Slashdot is powered by your submissions, so send in your scoop

 


Forgot your password?
Close
typodupeerror
×
The Internet

New P2P tool Using... IRC? [UPDATED] 97

Posted by Hemos from the interesting-idea dept.
SupremeOverlord writes "A new P2P file-sharing tool called "BitHive" is entering public beta soon. This one uses IRC servers to connect nodes to avoid the scalability problems Gnutella suffers while not having a centralized server like Napster. Check out the press release at BitHive.org, and sign up for the upcoming public beta here. At the very least, it's going to be an improvement over automated fserves." Update: 12/11 4:09 PM by michael: See also this article describing file-sharing over AIM - Aimster.
This discussion has been archived. No new comments can be posted.

New P2P tool Using... IRC? More

New P2P tool Using... IRC?

Comments Filter:
  • As someone who has just recently gone through the napster protocol piece by piece, I can tell you that not only do you already get a connection IP with every search reply (how do you think the windows client gets ping times?), but that it is also easy to track the files as move from user to user, with the md5 field.

    Granted, the md5 can be easily changed each time you upload your file list, but with most people using the same md5 hash length (last checked it was the first 299008 bytes) tracking isn't that hard either.

    I thikn that IRC is one of the natural mediums for this type of sharing network. The only problem I can forsee though is how strict P2P/IRC servers will be in allowing other servers to connect in. That is the essence of maintaining a decentralized network, and if you start setting up strict limits on who can connect back to your server, then you risk the health of the network.
  • I believe that CuteMX uses IRC as it's connection and transfer system already, or at least that's what they lead me to believe, based on their text within the program. They refer to transfers as DCC and integrate IRC chatting directly into their GUI. I could easily see how their querying and cataloging systems could be set up with a relatively simple IRC script. Does anyone have any info?
  • Here is a communication log: p_a: "Hello, anybody file XYZ?" p_b: "have it" p_a: "pls send it" p_b: "/DCC SEND XYZ p_a" p_a: "thx"
  • I don't claim to know anything about foriegn laws, but why can't one of these centralized services like Napster run in a foriegn country that has no real laws that would be broken?

    Can anyone who knows more about this comment?

    It seems like there will never be a decentralized file sharing service that works as well as Napster or Scour...

    FoonDog
  • I'm not a CS major.

    but from what i've read. O(n^2) would spell out too "Big Oh, N to the power of two" Big Oh -or- O() means the worst possible case for an agorythm to run. N is some variable. Big Oh notation is used to expresses the effenctentcy of an agorythm.

    another case of O(n^2) would be a bouble sort. in it's worst possible case, if you had N items it would take N^2 time to sort those items.

    What I don't understand is how is compares it to IRCnet, I can understand expressing it to expressions, and loops. but to a whole protocol is much more vague? What's N? in this case, the number of clients? what's ther result of n^2 apply too?

    -Jon
  • It means that the time required for the process grows at a rate that is similar to the curve generated by squaring the primary value that increases computation time (represented as n). "Big-O" notation is one of a couple different ways to express how a particular process' time requirements increase as the size of the job increases. If I remember right, Big-O represents the worst-case scenario (the time growth curve will be no worse than this). Combined with Big-Omega and Big-Theta, you get worst-case, best-case, and average curves.

  • How does this overcome the fundamental problem? (Score:3)

    by donny ( 165416 ) on Monday December 11, 2000 @07:33AM (#567475)
    Okay, I'm always skeptical when a couple of goonies manage to wrangle some VC money with a idea that has lots of buzzwords and promises to be really really hot.

    I'm not saying that this is the case here, mind you. I haven't looked into the particulars of this company at all (other than the fact that its home-page looks home-made).

    Here is my question:

    The thing that has been plaguing these glorified FTP clients is the real-time searching utility. It has an inherent O(n^2) feel to it. 50 users want to use the network, and search for files on 49 computers. 50000 users want to use the network, and search for files on 49999 computers. Seems O(n^2) to me.

    The Napster (Scour etc) solution is the central server, which people dislike, since these "P2P" fly-by-night companies are turning to the dark side at a rate not seen since the original Star Wars flicks came out.

    The Gnutella (Freenet etc) solution is "maybe n won't get so big". Or, in desperation, "let's not search everyone's else's computer". Doesn't seem like people would go for that. I'd like to find that one guy who likes the same obscure stuff that I do. Isn't that supposed to be one of the advantages of this type of file exchange?

    So that's the (billion-dollar) problem. Here's the proposed solution:

    Let's use the IRC network instead.

    Does something seem fishy here? How would this fix the inherent O(N^2) problems? It seems to me that these people are solving the O(N^2) problem by finding an unused network so N can start at 0 again. Does anyone else out there sense this?

    Donny
  • ANY irc client has the ability to share files quite simply through DCC.

    Why is this such a big deal?

    I could hack together a file-sharing bot in about half an hour using the Net::IRC module. Just let the bot idle on no channels, and respond to certain command strings.

    All this does is bring the litigation happy corporations down on the IRC...next thing you know, DCC will be banned. Good thing I run my own server...

  • Almost every single IRC server in existance has "Terms of use" that explicitly ban bots

    bullshit, and if they do. trust me that don't do a thing about it. damm near anyone can eggbot a channel as long as they want. look at google's [google.com] listing of IRC bots.

    Now I've never seen a shell account that allowed IRC bot's. but as for the IRC serers? every channel is kept open from a bot. hell they way IRC is hacked together, it's the only way to really keep a room/channel open. You have a bot sit there acting like a user. then you can log in and pipe commands to it like "/kick 33l3le_lame_ass" of whatever.

