×

Welcome to the Slashdot Beta site -- learn more here. Use the link in the footer or click here to return to the Classic version of Slashdot.

Thank you!

Before you choose to head back to the Classic look of the site, we'd appreciate it if you share your thoughts on the Beta; your feedback is what drives our ongoing development.

Beta is different and we value you taking the time to try it out. Please take a look at the changes we've made in Beta and  learn more about it. Thanks for reading, and for making the site better!

Self-Assembling Networks

michael posted about 11 years ago | from the sysadmins-out-of-a-job dept.

The Internet 112

prostoalex writes "Researchers from Humboldt University found a way to build self-assembling networks. By emulating the behavior of ants and insects the team, which is led by Frank Schweitzer, demonstrated a simulation where agent-based architecture was able to quickly assemble itself into a network and quickly react to a broken link or damages. Schweitzer's research papers are available off his personal Web site. The scientific paper referred in the original article, Self-Assembling of Networks in an Agent-Based Model is available off Cornell server."

cancel ×
This is a preview of your comment

No Comment Title Entered

Anonymous Coward 1 minute ago

No Comment Entered

112 comments

How interesting (-1, Troll)

unterderbrucke (628741) | about 11 years ago | (#5606096)

Re:How interesting (0)

Anonymous Coward | about 11 years ago | (#5606124)

Is this a joke? If so it's not very funny. That link has nothing to do with this subject - don't click on it...

We already have one of those. (5, Funny)

Maradine (194191) | about 11 years ago | (#5606099)

My network team looks *just* like a swarm of ants when the network goes down.

Re:We already have one of those. (1)

Maradine (194191) | about 11 years ago | (#5606128)

And come to think of it, management strongly resembles a swarm of bees . . .

Re:We already have one of those. (1)

fubar1971 (641721) | about 11 years ago | (#5606670)

Yeah I know what you mean, the scary thing is, I've noticed a change in there voices. My lead engineer is starting to sound like Woody Allan, my structured cable guy is starting to sound like Sly Stallone, and for some reason I noticed my voice is starting to sound like Gene Hackman. I just wish my wife would sound more like Sharon Stone :)

ANTZ [movieweb.com]

Re:We already have one of those. (0)

moominpapa (193163) | about 11 years ago | (#5607107)

D*mn, where's the mod points when you need them? This should be +5 Funny already.

Enormous Benifit (1, Interesting)

Anonymous Coward | about 11 years ago | (#5606115)

Imagine if you will being able to create and configure a LAN using technology like this? How long till we see it in Linux.

Setting -> autolan configure -> select yes -> give network a name -> done!

Re:Enormous Benifit (1)

Thyrhaug (536821) | about 11 years ago | (#5606223)

That sounds more like a windows-networking-thing. In linux i guess it'll be a configure file with at least 100 lines, exluding the comments.

Re:Enormous Benifit (1)

42forty-two42 (532340) | about 11 years ago | (#5606315)

Yes - and the only one you need is at the top.

Re:Enormous Benifit (0)

Anonymous Coward | about 11 years ago | (#5606646)

Which config files are you looking at? All the ones I've ever seen bury the important stuff somwhere in the middle of the file, usually segmented between two totally unrelated sections or even better, in a totally unrelated section. If there are any comments, they are usually cryptic and/or useless.

Example:
# If you need EQGS profiling, leave this option
# set to "No"
use_eqps = yes;

# Every FTR node must have a name. Each name must
# be unique, unless it is inside, or outside, of a
# HTO cluster. If you are using HTO clusters,
# then do not use this option, instead see the
# above section "My Favourite Pets"
primary_node_name = Node;
secondary_node_name = Node;

#
# My Favourite Cats
#
fffg = 3;

# Name your HTO nodes in decending order, starting
# with the top item last. Ensure that this list
# has been alphabetised in numerical order, but do
# not include non-active HTQ domains.
hto = fluffy;

So on and so forth.

