Slashdot Mirror
Ants in your P2Pants
Tim Finin writes: "Anthill is a framework being developed at University of Bologna to
support the design, implementation and evaluation of P2P applications,
viewing them as instances of Complex Adaptive Systems, typically found
in biological and social sciences. In Anthill, desired properties such
as resilience, adaptation and self-organization correspond to the
"emergent behavior" of the underlying CA system. An Anthill system
consists of a dynamic network of peer nodes; societies of adaptive
agents (ants) travel through this network, interacting with nodes and
cooperating with other agents in order to solve complex problems. The source code for Anthill v1.0 is available for downloading. MORE on this is at ebiquity.org."
Nice to see a more practical application of ant systems. The past year I have been working with ant systems in a more academic setting: optimizing dynamic problems (problems that change over time) using ant systems. Travelling salesman to be more precise, but my salesmen (sales-ants?) encountered traffic jams. :)
The research can be found here
Next idea: ant based routing. Get rid off BGP, use ants
Sooner or later more of our Technological systems will emulate nature. For those who haven't read Kevin Kelly's masterpeice - Out of Control should do so. Our tecnhological future will become more and more alive as time progresses.
www.enthea.org
"Anthill is a frameworkDING! being developed at University of Bologna to support the designDING! , implementationDING! and evaluationDING! of P2PDING! applications, viewing them as instances of Complex Adaptive SystemsDING! , typically found in biologicalDING! and social sciencesDING! . In Anthill, desired properties such as resilienceDING! , adaptationDING! and self-organizationDING! correspond to the "emergent behavior"DING! of the underlying CA systemDING! . An Anthill system consists of a dynamic networkDING! of peer nodes; societies of adaptive agentsDING! (ants) travel through this network, interacting DING! with nodes and cooperating with other agents in order to solve complexDING! problems. The source code for Anthill v1.0 is available for downloading. MORE on this is at ebiquity.org."
Ladies and gentlemen, we haaaave a winner..
Bowie J. Poag
Draging a hazy memory out of my brain... I'm sure that I saw an article about 'ant-modeling' in New Scientist quite a lot of years ago (at least 2/3 years ago) where they were using ant's and their scent trails (modeled of course...) to find the optimum routes across a network.
I think it worked as followed:
Put a bunch of ants on the start node, all they know how to do is travel from node to node and they know when they reach the end node.
Each ant will go down a path to a connected node, except that they will not backtrack to their previous node. Each of these ants is releasing an electronic scent and they are more likly to go down the path with the strongest scent.
Repeat with several thousand ants and you should have your optimum path across the network.
I seem to recall that the article said that BT (British Telecom) was looking into this as a way to optimise the routes that a phone call takes from switching station to switching station... but it was a very long time since I read it.
[No ants were harming in the writing of this post]
Those words do have meaning. And while they can be used for empty hype, there existance alone is not an indicator of vapiedness.
And besides how can you consider "implementation" and "complex" buzzwords but not 'dynamic', 'peer nodes' or 'downloading'?
autopr0n is like, down and stuff.
Unlike Anthill we're addressing a wider variety of needs (in fact, any kind of app you want to scale up), so you can build all kinds of things on top of Jtrix, from traditional P2P (Freenet, Gnutella, etc) to other apps needing scaling and security (Web-based mail, Passport-a-like, and so on).
We've been at it 18 months now and have things like an HTTP server and servlet engine which work in this environment. It's great to see these kind of frameworks growing in popularity.
Nik
Anthill is full of bugs!
-
- - You can't take something off the Internet! That's like trying to take pee out of a swimming pool.
They have the right idea when it comes to studying large distributed networks. Network emergent properties is absolutely the most interesting thing in that field, and I strongly believe that it is a much better model for trying to handle highly dynamic networks than trying to route within a rigid structure as lot of object location networks have been designed to do.
That said, peer to peer network simulators and frameworks are a dime a dozen - anybody who was at the last O'Reilly "P2P" conference would have noticed that everybody, their mother, and their great-aunt Petrunella seems to have one. Setting up a bunch of java interfaces is not that hard, and calling a Node a "Nest" and giving it the following methods:
public interface Nest {
void request(Request request,
ReplyListener listener);
void addService(AntFactory factory);
void addNeighbor(NestId nid);
void removeNeighbor(NestId nid);
NestId[] getNeighbors();
}
is not exactly a radical new design.
What is really needed are more examples of analytical work regarding the emergence of and utility of routing properties in the known algorithms, as well as other radically new routing schemes besides Freenet. I'm glad that these people have their framework in place, but it really isn't news until they have something to show for it.