    -Jon
  • your connection IP is available to anyone online
    Not true. There are several IRCDs that support address masking:

    - CitizenC . AssMonkey@EliteOrbit-58437.ivideon.com (com)
    - realname í rm -rf /* - server øù. Drow.Portland.OR.US.EliteOrbit.Net (Refuge for all weary Travelers)
    - idle í 1m 7s (signon: -1h 6m 6s ago)

    /mode CitizenC -x

    - CitizenC . AssMonkey@64-59-157-35.ivideon.com (com)
    - realname í rm -rf /* - server øù. Drow.Portland.OR.US.EliteOrbit.Net (Refuge for all weary Travelers)
    - idle í 2m 6s (signon: -1h 5m 4s ago)


    ------------
    CitizenC
  • Oh damnit, it's not your underground Warez/mp3 scene at stake here. This is about the future of IRC. IRC servers are shut down one after another due to DoS attacks and massive bandwidth requirements already. And now someone comes up with something to increase the load of these already heavily used server even more. Great idea.

    The warez/mp3 scene will survive, but IRC might go down, and that's what at least I am worried about.

  • The thing you are all missing with concerns over server load is that the transfer is PEER TO PEER. DCC, for example, does *NOT* travel through the server. It's a direct link between two clients. That's why if you are lagged with somebody, you can always /dcc chat them, and if the link is good between the two of you, lag go bye-bye.
  • Where in my post did I say I have a problem with unispaced text? I think the CODE-tag is very handy when you need to display code, and I won't be happy if ROb feels the need to remove it. Just like SUB and SUP were handy (H2O, E=mc2), but are now gone because people felt it was ok to fuck around with them.

    And yes, you really should stop using Lynx, w3m is much better.

    --

  • Hmmmm irc..... (Score:4)

    by GoRK ( 10018 ) on Monday December 11, 2000 @10:29AM (#567482) Homepage Journal
    In a brief email exchange about 1 week after Napster existed (at least in a public sense), I learned that the original Napster servers were set up using essentially a modified ircd. A modified services handled search requests and the database. The napster protocol itself is a lot like irc as far as the authentication and message passing goes, and the hash's (#) in the channel names are a holdover that stayed in because it was convenient and familiar. It is interesting to note that he also told me that the ircd based system showed such poor performance in testing that they rewrote everything from scratch even before they opened the service to the public.

    I am skeptical as to how well this is going to work in practice. Anyone who remembers EFNet #quake on the day of the release of quake 1 should know how problematic having several thousand people in an irc channel can be.

    ~GoRK

  • Re:Gah, IRC protocol bad... (Score:1)

    by Anonymous Coward
    All I can say is take a look at the very cool filetopia [filetopia.com]. It combines a collection manager (including a diff like feature), BBS system (messages that are held on the server), IRC, and a few other P2P ideas.

    Note, it is closed source for the moment, but you just have to love their philosophy (read their philosophy page, duh!). Oh yeah, and it encrypts most traffic with strong crypto.

  • Take a look at filetopia [filetopia.com], and look at how they combine file collection management, friends lists, and IRC. Though the userbase is still small, it seems to work well.

    Looking for pr0n? Go to the xxx channel, download file lists from people, click on the tab that lists only those files that are missing from your master file list, then add files to hotlist/ignorelist/download list. It all works very nicely.

  • If research to-date is representative of reality (and it probably is), Freenet has log(n) scalability, in this regard it essentially solves the scalability "problem" that supposedly plagues P2P.

    How come whenever P2P is mentioned everyone starts to spout half-truths (or just complete mis-truths) as if they knew exactly what they were talking about?

    --

  • http://www.mp3rc.com We released December 5th. There are a few issues.. but we are working on them and will have a very functional release late this month. We are also going to release the server prog so anyone can run their own.
  • I know addresses can be spoofed, I'm simply saying that if someone owned the connection (Extended Domicile Act needed perhaps) then they could sue people for trespassing.

    Maybe a net watermark could work.
  • Almost every single IRC server in existance has "Terms of use" that explicitly ban bots. This is very much a bot. It won't scale well either because if a user on IRC is transfering too much data at once, the server kicks them off for flooding. In other words if three modem users request a search on "slashcode.zip" and it replies to all of them with 5 found locations, they go down. You also have to solve the problems where the servers are bogged down and users can't connect, k-lining the main bot, the RIAA using some l33t mIRC script to take out the bot, and more. Why someone just doesn't copy the napster service, generalize it, and run it in a third world country not regulated by US copy right laws is beyond me. Why not use Cuba, they're preety pissed at us now?
  • How are they going to route all those packets around the sniveling 14 year old netsexers?

    krystal_blade

  • "explained Alan Greenspan, Chief US Economist" Point of interest here; Allan Greenspan isn't Chief US Economist, he's the Chairman of Board of the Federal Reserve Bank of the United States. This is arguably the most powerful economic institution in the world. He has his hand on your wallet so to speak. Take a few minutes to find out what he actually does. I'm sure it will be worth your time. If you think Bill Gates commands a financial empire you haven't seen anything yet. Funny stuff!
  • While IRC servers usually say they ban bots they don't usually actively seek them out, so you've got a ton of them on servers where they're supposed to be disallowed.

    There isn't the risk of flooding due to data transfer because the actual file transfer is done peer to peer. It's not like the client is going to barf UUencoded files into channels.