Re:Enormous Benifit (1)

multi io (640409) | about 11 years ago | (#5610037)

Which config files are you looking at? All the ones I've ever seen bury the important stuff somwhere in the middle of the file, usually segmented between two totally unrelated sections or even better, in a totally unrelated section. If there are any comments, they are usually cryptic and/or useless.
How does that differ from the average Windoze "preferences" dialog burying important options in 85 checkboxes and textfields, each one with a "cryptic" one-line explanation that can be understood only by reading the fine manual?

Right: The latter can't be easily created by scripts, nor easily managed using a version control system, nor easily augmented with self-written comments.

Re:Enormous Benifit (-1)

Anonymous Coward | about 11 years ago | (#5606346)

What a stupid troll.

Re:Enormous Benifit (1)

mmol_6453 (231450) | about 11 years ago | (#5606352)

I think it has the best immediate potential for things like ad-hoc 802.11 networks that consist of live cars on the freeway, and buildings nearby.

Traffic jams would be ironic...your data could be moving faster than you. :)

Re:Enormous Benifit (1)

iMMersE (226214) | about 11 years ago | (#5606478)

I dunno about what you've made your network out of, but my data is always moving faster than me.

True. That's why there already using it for that. (1)

fireboy1919 (257783) | about 11 years ago | (#5607097)

While it's nice that this guy wrote a paper about this, I already know of one company that is putting it into practice (obviously because they're not done with it yet, I can't give you a link - but I can tell you that I heard about the project three years ago). Considering that this is already in the hands of a corporation, it's been in the academic world for quite a lot longer.

In fact, I had a prof who wrote a paper about that. In fact, he got accepted as a professor at my old school because of his network knowledge in that area (that was five years ago). Can anybody point out why this guy's writing isn't redundant?

Re:Enormous Benifit (1)

paskie (539112) | about 11 years ago | (#5609269)

That resembles IPv6 router advertisment and node autoconfiguration to me. Actually you don't even need to select yes, just feed the router with a simple configuration (and it _is_ much less than 100 lines ;-).

I think this article wasn't exactly focused on IP-layer networks, though.

Self-assembling intelligence next? (4, Interesting)

wiggys (621350) | about 11 years ago | (#5606122)

Taking this idea one step further, what if each computer node on the network was given a basic set of rules so that it emulated a bunch of brain cells. Would the network self-organise to create some sort of intelligence?

Re:Self-assembling intelligence next? (1)

Psiren (6145) | about 11 years ago | (#5606181)

Would the network self-organise to create some sort of intelligence?

Ignoring the ongoing debate about what is or isn't intelligence, they have been doing this sort of thing to a limited extent in software for a number of years. Neural networks are actually used for specific applications. Don't expect your network to suddenly gain an IQ of 200 though... ;)

Re:Self-assembling intelligence next? (5, Informative)

Neuronerd (594981) | about 11 years ago | (#5606188)

It turns out that already today all successful applications of socalled "artificial intelligence" are self assembling.

In the first approaches to artificial intelligence [mq.edu.au] people used programming languages to obtain systems that generate intelligent or at least apparently intelligent behavior.

All newer [utexas.edu] approaches to artificial intelligence start with a large number of very simple units that, learning from data from the real world, develop specific patterns of connections. Many models even develop their own structure in such a way.

From my perspective is intelligence as well as artificial intelligence only possible in a system that can self-structure.

Re:Self-assembling intelligence next? (1)

oliverthered (187439) | about 11 years ago | (#5606373)

If a node came across somthing it didn't understand or was unsure about it could ask the other nodes for their opinion.

Re:Self-assembling intelligence next? (0)

jeanicinq (535767) | about 11 years ago | (#5608599)

..., we already wondered about how the brain of mother nature works. The ants are just the links between neurons. Next, they will figure out that ants are totally blind and work entirely by sense of smell besides their antennae.

They didn't react to this one very quickly /.'d (1, Funny)

DrSkwid (118965) | about 11 years ago | (#5606129)

okument nicht gefunden

Die von Ihnen gewählte URL ist auf unserem
Server nicht bzw. nicht mehr vorhanden.

Gehen Sie bitte zur AIS-Homepage oder
benutzen Sie das Navigationsmenu links.

