Finally Real P2P With Brains

dfelznic writes: "The mp3 archives of CodeCon are now availble, which is news in itself. But what makes this real interesting is that they are being distributed by BitTorrent. BitTorrent allows users to download a file from multiple different people. Instead of everyone nailing one server, users get the file from other users. Furthurnet uses a similar technology to distribute legal bootlegs of concerts. The archive is available at the BitTorrent demo downloads page. As soon as I started downloading (cable modem) at around 300k I got a request for the file and began uploading at 40k. This could be the answer to the slashdot effect;) Now, who is going to be the first to complain about the use of mp3s instead of oggs?"

Comments

Nice.

[Renraku]

Nice idea, I have to say, but my biggest problem with file-sharing utilities is the fact that the file you're looking for isn't going to be the same with everyone. NudeCheerleader(part1).mpeg isn't going to be the same as NudeCheerleader(part1).mpeg on someone elses comp. There's not a way I know of besides implimenting CRC to prevent people from just renaming files into other things. Maybe NudeCheerleader(part1).mpeg is really GoatseLiveVideo.mpeg, just renamed.

Re:Nice.

[shankark]

That's true. But I think a workaround for this would be to have md5 signatures computed for each of these parts and verify them before they are downloaded. I'm not sure if this isn't being done by others already.

Re:Nice.

[Anonymous Coward]

This is what MD5 checksums are for.

Re:Nice.

[stype]

furthur does use md5's to tell concerts apart. I've seen a lot of incomplete files on other programs with the same name that get mixed up with the complete ones, but furthur makes it really hard to access files before the download is done. It doesn't really get the files in order necesarily. Its possibly the greatest piece of code made my hippies since unix.

Re:Nice.

[dytin]

The size of the file is often used to determine if two files are the same. Even if your two files are similar in size, they will not be the exact same amount of bytes, therefore size is actually a very accurate way to find if two files are the same. I believe that Morpheus used this method.

Re:Nice.

[ironfroggy]

I think they compare other things, such as the extra info (title, author, etc) as well as dates and file sizes. I've seen (on morpheus, gnutella, etc) many times when the same filename comes up as seperate results.

Re:Nice.

[jerryasher]

I believe this is what Bitzi provides (or was supposed to?) -- a way to register files and lookup various pieces of information:

With Bitzi:
* You can look up descriptions, comments, and ratings about your files - or contribute such info yourself
* Our precise digital fingerprints match info to exact files, so you can distinguish between similar files and search for the very best versions
* Future file-sharing tools can assure you of a file's contents before you begin downloading
* Infected or mislabeled files can be flagged, and so discovered or ignored before doing any harm

The Bitzi catalog is an open resource built by a community of fans, developers, and creators. To get started:

Re:Nice.

[Anonymous Coward]

http://furthurnet.com

Re:Nice.

[Anonymous Coward]

The problem with that method was that you could insert files of the exact same size that were different. For instance Morpheus used Morpheus itself to update itself. It would have been trivial to write a virus and then padded it until it was of the proper size. Then as long as the file was downloaded from only the infected source it would have installed the virus. Otherwise it would have corrupted the file, which would not have been desirable either (although not as dire as a virus).

A MD5 sum would have been a better way to do things... Of course none of this matters now, at least for Morpheus.

Re:Nice.

[NorthDude]

It's really interesting, but there's is still something which I don't get. Despite the fact that P2P is really interesting for Joe User who wish to download gigs of MP3's and x movies of all kind (X as a variable not as XXX :-) I can't find a real use for this technology yet. I mean, if I need to find valuable information, as in documentation, data, specs or anything, it won't help me at all?!? It is mainly used to share "binary files" for now, nothing prevents it from sharing "text file" like web page, xml files (if ever I would need one of any kind) or even PDF(I know, those are not text). But currently, one have to search based on files name/meta informations. Not really helpfull if I search for a websit talking about block cipher for exemple. I understand the need for real sharing of data as in information, and it includes media files as well, but there is a lack in my sense which will need to be filed. I think the world could relleay benefit from a P2P system which would allow users to share ideas, knowledge and you name it, in a much more open and accecible way then the net as we know it. I said as we know it cause it's not true that the guy at his computer down in montreal for exemple can easily acces "the mass" without having to really show the money to either "promote" is web site, paying a good connection and etc etc. P2P could solve this as a way for the information to freely ride throu wires and be publishable to everyone. Then, using a standard way of indexing those information, one could easilly browse throu it and find what ones really search for, not 275 000 sites dedicated to sale a product which I don't care and ) sites about what I want because those pages whill probably never get indexed in any major search engine.