    I assume how it would work is that the client would join some channel(s), and ask something like "Anyone got beautiful day by U2?". All the other clients in those channels would then do a quick check for matching files and message (preferrably directly, DCC-style) the requesting client with the names of matches so that they can initiate a direct file transfer.
    ---
    Where can the word be found, where can the word resound? Not here, there is not enough silence.
  • ...seeing that IRC is based on FTP anyway, the idea itself was probably a no-brainer. Let's just hope that it doesn't get too popular and mainstream (like Napster did).
  • That's not a useful measure of scalability. Last I checked, Freenet isn't that searchable. If you wanted to search it, you would need to keep lists of what's around. As more people put more content on, and more lists get circulated around, then, we need bigger lists of lists and then that makes it... um... uh... n^2.

    I'm not sure what kind of scalability you're talking about, but it sure isn't searching capability. (If you didn't have to search for these damned things, wouldn't you just connect directly to whereever the thing was with FTP or something?) In this sense, Freenet's aim may be to secure documents in a non-removable and dynamic way, but search scalability is still n^2, and if 6 billion people decided to put their mp3's on, good luck building your collection of South African jazz mp3's.

    How come whenever P2P is mentioned everyone starts to spout half-truths (or just complete mis-truths) as if they knew exactly what they were talking about?

    Good question. You tell me.

    Donny
  • Shouldn't this be expected? With universities banning napster, Scour closed, CuteMX sucking, Gnutella toasted by modem users, many people who have been in the market of exchange since the early 90s went back to IRC chat channels for file exchange. The plague of pirates that once flooded the effnet network let up a bit when there were other, simpler alternatives (which didn't have a 40:1 ul:dl ratio). When their alternatives were taken away, what, they stopped what they were doing and found something else to do? No, they went back where they came from, but now they are spread across more networks and Dalnet has a bigger community. So it really shouldn't surprise us very much that IRC is the choice of programs for fileswappers.

    Regardless, it sounds like a good idea. But IRC servers are fragile things. Look at the dal-net network, and watch it's servers go up and down, up and down. Netsplit here, server crash there, services toasted for another 13 hours, and four thousand random messages from 12 year olds pushing porn down your throat. The problems of speed and stability faced by Gnutella may not be completely solved as it may seem, unless someone feels like shelling out some serious cash and setting up some slashdotable servers on irc.

  • Comment removed based on user account deletion
  • ...Got the t-shirt [www.var.cx]. Then killed it off, because there were better alternatives.....
    It's called Distributed FileServer over IRC, and is basicly like gnutella but using irc as a carrier.
  • This is almost exactly what the opennap network is, a setup of interconnected servers and clients that connect to a server(Like IRC). But the napster network was designed from the ground up to support file sharing.

    However, the existing napster network was a corporate entity, which was its major problem.

  • it is simple repackaging of existing ideas which look suspiciously like FTP ...

    You could have said similarly that http & browsers are "simple repackaging" of ftp. After all, using ftp and text files (containing addresses), with some cut & paste you could do with ftp anything you could do with http, html & browser. Yet, the internet didn't quite take off until the http & browsers made it easy to perform those operations (hyperlinking, viewing text & graphics, downloading files).

    In other words, repackaging which substantially lowers the usability threshold makes all the difference. Similarly, the aim of various P2P designs, mashrooming all over in recent months, is to lower the threshold of turning every computer connected to the internet into a a web server, chat server, file server, message board,...

    While you can certainly do this already with most connected computers (other than for some NAT/proxy configurations), and while dedicated servers have been doing it for years, you need much greater technical expertise than an average web user has (and often the additional ISP cost for fixed IP). When someone manages to convert that into a few simple, visually intuitive point & click operations, the P2P field will take off, just as Web did.

    This evolution will inevitably occur since the increase in the bandwidth, the CPU speed & the storage capacity creates a void calling to be filled.

  • It's called a blatant attempt for hits. Anyone on earth can see how quickly this is going down...IRC servers split all the time! IT'S JUST OVERHYPED. Leave it at that...and if you really want some detailed information about it, mail me your questions. [mailto]
  • Something a 15 year old would think of. O wait, it is a 15 year old.

    Using IRC, besides the fact that it has it's own scalability issues, is not the way to implement a P2P network! If the goal was to avoid 'central servers', using an IRC server doesn't get you any closer to that goal. Instead of placing the load on your own systems, you are just outsourcing it to public chat systems (without permission might i add since most servers excplititly state no non-IRC client 'bots').

    (http://www.dIRCchat.com - new Windows IRC client)

  • There are much worse algrorythm scalings than O(n^2). A trivial one would be O(n^n). At 100 items a O(n^2) would have increased by 10^4 times, while a O(n^n) would have increased by 10^200.
  • piracy is one thing, but any tool that will let you freely exchange information with out the fear of being smacked down by the man is protecting your rights as a human. do you like the idea of your every move being monitored by someone some where? do you like the idea of government storm troopers knocking down your door because they didnt agree with what you said? move to cuba or something.
  • Hmmmm... What exactly is a router? Let's see, operating system, listens for requests on ports, transmits replies. Yep, it's a server. So really, any time you are using the internet you're using "someone elses server".

    BW overuse? Bah humbug, responses to searches can be a dcc chat.

  • Quote from that page: "Please note that it is our policy only to maintain information from sites whose system administrators have given their approval. "

    Hmmmm


  • Dear sir, Im having problems trying to transfer a file over your mIRC network. Its only 14 Meg and I am using a 56k line. It gets to about 400K and keeps dropping out, stalling, or just crashing. Why does this happen?