Shit (1)

Lolaine (262966) | about 11 years ago | (#5606134)

My girlfriend was going to study Networks Cabling and Construction (wires , switches ...). It will be funny to say her she will be messed with ants and spiders ... I think you will hear her scream from USA :D

This isn't new (0)

Anonymous Coward | about 11 years ago | (#5606135)

MS platforms have had self-assembling perr-to-peer networks for years. The latest one even worked with SQL Server!

Maximum Comments Exceeded! (0)

Anonymous Coward | about 11 years ago | (#5606140)

You've reached your maximum number of comments you can post: 365 comments over 4 hours.

Chances are, you're behind a firewall or proxy, or clicked the Back button to accidentally reuse a form. Please try again. If the problem persists, and all other options have been tried, contact the site administrator.

I wonder (1)

compubomb (612155) | about 11 years ago | (#5606142)

i have nothing informative to say.. but hell, i think this is really great software-engineering development.

centralized? self asembly? (2, Insightful)

IAR80 (598046) | about 11 years ago | (#5606145)

For now I stick to OSPF. And it is not centralized also. And so are BGP, RIP an ISIS.

bit = binary digit?! (-1)

Anonymous Coward | about 11 years ago | (#5606152)

01100110 01110101 01100011 01101011 01111001 01101111 01110101
01100001 01101100 01101100

Maybe we should use the borg icon for this one... (3, Funny)

MeanE (469971) | about 11 years ago | (#5606156)

"We are your network...ect..ect...we will adapt"

this isn't news (2, Informative)

potaz (211754) | about 11 years ago | (#5606157)

Last Update: 18 March 1999
The article was posted to his web site in 1999 and this is front-page stuff? And the article itself was published in 1997. Stop the presses!

Re:this isn't news (1)

flokemon (578389) | about 11 years ago | (#5606217)

Where did you find this? The main document linked to dates from March 26th, 2003, and the reports mostly have dates around November 2002. Maybe not exactly news, but not as outdated as you seem to suggest.

Re:this isn't news (1)

tincho_uy (566438) | about 11 years ago | (#5607057)

The method they talk about is some form of ant-system optimization, and it's most definitelty not new... I think there was a recent article that talked about that (swarm intelligence), but I don't feel like searching the link...

Too bad for me.. (1)

sokkelih (632304) | about 11 years ago | (#5606159)

Now it seems that my time in college is wasted. No need for network admins.

Re:Too bad for me.. (1)

hatstandman (466901) | about 11 years ago | (#5606177)

That's what I thought, then I looked at the patch cabinets...this software'll have to get much better before it can sort out a cable that plugs into 3 different ports :S

I can see it now! (5, Funny)

Neck_of_the_Woods (305788) | about 11 years ago | (#5606175)


HUB, "MALFORMED PACKET!!!! AHHH!!!! - HELP HELP HELP! I am lost!"

Router "Calm down, this is nothing compared to the broadcast storm of 93. Everything will be alright."

HUB, "Thank you,"

Router "These simpletons, when will they ever learn just to ignore that packet."

ala - bugs life.

Ant-like-technology (2, Funny)

LamerBunny (613373) | about 11 years ago | (#5606176)

"By emulating the behavior of ants and insects..."

(I wonder who played the Queen...)

I didn't know ants were this advanced! This must be the final proof that indeed insects are super-intelligent aliens come to earth to eat our... ehm... sugar-water... If only we can harness this power elsewhere! Maybe we should try milipede power-plants next... All that static electricity from all those legs must be harnessed!

Re:Ant-like-technology (1)

downtrader (649720) | about 11 years ago | (#5606581)

Damn, that flies in the face of my years of belief that ants are really the tanks of a much smaller super-intelligent being or wait was it that ants were the only appendage we can see of another dimensional being. Nothing like a slashdot networking discussion to ruin my years of ant theory.

another great example (0)

jsinger (415774) | about 11 years ago | (#5606195)

...of how we can study millions of years of evolution and use it to help our daily lives.

go ants!

