Net, with some killer application the normal network infrastructure cannot support. I.e. you need to carve out your own space much in the way Big Business intends to. Some of the stuff the IETF routing people are coming up with might be useful. I'm thinking Cisco's MPLS (for VPN sorta configurations and QoS guarantees), what the manet WG is doing (for decentralized routing infra), and broadcast/multicast routing (for replication and anonymity). Sampo Syreeni, aka decoy, mailto:decoy@iki.fi, gsm: +358-50-5756111 student/math+cs/helsinki university, http://www.iki.fi/~decoy/front
-- On 26 Jul 2001, at 13:54, Ray Dillinger wrote:
The problem with Plan D, if implemented over the current Internet, is that the low levels of the internet are a tree rather than a proper network. There are choke points and listening points at which all of a particular person's traffic can be guaranteed to be intercepted.
That is true of the typical home connection. It certainly is not true of many of the networks I have access to. To crack down on the internet, the government has to construct a mapping from packets to true names. In general that is hard, even though it is easily doable for lots of packets and lots of true names. --digsig James A. Donald 6YeGpsZR+nOTh/cGwvITnSR3TdzclVpR0+pr3YYQdkG RLqtJ72UcIIsUjZOwHNOsz/YzK+IpZPk8CuF2Jti 4ePgVwggx855XmBvPnEuSTEGX+J9Etf0m+xAq4qLa
On Thu, 26 Jul 2001, Ray Dillinger wrote:
The problem with Plan D, if implemented over the current Internet, is that the low levels of the internet are a tree rather than a proper network.
Which is where a distributed grass roots wire-less network based on 'small world networks' comes into play. In general this new network layer will have two components. The first are 'static' within a region. This includes both fixed point and mobile users in that locale. The second, and to date pretty much ignored aspect, is long-distance long-haul. The answer here is automobiles using the same sort of technology as above but tied to highways and byways. Small Worlds: The dynamics of networks between order and randomness. D.J. Watts ISBN 0-691-00541-9
There are choke points and listening points at which all of a particular person's traffic can be guaranteed to be intercepted. Every packet that traverses the internet can be queried to see where it's going, where it came from, how many hops it has left to live, etc.
Only if we keep letting 'them' take the lead in the standards definition and implimentation. This is the true power of Open Source software, it allows the individual to select what they will use, as opposed to some dicta from 'them'.
Every "solution" to these problems requires identifiable nodes to traffic in detectable types of packets over an increasingly monitored and controlled infrastructure.
Beep, wrong answer. -- ____________________________________________________________________ Nature and Nature's laws lay hid in night: God said, "Let Tesla be", and all was light. B.A. Behrend The Armadillo Group ,::////;::-. James Choate Austin, Tx /:'///// ``::>/|/ ravage@ssz.com www.ssz.com .', |||| `/( e\ 512-451-7087 -====~~mm-'`-```-mm --'- --------------------------------------------------------------------
participants (6)
-
David Honig
-
jamesd@echeque.com
-
Jim Choate
-
Jim Choate
-
Ray Dillinger
-
Sampo Syreeni