    On, and yeah, since youre reading this already, can you also do something about Galt? He keeps kicking and banning me from #hottub but I did nothing wrong. He wont even acknowledge my numerous private messages to him after the fact. What is wrong with him?

    Thank you,
    Signed - lamer newbie #1038801

  • splooge.com? Is this porno related?
  • It's a counting argument more than an algorithmic one.
    It's not about how much work you need to do, it's about how many oppertunities there are for needing to do work.
    The "work" is synchronising server state between servers. There are approximately n^2 pairs of servers (n*(n-1) more accurately). So there are n^2 chances of a 2 servers being out of step with each other.
    There's a far better explanation of this "communication/interface" scaling in Fred Brooks' Mythical Man Month, but that applies to humans rather than servers.

    FP.
  • I have a better explanation.
    If a server changes state on average X times an hour, then one of N servers will on average change state NX times an hour. This change needs to be propogated to N servers. Therefore the work required is N.NX communications per hour.
    O(N^2)

    FatPhil
  • IP addresses are easily spoofed, bubba. Try again.

    IPSec and IPV6 work to remedy this, but if you are using vanilla tcpip with IPV4, your data simply cannot be trusted, nor can you inherently trust data from other sources.

  • Comment removed (Score:5)

    by account_deleted ( 4530225 ) on Monday December 11, 2000 @06:19AM (#567510)
    Comment removed based on user account deletion
  • If I was to generate something like this - and this is an excellent idea - I would do the following:

    Set up an irc server. Just a plain-vanilla irc server.

    Then, I would create a program that will join a specific channel on that server. The user would NOT be able to see what goes on in that channel.

    When the user searches, it pops up in the channel as "request titney spears". All other clients search their internal share lists.

    If one or more clients finds that file, they will send a private message to the requesting client with a specially formated response.

    The requesting client will then display a list of the responses as they arrive.

    You might also want to include a simple irc client, and let the users talk on different channels.

    File transfers can be done via DCC, or a separate system, like an ftp server/client on a customized port.

    So, all you need is Glade, some libs and a bit of elbow grease. Who's with me?
  • Two things:

    1) IRC is already muffed up enough already. Look at dalnet and efnet for christ sakes. Dalnet has always had lag issues between long distance servers, meaning people wishing to chat in regular groups usually use the same server (defeating the purpose of a network). EFnet had it's "OMG we're dissapearing" scare earlier in the year. And it's overrun by lamers.

    2) There are ALREADY mp3 channels. They are not that hard to find, or use. To return idiot newbie and/or idiot industry attention to these souces, which have improved since everyone started using Napster, would be brutal to say the least.

    3) (I know) OpenNap, Audiogalaxy, gnutella, freenet, do we really need another one?
  • I've been on IRC for years and all sorts of stuff trades hands there. Anything from MP3s, pictures, and warez can be had in plenty. The problem with IRC has always been that it's too damn hard to figure out how to get that stuff - at least for the casual user.

    Interfaces that rely on IRC (and DCC), make it easier for the average Joe to use.

    Real-time indexing of the Internet coming soon [grub.org]!

    Kord

  • Always sketpical when I see "revolutionary" (Score:5)

    by Mike Connell ( 81274 ) on Monday December 11, 2000 @06:21AM (#567514) Homepage
    There's no information about how it uses IRC in the PR guff or on the site. Does anyone have some *real* information?

    If they're passing messages over IRC, either they'll have to run their own (and risk being shut down), or put the messages out over existing IRC networks (EF,Dal, whatever). In the second case, I can't imagine the guys running the IRC servers are going to be too impressed - the onus of fending off irate lawyers will then be on them, and most of the IRC servers seem to be run out of kindness as a hobby.

    0.02,

    Mike.
  • As a matter of fact, I am a Windows user myself (I recently upgraded to ME, which was a mistake) but only because I can't wrap my head round Linux. Windows, on the other hand, is fairly (I said fairly) intuitive and users tend to catch on quickly. This is what I am getting at - if it was made easier to use it would become more popular amongst people such as myself who only use Windows because: (a) they can't be bothered with the complexities of Linux, or (b) lack the technical knowledge to use it. (And I have tried BeOS, and I don't like it. It is easy to use, but multimedia support seems rather flaky to me.) But having said that, if it were made more user-friendly, functionality would probably be compromised, and in the end we'd just end up with...another version of Windows. Eugh.
    To hell with PCs. You know what I think? We should all go back to using Amigas. They're much better than PCs.
    No, seriously! Amigas kick ass!
    _____________________________________

  • Isn't IRC overloaded already? (Score:4)

    by fatphil ( 181876 ) on Monday December 11, 2000 @06:21AM (#567516) Homepage
    I got the impression that the O(n^2) nature of IRC made scalability a problem (every change at every server needs to be propagated to every other server). Similarly, this was initially designed for sending 50 byte packets of gibber-gabber, not whole files.
    I suspect I'll hear it start creaking at the edges fairly soon...