Re:another great example (-1)

Anonymous Coward | about 11 years ago | (#5606578)

Yes, I enjoy reading fiction too

It's about time. (1)

oneiros27 (46144) | about 11 years ago | (#5606205)

Wasn't Skynet [imdb.com] supposed to be self-aware like 5 years ago?

I mean, it's 2003, and we don't even have systems that we can't leave alone over the weekend. Where's the AI that's supposed to do all of the thinking for us, so we can actually get some free time? [Okay, there's that little problem with it trying to kill off all humans, but well, I'm sure they'll fix that in release 2]

Re:It's about time. (1)

Sarin (112173) | about 11 years ago | (#5606350)

First of all, I bet there's a logical explanation for that, watch the new terminator movie (it's supposed to be released somewhere in june) and there's no such thing as skynet ofcourse, you are safe.

Where's the AI that's supposed to do all of the thinking for us, so we can actually get some free time?

the problem is that everytime ai comes up with a new findings it's quickly adopted in all kinds of automation processes, people don't consider it ai anymore when they know how it works: "hey, that's not ai, it's just a mathematical formula that does things this way or that way". Many people don't realise how broadly ai is being used, for instance consider a heart-monitor used in hospitals, it's being driven by a state of the art (knowledge based) ai system

There's a lot to be learned from selforganising systems like the behavioural processes of ants, bees and other insects (in the core it's all just mathematics evolved over millions of years, so there must be something good about it, right), it's really nice to see how people copy and adapt certain ideas.

this comment was auto-generated by skynet 3.0

Re:It's about time. (1)

WasteOfAmmo (526018) | about 11 years ago | (#5608708)

the problem is that everytime ai comes up with a new findings it's quickly adopted in all kinds of automation processes, people don't consider it ai anymore when they know how it works: "hey, that's not ai, it's just a mathematical formula that does things this way or that way". Many people don't realise how broadly ai is being used...

This brings up an interesting point. When many people talk about some self-aware 'evil' computer systems they generally think of some large project having gone wrong (i.e. WOPR in Wargames, skynet in Terminator, or HAL in 2001, or the computer in Two Faces of Tomorrow [jamesphogan.com]). Discussions usually center around the concept of a single project or thread of projects that ends unexpectedly in a self-aware system which goes haywire. The solutions to these scenarios typically include those implemented in various movies (somehow take out or prevent that single project from completing and the world is saved) or at least along the theme of dealing with a single instance of the problem.

In reality as the parent poster mentions most innovations are quickly taken up by the masses and incorporated into existing systems on a continual basis. This means that over time most systems and processes continue to evolve at roughly the same rate. Hypothetically speaking, this could lead to a point of critical mass in more then one production system where the last incremental upgrade/enhancement was enough to initiate the final evolution of a self-aware system (ignoring for a minute the whole argument as to what self-aware means).

When, if ever, a system does become self-aware it will likely happen:

  • First in a lab in a project that pushes existing technology "over the threshold". But many other systems will be extremely close to the threshold also.
  • Accidently in one or more systems simultaneously (or relatively) as by-products of projects pushing the technology.
  • or, Worse case scenario, accidently in one or more production systems due to incremental innovations pushing the systems over the threshold.
IMHO the first case is more likely but along with it comes the realization that most production systems will be extremely close to the threshold at the time also. This scenario would more then likely cascade into either the second or third scenario.

Of course all this conjecture does not touch on the personality traites of said self-aware systems. This I leave to the philosophers and Sci-Fi writers in the crowd.

--
WhatSpellChecker?.Merlin.

*waves* (1)

DaLiNKz (557579) | about 11 years ago | (#5606215)

...and then all the system administrators disappear.. But then again, most sysadmins arnt even around their networks. instead they are answering calls from room 0139 because her computer can't boot up.

With your connections combined... (0)

Joe the Lesser (533425) | about 11 years ago | (#5606220)

I am Captain Network!

*theme song*

Captain Network,
He's our hero,
gonna cut packet loss
down to zero!