Mod me Up, this is recursively on topic!

Re:Nice.

[ffatTony]

NudeCheerleader(part1).mpeg isn't going to be the same as NudeCheerleader(part1).mpeg

This is where a good hashing algorithm would be great, e.g. md5 hashes to determine if different users have the same item. This seems mandatory, they'd be silly not to have something like this in place

Re:Nice.

[King of the World]

Wow. It's like your some kind of programming genius!

Hey - do you think my code would be better if I use those function dohickeys?

Re:Nice.

[Quixote]

NudeCheerleader(part1).mpeg isn't going to be the same as NudeCheerleader(part1).mpeg on someone elses comp.

Then download both! You will have to visually inspect the contents of both to really tell if they are same or not. It's called "research".

Re:Nice.

[Anonymous Coward]

Edonkey (http://edonkey2000.com) does, and always has done, exactly that.. allows you to transfer the files in chunks.. upload chucks that you have before the download is complete etc etc.. it creates a hash (read crc if you don't understand hashing) of the file so it has a unique signature of the file so it can determine if the contents of various copies of the file are the same.

Re:Nice.

[bramcohen]

I think a workaround for this would be to have md5 signatures computed for each of these parts

Not coincidentally, BitTorrent is doing this already :-)

It's actually using sha1 hashes instead of md5, since sha1 is the more generally accepted standard for new protocols

Re:Nice.

[Zurk]

shouldnt you be in university getting a degree rather than messing around on open source contract projects with minimal pay ? get a career and quit living in a basement.

Re:Nice.

[Anonymous Coward]

edonkey2000 creates a hash for each file you're sharing. you CAN do a keyword search for files, but most people are downloading from links found on forums. the link has the hash embedded, so you know when you're clicking on the NudeCheerleader(part1).mpeg link you are going to get the file the poster was referring to. if you have that file in your share directory and you rename it, the hash based search technology donkey uses will still find it and feed it to other people even if the filename of their download is different. thats where users are running into problems -- if you download based on keyword search you could potentially start downloading a hoax file when you thought you had the new LOTR trailer - if you look in your download queue closely you'll notice that some people have the file names "lotr trailer" and others have it names "HOAX" or "nude cheerleader" but regardless the hash (crc) system at least ensures that its the same file, regardless of name. of course, this gets back to the classic problem of trust, and i doubt we'll see a certifying authority for pirated movies/software anytime soon, so its best to trust the links from the forums. either way, at least you're getting the same nude cheerleader. and just like bit torrent you're downloading it from other users who might only have part and who are simultaneously downloading it also.

Re:Nice.

[Anonymous Coward]

WinMX 3 Beta 5 (win32! gasp) does issue a "File Mismatch" error when performing multi point downloads which seems to indicate some kind of checksum / comparison ...

whoopdeedoo

[Anonymous Coward]

The latest version of BearShare does this. Of course gnutella is horribly broken and doesn't scale blah blah blah

edonkey2k

[Anonymous Coward]

edonkey has been doing this for ages..

Re:edonkey2k

[Anonymous Coward]

So has kazaa(and it's clones grokster and morpheus), peergenius, scour, and many other p2p applications. This was new 2 years ago, it shouldn't be news.

Re:edonkey2k

[Supa Mentat]

I'm sure it's cool and everything, but when I am looking for file trading software anything that remotely reminds me of goats is out of the picture, if ya know what I mean.

Comment removed

[account_deleted]

Comment removed based on user account deletion

Am I missing something?

[duren686]

Doesn't Getright do this already?

And with filesharing, Kazaa/Grokster?

Re:Am I missing something?

[dytin]

Getright does do multiple downloads, but they are generally from the same source, which helps speed things up, but not that much. Getright can do multiple sources, but it doesn't find the other sources automtically, you have to find them yourself. (Yeah, it has a built in Lycos FTP search for files, but that never returns any results).