    FP.
  • I can't understand why anyone would say this system does not use a centralized server, it most certainly does, and it's IRC. Remove access to IRC, and the system no longer works. However, I don't disagree that its not an innovative solution to the problem. We used IRC for a similar purpose from around 1990 to 1993. We wrote an encrypted protocal using IRC as a network connection point for our chat systems. It allowed us to link remote chat systems without having to pay phone bills. We basically hacked user accounts on university systems to gain access. Was around the the time pc-pursuit went down. If anyone was around back then the chat system/bbs was STS-12, and in case anyone was wondering, the channel was #freedom. Anyhow, the only real way to do this efficiently and reliably is with a central server based system. We did alot of research on Gnutella and non-server based networks, and we came to the conclusion that while it works great if the point of the system is anonymity and piracy/copyright fraud, it just doesn't work in any business models. "We" are the engineering crew at http://www.splooge.com [splooge.com], a new (beta) media-sharing application. Don't really want to use this as my own advertisement system, but check it out, we've spent a tremendous amount of time on it.
  • maybe this will finally be an underground, non-mainstream type deal that people can use without fear of the government coming in and intervening. that was the big problem with napster, it spread too far. let's hope this one stays concentrated to those who will use it responsibly.
  • I don't see this new p2p software as a overuse issue for irc servers, people are already talking on them all the time:). I think the software must open up discussion groups named after the software or mp3 you are seeking (irc chat room Metalica - Battery :) with the info tag of the group containing the information of where to find your software/info. In the info part, there would be several addresses, such as 198.58.113.7 ... to the peer that has the software/info you requested. The client would than proceed to attempt connects at the addresses until it hit one that worked.

    This is all speculation. It would be interesting to see how it actually works. (It could work by continually posting address's in the discussion room from people with the ware/softwares computer, overuse of the ircs I guess)

    It is shady to use someone elses server to steal warez/music though - Even from people who encourage theivery among youth with their music (like Dr.Dre).

  • This Red Herring article [redherring.com] lays it out pretty nicely. P-2-P is not any new innovation, it is simple repackaging of existing ideas which look suspiciously like FTP or, in this case, IRC. The specific uses might seem interesting, but now that people are starting to see that there isn't really any great leap forward here, they can get back to just designing useful file transfer systems and leave the gee-whiz hype at the door.
  • We need to stop allowing quacks like you to yack beeven if it is legal because it will rot the net.
  • Alright, I'll take your advice and post rather than modding.

    I still think you're wrong. And possibly trolling.
    "We need to stop making new tools, even if they can be used for legal purposes." That is just patently ridiculous. If people had that attitude, then Nuclear research would have stopped before it started. Sure, we wouldn't have nukes, but then we'd also have nothing like chemo therapy and the advanced Diagnostic imaging techniques we do now.

    Think about your logic before you post.
  • Other's have already defined big-Oh so I wont do that. I'll show you (in a simple case) what it means as far as runtime.

    Suppose you have:

    for (i=0; i<n; i++)
    // something that takes 1 second to do;



    Assume it takes 1 second to complete each iteration. Now if n=1, runtime=1 sec; if n=2, runtime=2 sec; etc. Thus, the runtime is equal to n*1 or more generally, n*(time to complete one iteration). Thus, this algorithm's runtime is O(n).

    Now lets say you have this:

    for (i=0; i<n; i++)
    for(j=0; j<n; j++)
    //something that takes 1 second to do;


    Let's look at the inner j-loop first. Notice that its identical to the above example that we know to be O(n). However, its enclosed within another loop that iterates n times. So the runtime is (outter loop iterations) * (inner loop iterations) * (time to complete inner loop iteration), thus n * n * 1 which is (n^2) * 1. Thus, this algorithm's runtime is O(n^2).

    The important thing to take out of this is how each algorithm's runtime increases as n (or the amount of work) increases. For O(n^2) algorithms even small increases in workload can mean pretty nasty performance hits. Of course, it's even worse for cubic (O(n^3)) algorithms, etc, etc.
  • If everyone had an IP address they could use (not assigned just available) and if people OWNED said connection then it would be possible to work out internet voting. But since no one owns the connections (IP to IP communication not wiring) then it's impossible.
  • Gee, nothing like building a massive data transfer system on top of an incredibly stable, dynamic, up to date system like IRC.

    Everyone knows that IRC is a quick, responsive way to communicate, as well as very stable and dynamic.

    -------
    CAIMLAS

  • 2 or 3 months ago, wasn't it? I don't remember the article's name, but slashdot covered decentralized and *encrypted* networks over irc --- all this current article is about is that someone is actually putting this idea to use, formally...
  • Yea, if you need, we can get back to you with whatever legal information you need. Our system is most certainly centralized. It's really the only logical option. We know what O is, and we respect her. Let me just add.. We are a Zero VC company, thats right, $0.00 VC, we have no investors, and to date have only spent some $10,000 out of pocket on hardware and bandwidth. That pretty much makes us one of the most successful internet companies around. :) -James

  • The other day I tried a new napster clone and was supprised to see it connected me to dalnet after installation. VNN 2000 [vnn2000.com] is the Java based p2p that I'm referring to. It suffixes your name with _VNN or somehting and tosses you into #vnn on dalnet. It's not supremely obvious at first, but you do get their motd and such. Just recently I believe they have had some trouble with the irc end and may have dropped it.

    On another note, what happens when they take it to the next level? Will we ever have to worry about ftp and such?

    -XtAt
  • While I am not a big fan of IRC, especially the late EFNet, I am no bigger fan of ICQ. ICQ uses a lot of client-side security, which means that a simple patch of the program (or an alternate version of it) allows you to do things which you should not be able to do, such as adding people to your list without their permission and seeing people on your list even though they set themselves to be "Invisible".

    And now, a few comments on this "revolutionary" file sharing idea. Napster (the program) was originally written because Napster (the person) was annoyed with having to manually go into IRC channels with offer bots running in them and then manually picking files to download and keeping track of it. Napster was, in fact, designed to be a more convenient version of an offer bot network. Does anyone see the irony of going back to IRC?