Anyone else thinking what I'm thinking? (0)

masq (316580) | about 11 years ago | (#5606230)

[with awe]
The. Ultimate. Napster.

Re:Anyone else thinking what I'm thinking? (-1)

Anonymous Coward | about 11 years ago | (#5606840)

Nope. I'm thinking about Aleesha Dixon from Mis-teeq rapping while I sink myself in... CONKERS DEEP.

Re:Anyone else thinking what I'm thinking? (0)

Anonymous Coward | about 11 years ago | (#5607643)

Wow, good work mods. Using a self-healing network that routes around censorship would NEVER work for Napster. Right. Thanks for your insight - Napster and Kazaa are obviously in no danger of being taken out by the RIAA. Like Dubya says as he monitors your reading / buying habits, "YOUR Freedom is overrated, mine is absolute."

Just don't let Al-Jazeera have this, we could have a PR "incident" on our hands that we can't contain if they show the aftereffects of our depleted uranium bullets. Shoot an Iraqi, he gets cancer. The bullet casing drops to the sand, the kids who live there get cancer. Drink the water, get cancer. Great. I bet George feels like a real man now. And unthinking, you mods toe the Rumsfeld / CIA party line.

America only has the freedom it fights for, and I don't see you putting up much of a fight right now. I hope you're not trusting A POLITICIAN to fight for your freedom. You should have learned your lesson by now....

Prey (1)

henrod (661974) | about 11 years ago | (#5606231)

Anyone read Michael Crichton's new book, Prey? This kind of thing is a little scary after reading that one!

Survey: Mod this down if you molest animals. (-1, Offtopic)

Anonymous Coward | about 11 years ago | (#5606243)

I'm doing a survey to see how many people on Slashdot molest and have sex with animals. To register a vote in the positive, please mod this post down. If you do not have sex with animals, mod it up. Don't mod it at all if you are undecided or 'curious'.

a solution for Mesh-Networks??? (0)

Anonymous Coward | about 11 years ago | (#5606258)

So I wonder how long it'll take before we see this idea applied to mesh-networks? I mean that really does seem like a logical progression for this It was my understanding that 'keeping the network up' was a big problem for the current implimentation of wifi meshnets...

Actual info? (1)

evilviper (135110) | about 11 years ago | (#5606272)

This article is beyond ridiculous. It is more like a pointless press release from Dr. Seuss than actual info.

A node does this, then it does that, that somehow attracts other nodes doing something else, and POOF, the world is a great place to live in once again...

Give me a break. I'd rather read about magic, self-healing, server pixie-dust.

On a similar note, look for Dr. Seuss' latest book in stores soon: "One Node, Two Node, Red Node, Blue Node"

Actually uninformed? (0)

Anonymous Coward | about 11 years ago | (#5606954)

You apparently didn't read the PDF. Or (more likely) you skimmed over it without understanding any of it. Either way give the PDF a read.

P.S. I know I'm responding to a troll but its been a long time since I've done so and at least I'm doing it AC

Re:Actual info? (1)

Mignon (34109) | about 11 years ago | (#5607943)

This article ... is more like a pointless press release from Dr. Seuss than actual info.

Here's Dr. Seuss' explanation of how self-assembling networks handle errors:

If a packet hits a pocket on a socket on a port,
And the bus is interrupted as a very last resort,
And the address of the memory makes your floppy disk abort,
Then the socket packet pocket has an error to report!

From here [sydney.net].

emulating insects (0)

Anonymous Coward | about 11 years ago | (#5606284)

where agent-based architecture was able to quickly assemble itself into a network and quickly react to a broken link or damages


Where can i buy one of these cute little critters.
I would put it with those sea monkeys i have...

Potential (4, Interesting)

perspex (635004) | about 11 years ago | (#5606288)

This could be really cool for ad-hoc wireless networks.

Re:Potential (1)

c64cryptoboy (310001) | about 11 years ago | (#5608505)

How is this any better than existing ad-hoc network protocols?