Re:Am I missing something?

[optikSmoke]

Last time I checked, Kazaa does this, but not well. It doesn't seem to be very intelligent about choosing what connections are good (and therefore should be given more of the file to download after they finish a currently-assigned chunk) and what connections are bad (and therefore should be dropped, not let to continue at 0.01K/sec!).

[ On a side note, GetRight allowed for more control over where to download from (and did allow multiple sources, last time I used it -- about a year+ ago). It fact, I used GetRight to download linux ISOs from multiple sources at once :). ]

Anyway, does this system offer *better* multi-connection filesharing (ie, more intelligent?), or does it keep slow connections, and fail to recognize that a fast connection just finished and should be given more of the file to download?

Re:Am I missing something?

[Anonymous Coward]

It only speeds things up until someone notices what you're doing and blocks you from the web server.

I've seen these idiot things download 800-900KB just to get a 300K file. The pipe going out of the web server is only so big. The more you request, the slower it goes. But they try!

Twits.

Re:Am I missing something?

[Anonymous Coward]

Kazaa does that, but not smart enough to figure that there are 5 other files out there with different names same size are actually the same file. :(

Generally you can grab the first XX bytes

[Anonymous Coward]

and compare them to the other files. That would suffice along with file size.

Been there done that...

[mo]

I hate to rain on the parade but Morpheus et al. as well as the latest version of BearShare both do this, and have for some time.
When you say p2p with brains, to me it means somebody has come up with a elegant balance between centralization and search speeds.

Re:Been there done that...

[rtaylor]

Ok.. I knew both of those clients would do parallel downloads, but I didn't realize they were capable of sharing 'chunks' of a file. That is, sharing a file that has not yet completely been downloaded.

Next time I run into something EDonkey doesn't have, I'll have to try out Bearshare.

Re:Been there done that...

[Wintersmute]

When you say p2p with brains, to me it means somebody has come up with a elegant balance between centralization and search speeds.

Ditto, Holmes. The real question is the scalability issue, and I'm not convinced that the traffic cop features implemented by Gnutella front-ends have really sorted this out.

When that's the case, that will be some p2p with brains. Right now, the networks only seem to be hanging on because the critical mass of crash-inducing traffic hasn't hit the super-peers yet; at least not on the permanent basis.

What would really make my evening interesting is if someone would be kind enough to contradict me.

Re:Been there done that...

[bonch]

The next version of eDonkey is supposed to implement a new method of decentralization that requires no servers for search requests. I sure can't wait, since the current eDonkey network is beginning to succumb under the strain of its popularity.

Re:Been there done that...

[Anonymous Coward]

What would really make my evening interesting is if someone would be kind enough to contradict me.

No it wouldn't. There. Don't mention it.

Re:Been there done that...

[Rew190]

Actually you're right and the original poster is wrong. I think the process works by scanning your "My shared folder" directory and basically putting out all of the files you have that aren't .dat files (although i've never done a search on a .dat file, so they might turn up). The .dat files are your uncompleted downloads.

I'm pretty sure the software just searches for names, so unless you force a different name on one of the .dat files then you won't be sharing the file. You can only share after you've got the whole shebang or fake it.

And this is new?

[Edgewize]

Could someone explain how this is different from FastTrack (Kazaa), eDonkey, or the more reputable Swarmcast?

Peer broadcasting is hardly something to write /. about, I'd say.

Re:And this is new?

[mo]

They're marketing BitTorrent as a solution to web providers with bandwidth limitations. The client registers a mime-type so when you click on a BitTorrent download link it hands it to the p2p client which then downloads it from the network.
The technology is nothing spectacular, but it's nice to see a simple install method that integrates nicely into the browser.

One interesting side-effect of this implementation is that there is no searching. You only download stuff from BitTorrent if you find a link on a web page for it. However, without the requirement for searching, Freenet would be a great replacement for this role of browser-download accellerator. All you really need to do to implement this would be to provide a nice installation .exe of freenet that could parse a meta-file pointing to the freenet key of the object you wanted to download.

Re:And this is new?

[PureFiction]

Sure. Fast Track is decentralized file sharing network where there is no guarantee that what you ask for is what you get. They may be codecon mp3's, they may be nasty midget porn incognito.

eDonkey likewise is more of a filesharing (aka, keyword search, then dowload hits) method.