    As for the comments that Napster is in danger of being extinct, take a look at Napigator [napigator.com]'s server listing page [napigator.com]. There are at least 10 different networks, all of which are pretty large. Though they cannot yet be compared in size to the Napster network, they will explode the second that Napster is taken down.

    --

  • Sure, we wouldn't have nukes, but then we'd also have nothing like chemo therapy

    I never said give up all research, I was specifically referring to P2P filesharing tools. If people keep making more of these, the internet will be heavily regulated by the government. They will see all the piracy from the "open source geeks", and get laws passed to stop it.

    One of the only ways I can see to prevent it is to stop making the tools. If we stop making the tools, we will have a better name, and they will leave us alone. How many legitimate purposes could Napster have anyway? Wouldn't it be easier to go to mp3.com or something? P2P utilities cause nothing but problems.
  • Last time I joined an mp3 channel in IRC (been I guess about a year ago) everyone was using an mIRC script that generally does what you are talking about. If I remember right it was called SailorScript or something to that effect.
    • Algorithm, not algorythm.
    • Omega(n^2), not O(n^2).

    Informally, a function is O(n^2) if it grows no faster than n^2. A function is Omega(n^2) if it grows at least as fast as n^2. The latter is at issue here.

  • A bot to keep the channel open? That's what chanserv is for :). However, bots are good for logging and topic management (ie, using the topic as a message board, six lines of topic, anybody?:). Bots have their place on IRC, but keeping a channel open is not it (what if the bot PTOs?).

    Bill - aka taniwha
    --

  • I think some of the people saying IRC couldn't handle it may be forgetting about DCC and CTCP. DCC "Direct Client to Client" is a protocol for IRC clients to open connections directly to each other, bypassing the server. CTCP is a protocol for sending non chat message type signals between clients (relayed by the server), for example, the /ping command. To initiate a DCC connection, it's just two little CTCP messages going across the server, and the rest never touches it. Want to connect to a bot that is acting as a directory? It's only 2 little CTCP signals through the server, probably requiring less resources than a typical conversation. Want to download a song from someone? 2 more CTCP messages per request.

    What's even better, IMHO, is that we don't even need to write new IRC clients. Many popular clients already have ample scripting facilities; All we need to do is define a protocol for indexing and searching and then write some scripts to support it in existing clients.

  • Um, if I'm not mistaken this has been possible for years.

  • People once thought Napster was untouchable, too.

    Ahem! People once hoped Napster was untouchable. Any person who thought they were untouchable probably thinks that if they remove the numbers from their house the cops won't be able to find them.

  • actually, if you pay attention, when it looks like The courts are starting to believe Gore's Bullshit, the stock market begins to crash. 'Nuff said

  • It just goes to show you that you can't block peer to peer filesharing without requiring end-to-end authentication for all Internet connections. Don't laugh; the MPAA/RIAA is probably working on this right now.

    Change the CTCP (client to client protocol) VERSION response to mimic that of a standard IRC client, make the protocol look like a few people talking, and change the channels you use, and it will become reasonably hard (albeight not impossible) to systematically identify these programs.

    If I remember right, Napster originally was based off of the IRC idea as well. Unfortuately, IRC offers little anonymitity as to the source. An IRC server put in debug mode captures all the messages sent through it, and your connection IP is available to anyone online. We'll see how this one works. Hopefully we won't end up with a bunch of additional people flooding the system to take control of channels; IRC is overburdened with DoS attacks against servers already.

  • gnutella can be used this way. there are several private and underground gnutella "nets" already out there. my friends and I have one (invite only, sorry.) and with 20 people sharing everything instead of one person sharing and 19 moochers it works great.

    gnutell works for the "underground" idea already. and it can handle it.
  • I used to use IRC fairly extensively and servers would regularly go missing creating a fractured and fragmented network - response times were laughable - 2 or 3 minutes between someone sending a message and the recipient getting it was not uncommon. Now I find ICQ a much better way of contacting the people I want to chat with - they all have ICQ numbers and I have much more flexibility - even if the software is pig ugly and eats memory faster than Outlook.

    I hope that these guys are not planning on layering this service over existing networks becuase they're going to have to make sure that sys admins can't identify the clients to K-line their users. Adding Napster and Gnutella amounts of traffic to already busy IRC networks is just asking for disaster - I know of a few ISPs that would like a good reason to remove their IRC servers.

    Personally I think that file sharing en-messe needs dedicated protocols and servers - or at the very least extensions to existing protocols and dedicated servers.

  • ALL it is, is another company trying to get their piece of the P2p pie. Look at aimster, songspy, cutemx, etc. (Do I really need to go on? ). I think, instead of coming out with a new protocol, we should embrace one as a standard. Opennap for example, is a great, OPEN., standard.
  • "promises to address problems which have plagued other similar applications both legally and structurally."

    The article cites Gnutella's centralized load problem. How will this differ on IRC? Aren't you just adding massive load to whatever network you utilize, be it Dalnet, Efnet or what have you?