Here's a paper [uci.edu] that compares (in the context of wireless networks) DSDV-SQ, TORA, DSR, and AODV-LL protocols for how well they make use of shortest paths, number of packets successfully delievered, ability to deal with dropped nodes/connections, and routing overhead (as either packets or bytes).

these guys don't watch enough anime (2, Funny)

andih8u (639841) | about 11 years ago | (#5606302)

Anyone who does knows this is just a step before the evil computer AI infecting all of the other computers in the world and setting about to destroy mankind. I will rise up to defeat this terrible menace right after I find a girl with blue hair and eyes the size of dinner plates.

Re:these guys don't watch enough anime (1)

Thing 1 (178996) | about 11 years ago | (#5609920)

I will rise up to defeat this terrible menace right after I find a girl with blue hair and eyes the size of dinner plates.

Ask and ye shall receive [amazon.com] .

Web Service Intermediaries... (1)

DigitalCH (582593) | about 11 years ago | (#5606343)

A lot of the web service intermediaries have these kind of capabilities...
You plug in their agents on the network and they slowly become aware of each other through message exchange. When one section of the network goes down the agents talk to each other to figure out which agent can be used to relay a message around the broken link.

It's really wierd to be up in layer 7 and see the same modeling of behavior of lower layers in the stack...

That's not how ants would do it (1)

HarmlessScenery (225014) | about 11 years ago | (#5606359)

If it was real ants there'd be none of this "You have the blue nodes and we'll have the red nodes" niceness.

The blue ants would be killing the red ants and vice versa - and the scent given off by the dying ants would attract more ants to the area until there was one hell of a war going on for territory (nodes). With the winners getting better connectivity for their network. And the ants would quickly specialise into scouts, soldiers and queens (to reinforce the army).

Come to think of it, that'd be much more interesting than plain old networking anyway :)

All this cuteness and collaboration all the time - do people not realise that the Mother in Mother Nature isn't mother as in 'caring nurturing type person' - it's mother as in M.O.A.B. :)

Coded (1)

PHanT0 (148738) | about 11 years ago | (#5606384)

I've help test, write and measure the success of code that does exactly this. Our implementation uses the ants to collect routing latencies and update routing tables. It's actually surprisingly efficient and deals very well with downed nodes depending on your timeouts for downed nodes, etc.

Poor choice of language leads to misconceptions (0)

Anonymous Coward | about 11 years ago | (#5606429)

They call this "Self Assembled" as if it has some sort of adaptive intelligence. It does not, it has some algorithms that will have to be understood and worked around when they fail. This should be called "Algorithmically Assembled" or "Programmatically Assembled" so more people won't anthropomorphize their electronic equipment, imbuing it with special properties it just doesn't have.

Big Deal (1)

Salamander (33735) | about 11 years ago | (#5606454)

Been there. [newscientist.com] Done that. [ulb.ac.be] These types of algorithms are not exactly new, and what this paper describes is no more "self-assembling" than any other distributed routing/discovery protocol - examples of which have existed for over twenty years. Of course, lots of things are new to the Slashdot editors that are old to the rest of us.

Re:Big Deal (-1)

Anonymous Coward | about 11 years ago | (#5606993)

Mod up parent, please. Just for the sig line.

--
Slashdot: news for herds. Stuff that splatters

Alright, where's the cleanup robot swarms? (1)

ahfoo (223186) | about 11 years ago | (#5606521)

The somewhat self-assembling nature of P2P networks got me thinking about little swarms of tiny clean up robots. Instead of a hunanoid robot, it seems what would be more useful and simpler for things like household or even commercial maintenanc is a network of small robots relying on each other for various specialized functions sort of like cells in a larger organism.
It seems like you almost have to forego the android approach and go this way to get automated maintenance workers financially feasible because there will be certain parts that will tend to wear out much faster than others. It's the nature of the clean up game that many of the parts are consumeables.
While my musings on P2P were rather far from the goal, this sounds quite a bit closer. I know some fugly buildings in a town not far from here that could really use a good scrub down.

Re:Alright, where's the cleanup robot swarms? (0)

Anonymous Coward | about 11 years ago | (#5607047)

Kid:Hi, Mom! I'm home!