Swarmcast is the closest relative to BitTorrent, but BitTorrent avoids the FEC encoding and cryptographically secure block verification in favor of a more centrally controlled broker that uses multi source downloading at various offsets to accomplish the same task.

In short, BitTorrent is a distribution system where a central server provides content, and peers requesting that content join a mulitsource downloading group where they also share offsets of data with each other (preferably) and download from the central server when necessary.

This isnt file sharing (really), this is content distribution in a fast and effective manner using peer networking concepts.

Re:And this is new?

[Anonymous Coward]

"Peer broadcasting is hardly something to write /. about, I'd say."

It is if you're hawking your wares...

;)

Re:And this is new?

[Edgewize]

OK, so it is similar to eDonkey but without the problematic public servers. And it is basically identical to Swarmcast, but less robust and potentially slower to complete a file. Both of those also support linking, though each in their own way - I believe that Swarmcast is through Java applet parameters, while eDonkey intercepts ed2k:... style links.

Just wanted to know where BitTorrent stood in the grand scheme of things.

Re:And this is new?

[ccweigle]

The client registers a mime-type so when you click on a BitTorrent download link it hands it to the p2p client ...

Again, edonkey already does this. Has for a while. I imagine the other mentioned programs can as well.

What makes BitTorent new is they are actually trying to get it used for a legitimate application instead of just arguing that "people could use it for something other than piracy".

Now, I don't know how long BitTorent has been around, but it appears not to be new (too many "I can download foo real fast with it" comments to think it's just out). This is possibly not their original goal and really just something to point at if a lawyer comes calling (or to try and get VCs to come calling).

Re:And this is new?

[Anonymous Coward]

there is no guarantee that what you ask for is what you get. They may be codecon mp3's, they may be nasty midget porn incognito.

Sounds like a win-win situation to me.

Re:And this is new?

[PureFiction]

Your missing the point. This is content distribution, where those "problematic public servers" are actually usefull. I.e indie bands sharing music (etree). And they server as Quality Control, you know your getting the files from a central source.

Regarding eDonkey, they provide no detailed information on how they implement multisource downloads, so it may be that they are actually very similiar. Who knows.

Lastly, swarmcast is designed for larger decentralized peer networks where no central broker is present. In that scenario FEC and cryptographically secure block transfer make sense, but its a lot of overhead for simply transferring a file. I dont see how your can say BitTorrent is potentially slower. BitTorrent will be faster, as it avoids that overhead entirely, and has a centralized broker/server to fall back on.

Re:And this is new?

[Paradise Pete]

Now, I don't know how long BitTorent has been around

From the author's site:

BitTorrent has been created by me, Bram Cohen as a full-time project over the last eight months.

Re:And this is new?

[Anonymous Coward]

Well, but the fundamental idea behind edonkey and this bittorrent is the same. The implementation is what is different (so, it is not really something new, just a variant).
Anyway, a centralized server seems like a bad idea to me, because we all know that centralized servers tend to be the bottlenecks of performance (so, yes, bittorrent *can* be actually slower. and, BTW, are checksums and integrity checks so slow as to impose an important overhead? I feel that in edonkey this overhead is close to be negligible.)

Anyway, there is something I don't like. While I am downloading, I don't want someone messing with my upload bandwidth - perhaps I am using it for something else, like serving web pages.... Other than the TCP ACK's, I don't want noone messing with my upload bandwith...

Re:And this is new?

[BCoates]

Freenet can be accessed through freenet: URI links (like &lta href="freenet:SSK@npfV5XQijFkF6sXZvuO0o~kG4wEPAgM/ homepage//"&gt) if you have a uri-handling plugin, which to my knowlege have been written for at least IE and Mozilla. Unfortunately, they are not going to be allowed into the default Freenet install for the forseeable future for political reasons (i.e., Ian doesn't like them).

You can still use fproxy to access freenet through a web browser, so I suppose you could use a http link to the fproxy presumably running on your user's localhost, but that's somewhat broken and unlikely to catch on.

--
Benjamin Coates

Re:And this is new?