    The article also cites Napster's legal problems. How is utilizing IRC going to change this? People once thought Napster was untouchable, too. You'll just move the battle to IRC and the people who run IRC servers. Passing the problem on to someone else doesn't help.
    ---
    seumas.com

  • I have to disagree with this article. Windows 2000 is obviously the answer to our computing problems and is most certainly superior to Linux (et al) in every way.
    Hmmm. I'm not too sure how you work this one out, and I must disagree. Any incarnation of Windows you care to mention is bound to be worse than Linux in many ways. If I could get my head round Linux, I would use - ah. I see what you mean. Linux is very complicated, to be sure.
    If Linux could just be made more user-friendly without detracting from its immense configurability (something that, as yet, Windows cannot offer) then I reckon it would rapidly become the most used OS in the world.
    We can only hope...
    _____________________________________
  • I was playing in a field once, and our baseball accidentally hit one-a them there bithives, and all them bits got real mad and were byte-ing us, and a-stingin' us... maybe if we-un's 'ad had some-a that there P2P we coulda killed all 'em bits...

    P2P, is that kinda like that DDT stuff?

    yes, it's a really lame joke but I've had very little sleep - roto-rooter is my friend, and I LOVE mopping the basement floor with bleach

  • Poor old IRC (Score:3)

    by yebb ( 142883 ) on Monday December 11, 2000 @06:56AM (#567545)
    Sweet jesus, as if some of the IRC networks arn't already overworked enough. Any of the last few battle scared EFNet servers will finally cut their losses, and say "To heck with you all" while they pack up their server, and twine up their twisted pair.

    Although maybe some people who have a commercial interest in this could setup some IRC servers that can handle serious load, and will scale with the amount of usage.

    If theres money to be had, it could very well mean the introduction of new IRC servers. Maybe this will be the best thing that has happened to IRC in a long time. Although, maybe it will just mean more splits, and slower servers.

  • I've been wondering when someone would figure this one out. Let's see. IRC may not have been the first information relay protocol, but it certainly one of the most tried and tested... Code reuse. What a concept!

    "I need eggdrops. Lots of eggdrops."

  • I remember finding an FTP site, digging around in messy directories hoping to find the MP3 I was looking for...

    With Napster, I can actually find MP3's of music I actually like and regularly listen to (punk/progressive rock), not the ABBA and 80's pop tunes I'd download and play just to get them out of my head.

    It's not about capabilities -- it's about making things accessible to the less tech savvy.

    (end comment) */ }

  • I think I'll run it over EFNet!
  • Ooohhh, poor baby. It's torture to look at unispaced text, isn't it?

    Wait a second, this is ALL unispaced to me.

    Maybe I should stop using lynx.

    you fucking crybaby

    -Chris
    ...More Powerful than Otto Preminger...
  • It seems to me that some people are confused about IRC's role in "BitHive". I am under the impression that IRC will be merely used as a host to keep track of active connections on the net (ip's willing to share files)...not to do the actual file transfers or to keep a running database of available files (like napster's central server).
    An earlier post made the comment about IRC servers just being a hobby of the SysAdmin's. I would agree that this is primarily true, so inorder for BitHive to be successful the current IRC model can't be compromised or changed.... maybe just "borrowed" - (taking the current technology and using it for other ways than it was intended.) -WebWiz

  • "Unlike Napster, which is becoming subscription-based, Scour, which has succumbed to legal pressures, and Gnutella, which suffers from scalability issues, BitHive relies on no central servers or corporations to run."

    Neither does Napster. With Napigator [napigator.com], users can connect to OpenNap directory servers [sourceforge.net] and share their files without needing some big corporation's help (unless you count VA Linux's SourceForge, which hosts the OpenNap website). And this To demonstrate the legitimacy of OpenNap, simply make a Linux kernel [kernel.org] tarball available on one of the servers, and run an OpenNap segment on your local network [everything2.com] to ease the bandwidth problem. With that kind of cred[?] [everything2.com], RIAA won't be able to touch it.

  • Anyone remember archie? (Score:3)

    by peterdaly ( 123554 ) <{petedaly} {at} {ix.netcom.com}> on Monday December 11, 2000 @09:35AM (#567552)
    Remember Archie? I'm only half joking...

    The archie information system is a network-based information tool offering proactive data retrieval and indexing for widely distributed collections of data. The archie Data Gathering Component automates the gathering, indexing and maintenance of information, allowing information providers to offer improved resource discovery and access to information.

    The archie User Access Component allows your users to locate and access your information using a variety of interfaces and search methods.

    Given the number of hosts being used as archive sites nowadays, there can be great difficulty in finding needed software in a distributed environment. You may know that the software that you need is out there, but it can sometimes be difficult to find. Perhaps the best known application of the archie system is to maintain this Internet Archives database. The database, already available from a number of service providers across the Internet, currently contains the names of over 2,400,000 files at over 1,000 anonymous FTP archive sites. Using this database, users can rapidly locate needed files without the need to log onto dozens or even hundreds of machines. archie servers offering this database currently receive over 50,000 queries per day.

    http://archie.emnet.co.uk/readme.html [emnet.co.uk].

    -Pete

  • Gnutella passes queries amongst all peers, providing anonymity in the queries and avoiding the need for any centralized indices of files. The problem is that now logging onto Gnutella swamps most clients with zillions of queries which your computer will have to forward to its peers.

    IRC decentralizes the indices. That's about all. I always figured this would be the perfect replacement for ICQ, but file sharing works too. When I first logged onto Napster, this is exactly what I thought it was... a pretty shell for some magic IRC channel.

    Neither system provides anonymous transfer of information. If I request a file from you on IRC, I can pick out your IP through the DCC connection.