Mom:Well, watch your clumsy feet. The carpet ants are out and if you step on any more of the vacuum cleaners, I'm going to take it out of your allowance.

Re:Alright, where's the cleanup robot swarms? (0)

Anonymous Coward | about 11 years ago | (#5609585)

Scrubbing bubbles! I can already hear the Wagner.

DNS (1)

CAIMLAS (41445) | about 11 years ago | (#5606539)

People have been tlaking for years about how kludged together the current internet infrastructure is; my question is, might something like this make for a feasable replacement, or at least a suppliment to what is already out there? I can see this being very useful indeed. You'd be able to de-centralize the root servers, and have them be distributed from

Rendezvous (1)

ajs (35943) | about 11 years ago | (#5606699)

I don't know a lot about the state-of-the-art in the area of network discovery/repair other than what I know as a socket-programmer and sysadmin, but I'm wondering if someone who does know can point out the differences between, say, this research and Apple's Rendezvous [apple.com] (not to be confused with Tibco's product by the same name [tibco.com])?

It seems to me that the basic goals are similar, but with Apple focusing more on the engineering side of solving a user-problem rather than passing the point of diminishing returns on "correct" solutions. Please, feel free to enlighten me though. This stuff is actually really promising, and I hope to live in a world 5 years from now where my laptop just "fits in" to the network that it's placed on in more ways than mere DHCP can accomodate.

Thanks!

Re:Rendezvous (1)

ajs (35943) | about 11 years ago | (#5608238)

And the answer is... RTFA ;-)

My bad, wrong kind of "network". While the idea may be applicable to computer-interconnectivity, that's not what this is about, and I would have known that, had I read the article.

Thanks anyway, all!

Humboldt University, eh? (1)

schnarff (557058) | about 11 years ago | (#5606742)

Why am I not surprised that the people in Humboldt found a way for themselves to do less work?

I wonder if we'll see a press release from them later saying they've designed something to emulate a particularly famous local plant. ;-)

Re:Humboldt University, eh? (0)

Anonymous Coward | about 11 years ago | (#5607177)

I just dont get this joke, maybe because this is about Humboldt University in Berlin. which Humboldt are you talking about?

Self-Assembly gone wrong (1, Funny)

lavalyn (649886) | about 11 years ago | (#5606789)

IT Guy: We're being nailed off our ABC uplink with a denial of service attack!
Manager: Well, we still have our DEF uplink in reserve. Drop everything from ABC!
IT Guy: Okay, much better now.... oh wait, the network reassembled to attack our DEF link!
Manager: I think I'll be cavorting in Arizona for a while...

Are there any agents? (1)

nanojath (265940) | about 11 years ago | (#5606938)

where agent-based architecture was able to quickly assemble itself into a network...


Good god, didn't you people learn anything from The Matrix?! Agent-based architecture is the most dangerous type of computer system you can design!

SPAM requires this repair (1)

4of12 (97621) | about 11 years ago | (#5607054)


I'm of the opinion that spammers represent an infection of the net and that we are watching how the network is adapting to fight it off.

no value (0)

Anonymous Coward | about 11 years ago | (#5607125)


It still requires humans to setup the architecture
before anyone or anything could use it.

I.e., someone has to set up the OSI level 1 and level 2 before any "thing" could use it.

self-assembling? (0)

Anonymous Coward | about 11 years ago | (#5607183)

Researchers from Humboldt University found a way to build self-assembling networks

Why did they have to build the network if it was self-assembling?

A good research work (3, Insightful)

varjag (415848) | about 11 years ago | (#5607683)

The paper is indeed very interesting and innovative, but keep in mind that it is very far from being suitable to embed into your next 802.11 adapter.

While this approach is indeed appealing, it has still some drawbacks, e.g:
- generally, you can't tell what your topoligy your network will end up having, so forget about architecting one
- it does not guarantee that all your nodes will end up being networked within a fixed number of attempts (see the fig. 3 in the paper)
- it tends to require significant redundancy of interchangeable nodes to function well

Such approach can work well, say, for military field communications, but would be clearly suboptimal for building a corporate network.