[Edgewize]

I was referring specifically to eDonkey's problematic servers, which rarely remain up, have a low user limit, and are are completely unsynchronized. A properly done central server is necessary for content distribution - someone has to have the content to start with.

FEC and secure block transfer is slower to compute, and it requires a bit of overhead in block size as well. But it is much faster because packets don't have to arrive in any particular order. It may be speed-equivalent or slightly behind for 95% of the file, but that last 5% under BitTorrent requires that someone else has already downloaded it and is currently broadcasting it. Or else you are back to the mercy of the server's upload speed for the last chunk. With swarmcast, there is a larger pool of available packets to pick the last 5% from, and they don't necessarily have to correspond to the lsat 5% of the file.

Re:And this is new?

[Anonymous Coward]

Red Swoosh (www.redswoosh.com) has been doing this for nearly a year, with many thousands of simultaneous users.

first to complain about the use of mp3s...

[ralian]

Why would these people use a closed format with possible impending royalty fees when they could use an uber-1337 open source format instead?

:-)

Free!

[GigsVT]

You have a great product, many customers, and are delivering your product to hordes of happy
customers online....

The key to cheap file distribution is to tap the unutilized upload capacity of your customers. It's free.

Emphasis theirs.

Free to who? You? Maybe not them. Nothing is free. I like the idea, but I really don't like the way they are selling it.

Re:Free!

[Gumber]

It does, however, seem like a fair trade.

Re:Free!

[GigsVT]

Unless you already paid for the program. They talk in terms of "customers".

yep

[xfs]

and kazaa and morpheus (the old one) have been doing the multiple-user download for quite some time

But I want to be a leech!

[Anonymous Coward]

I mean, really. There's *liability* when you're providing an upload. In this case, you can't grab warez without being a warez site, and you can't grab pr0n without being a pr0n site. And, when you download something with a virus in it, or that isn't what it seems to be, why you've contributed back to the community there, too. Could be a _lot_ harder to get sucker-junk out of circulation.

Th3 Fr33L04d3r

Comment removed

[account_deleted]

Comment removed based on user account deletion

From the oxymorons for dummies department...

[Anonymous Coward]

There's no such thing as a "legal bootleg". A bootleg is, by definition, an illegal product. If they let you make a copy, it isn't a bootleg. If you download the source to a GPL program, would you say you've just bootlegged the program?

eDoneky dudes!

[Baby Duck]

BitTorrent allows users to download a file from multiple different people. Instead of everyone nailing one server, users get the file from other users.

eDonkey does one better. Even if you only have parts of the file downloaded, you can immediately send parts of the file you do have to other users. And eDonkey has had a pretty good track record. I thought everyone and their mother knows about this, so why was this a Slashdot headline, especially when it's pretentious and untruthful?

Re:eDoneky dudes!

[bonch]

If you're having problems finding valid files, try tracker sites like Gowenna/Sharereactor and Filenexus. You can also find dynamic serverlists.

eDonkey's developers are aware of the network's current strain and are implementing a new p2p method for the next version.

Re:eDoneky dudes!

[jth1234567]

Actually, downgrading to an older version will fix the slowness... v57 was the last relatively fast one, and v56 is considered to be the fastest still working client.

Of course, some servers won't let you in with v56, but there are sites with constantly updated serverlists, with those you shouldn't have problems connecting to servers.

As a comparison, v58->v60 give me an average of 20-30 kB/s, v57 about 60 kB/s, and with v56 it's 90-100, _average_.

Re:eDoneky dudes!

[Anonymous Coward]

try eDonkeyBot from http://www.file-finder.com, it helps connect to servers and automatically updates the server list, also you will never get disconnected and its global search works much better than eDonkey's search.

Re:eDoneky dudes!

[Tha_Zanthrax]

eDonkey2k used to rock!!
It was great but for the last 3 months I'm having lot's of trouble connecting to any servers.

I left it running one night once and it connected after a couple of hours and then disconnected after 15 mins.

Re:eDoneky dudes!

[joyrider]

ShareReactor helps me out a lot when it comes to determining the real and fake files on EDonkey - it's worth a try at least...

^^very true^^

[mill5ja]

They use hashes of the file to compair.

-jason m

Re:^^very true^^

[Anonymous Coward]

Very true? You don't seem to fully understand how long an "age" is.

beer 2 beer wif fangs u speaka?