    If you were using IRC instead of ICQ you could maintain contact lists on servers... you create a password-protected channel and invite your friends into it. When they log on, they automatically log on to the appropriate channels. By querying people's info you could send them DCC chats, files, and other requests. Very simple, very standard, very open.. no more IDs and contact lists either. The only trick is verifying the identity of your friends, but Email works fine for that.

    All IRC lacks is a tallbar UI.

  • It was called IYoink. It didn't work very well. I don't think BHive will either. IRC servers are too unstable.
  • Running an IRC server costs thousands of dollars per month to (mostly) large companies with big legal departments. There is (to my knowledge) no such thing as a Public IRC Network: all servers are owned by companies who pay for them and rightfully reserve the right to throw out anyone doing things they don't like with their bandwith. In fact, most of these companies host an IRC server to provide service for /their/ clients, but allow others to use it.

    Now, will companies that host those supposedly "public" IRC Servers that are targetted by this software just stand there and watch while some other startup just wastes their bandwith with file trading bots? Does this make sense from a buisiness point of view?

    Somehow I think not. I know AT&T (disclaimer: I have no relation with AT&T besides being an IRC Operator on one of their servers) would not want its servers used to help some other software work while not getting anything in return. What will happen? as soon as these clients are identifiable, they will most likely get automatically banned, or the trading channels closed. I for one am not going to tolerate these on my server taking up connections instead of people who actually want to chat.

    In either case, moves by BitHive to try to go around these bans by say, pretending to be a standard mIRC client would probably get them in some legal trouble. I don't know on what basis, I'm not a lawyer, but legal departements can be so oooo creative =)

    In short, the idea is good, but forgets one tiny thing: in the end, IRC servers belong to companies who don't like someone else just jumping in and using their paid bandwith.

  • I think you are definitely onto something here. (Of course, I am biased since for the past year we've been working on a system that does exactly that.) Check out http://www.peermetrics.com or send me an email for more discussion about this (if interested.) Of course, email would be easier, but look at the project info at peermetrics; it might give you some ideas... Peter
  • I know! I was trying to add a little comedy into the conversation. splooge? I mean, come on!
  • Why? Because IRC still uses centralized servers. I've never seen the RIAA mention IRC,
    so basically it's like a haven for people who pirate music, software, and kiddie porn (gasp!).
    No one is going after IRC.

    Running software which claims to be non-centralized _on top of_ a centralized system is just silly!
    If decide they want to shut down the new service, all they have to do is go after the IRC servers.
    And you know that when it came down to it, the people who run the IRC servers would buckle in a minute.

    So.. If you use IRC to trade files on a regular basis, I wouldn't be supporting this idea.

    Just a thought..
  • I was thinking the same thing... In fact, I
    just posted a reply saying pretty much the same
    exact thing.

    IRC is and will always be the best way. No one
    in the media ever mentions it! It's like the internet
    used to be..


    moo [2y.net]

  • Gah, IRC protocol bad... (Score:4)

    by Masem ( 1171 ) on Monday December 11, 2000 @06:31AM (#567560)
    IRC is a rather antiquated and outdated protocol for what it needs to do. It's not the concept of multiple servers in leaf/hub connections that's the problem, but the protocol that is used to send messages between servers and the various types of messages that a user can initiate. There's a lack of encryption (in general, some servers have that hacked in), spoof protection, and various other details that while most people think should be added to IRC, no one wants to reprogram all the clients and servers from the ground up to add said features. Previous attempts to try to 'extend' IRC by various tricks, such as MS' infamous Comic Chat, tend to pollute the system more, and if these people set up a standard on their own IRC network, some other networks that are meant for chat only might see weird requests from people trying to leech from EFNet.

    If anything, instead of starting with IRC, just develop a new protocol that would add the needed features in a desired P2P setting; don't use the IRC protocol as the model, save for how leaf/hubs work and how a message is propigated across a network.

  • Yea man.. Next time your on, actually take a second and read the server message. 90% of the servers I log on to ban bots.

    Of course, that doens't mean people still don't use them.. I think they just have the message there so if you're ABUSING the bots then they can boot you while they get to feel justified and rightous..

    Note:This is not a link to goatse.cx [2y.net]
  • Anybody that uses IRC will see that this is a very bad thing. IRC has always been a breeding ground for the underground MP3 and Warez scene. Everyone on IRC just laughs at Napster and the likes. If you want to get full albums you know where to go and that does not include the word Napster. IRC has always been the best place for the hardcore bootlegger and gamer that wants to get off cheap. Why would someone make something like this to gain usage of IRC? This will destroy the underground scene and eventually lead to IRC being monitored more and more. This is a very bad thing for anybody that currently uses IRC as there means for getting stuff.
    >neotope
  • When there is already a freely available, generally used, constantly updated, and almost endlessly exstensible IRC client already available for Windows, why are you making another one? And if you do, can't you at least give it a name not governed by (lowecase consonant)IRC?
  • got the impression that the O(n^2) nature of IRC made scalability a problem (every change at every server needs to be propagated to every other server). Similarly, this was initially designed for sending 50 byte packets of gibber-gabber, not whole files.

    As I understand DCC, which is used for actual file sharing, is peer-to-peer. It's the directory service part that will work using IRC. In that light, it will not be too heavy weight but you are right about IRC having scalability problems.

  • Oh, this is just great...

    As if having 90% of the kewl warez lists on the IRC wasn't enough, NOW they're trying to monopolize the TRANSFER of warez, whether you "IRC" or not...

    It's a conspiracy, I'm telling ya.

    krystal_blade

Slashdot Top Deals

To the systems programmer, users and applications serve only to provide a test load.

Close