And of course, as most of agent research, this is still too far from established technology ready for production.

Re:A good research work (1)

mivok (621790) | about 11 years ago | (#5609865)

disclaimer: I have not (yet) read the paper, just the first linked article.

But the problems you describe wouldnt be problems for say an ad-hoc p2p wireless network, with each node forwarding for others.

- generally, you can't tell what your topoligy your network will end up having, so forget about architecting one

It doesnt really matter, as long as it works and can get packets from a to b, to me it seems the whole point is that you don't need to architect the network.

- it does not guarantee that all your nodes will end up being networked within a fixed number of attempts (see the fig. 3 in the paper)

In a peer-to-peer system, this wouldnt be a big problem, as you state, for a corporate network, this wouldnt be acceptable, but the solution is just to keep trying until you do get networked, or the user gets fed up of waiting.

I guess my point is, given time (your point about it not being established is obviously valid, considering it is just a research paper atm) this could be the solution to the peer to peer wifi network that people have been looking for, automatically finding the best route for traffic, without requiring any centralised control. I would assume they still have to work around the problem of malicious agents leaving strong 'chemical' scents, but it still sounds liek a really neat idea.

An observation (0)

Anonymous Coward | about 11 years ago | (#5608149)

The method outlined is an interesting one, but
it has a couple of weaknesses:

* It does not guarantee that every adjacent node
has an actual flow/connection in place (note that
the article says almost every link, not every
one).

* It accumulates information on a random basis
rather than an on-demand basis. While the agents
will in fact eventually find out what the network
looks like, there's no guarantee that they will
find the most needed path first, or even ever.

* There's a degree of implied overhead in the
actions of these agents, necessitated by the
repeated travel over various links, analogous
to a system where one polls to check a link's
speed/state.

* There is no mention in the article of how
routing to remote nodes is achieved; I can
speculate that a piece of information is carried
by an agent acting randomly until it reaches
the right place, but this seems exceedingly
inefficient.

It is quite possible to construct a network
protocol in which the protocol organises itself
without agents, based very simply on the presence
and number of adjacent nodes (active network
links). As a matter of fact, I have done so.

Using on-demand creation of routes permits one
to ignore failures until such time as they
affect desired routes, and only using on-demand
creation of routes means that a lot of preliminary
setup communication can be reduced.

It is also possible to, with very minor human
input, add both a naming system and an encryption
system into the same setup, giving one eventually
a truly decentralised network in which namespaces
can be assembled and dismantled as needed without
reference to a central authority, and in which
authentication can be set up and confirmed by a
trusted authority, or set of authorities.

It dawned on me while I was doing this, after
I'd successfully constructed proofs of efficiency,
that the people most interested would probably
be the mafia, so I stopped work on it. Pity;
it is an interesting challenge, but with the
predominance of IP, it looks like human-organised
networks (yes, IPv6 is also human organised, and
still pretty darn centralised in many ways) are
with us to stay.

Where are the 'pheremones'? (1)

The Raven (30575) | about 11 years ago | (#5608738)

Perhaps I'm not understanding something, but how do the researchers intend on having the nodes in a real life network emit 'pheremones'? The only application I can see for that is self organizing wireless hubs. But for hard wired networks it does not seem to make sense, since there is no way to estimate the ease of connecting two points based on distance alone.

Overall, this article only seems to apply to wireless networks. An interesting, but limited, usage.
Load More Comments
Slashdot Account

Need an Account?

Forgot your password?

Don't worry, we never post anything without your permission.

Submission Text Formatting Tips

We support a small subset of HTML, namely these tags:

  • b
  • i
  • p
  • br
  • a
  • ol
  • ul
  • li
  • dl
  • dt
  • dd
  • em
  • strong
  • tt
  • blockquote
  • div
  • quote
  • ecode

"ecode" can be used for code snippets, for example:

<ecode>    while(1) { do_something(); } </ecode>
Sign up for Slashdot Newsletters
Create a Slashdot Account

Loading...