[Anonymous Coward]

never eard of it

so when was it that little silly chips own your weak astas-_____?

It is NOT a p2p Network people!

[Com2Kid]

It is a browser plugin (IE) that creates mini distributed networks based around a website.

So say you start downloading the latest Counterstrike patch from some server. Well you know how servers giving out the CS patch get filled up quickly.

Well if the users were running this program (plugins to IE, no restart neccisary, look if there is a {browser here} version yourself!) then when they started downloading somebody ELSE could start downloading FROM them.

No file synch issues (same file, same source) the server just re-directs future downloaders to current downloads and has the original downloaders forward the files along.

Re:It is NOT a p2p Network people!

[jacoplane]

Perhaps something like this should be built into Mozilla. I think it's a pretty cool feature but it will only really work if browsers support it by default. I don't think many people will actually install this, and of course the usefullness of such a tool is small if there isn't a large user base.

Re:It is NOT a p2p Network people!

[Com2Kid]

Send letter off to moz team, source code for both projects is availble, no reason it cannot be implemented.

If enough people ask and / or do. . . . :)

Re:It is NOT a p2p Network people!

[dago]

There's a new CS patch !

and nobody told me !!!

I'm gonna get it on kazaa ...

eDonkey

[Anonymous Coward]

Now all they need is a better eName.

Wasn't Marc Anderseen's Company...

[AsbestosRush]

Loudcloud working on something like this for a little while... something called "bitcasting"?

Will not help the slashdot effect

[Xenopax]

Ok people, I know we all have this dream of distributed web serving, but as a web developer I feel I must explain why this will not work now:

1) Response Time

To make this work you need more than a fancy P2P network. Remember site like slashdot are database backed and update very quickly. Sure slashdot caches pages, but many things like user preferences and comments are updated way to quickly for a P2P network too distribute it.

2) Security

Yes you can encypt, but who other than a hobbist is going to put the content that represents them on several machine at once and expose themselves to someone breaking it. If someone was successful they could do things like change the slashdot homepage for those they are distributing to. You cannot be a credible source and distribute yourself like that.

3) Slashcode (yeah I know, slashdot specific)

Have any of you actually read slashcode? I'll tell you what, it is damn complicated. There is no way a simple patch is going to make a site like this distributable. The entire thing would need redesigned, which is no small job. I'd say that this would be the case for any database backed site as well.

4) Databases

Since I mentioned a few times already, I think I'll point out the flaw here. Name one database system that is able to handle and organic network of servers (ie constantly going up and down), keep all the data available, keep all the data available on a resonable connection (not behind 56k lines), give the response time you need, doesn't take up huge amounts of systems resources, and can easily be set up on one of the P2P nodes by even a reasonably competent user. Oh that's right none, and you have to have that in order to have a dynamic site on a P2P network, which is a huge portion of the web at this point.

Well, that's all I can think off right now on this, but I'm sure there are plenty of other reasons why this isn't feasible in the near future.

Cheers

Re:Will not help the slashdot effect

[Eloquence]

You make some valid points. Distributing dynamic web content is immensely difficult. There are some projects that attempt to do it, and usually, synchronization servers come into the mix at some point (data loss can be avoided by tracking server availability and assigning new servers as needed, you quickly get mathematical safety). The problems are not unsolvable, but I don't see any large corporation pushing the idea, so it will probably still take a while. Realize that this is about high load on static large files, though, where P2P makes very immediate sense.

Re:Will not help the slashdot effect

[Com2Kid]

Uh.. . .

The idea was to help sites that GET linked to BY Slashdot. /. itself is not, AFAIK, having bandwidth problems.

You know, those small user pages with some cool casemod on it?

This network would allow viewers of the site to download the images from EACH OTHER instead of from the main server.

Re:Will not help the slashdot effect

[dfelznic]

how many times has a slashdot story linked to a pdf on a server that got ./ed? How many dynamic pdfs do you see floating around. I got bad news for you, most of the web is not dynamic...

Re:Will not help the slashdot effect

[ryanvm]

The idea was to help sites that GET linked to BY Slashdot. [...] This network would allow viewers of the site to download the images from EACH OTHER instead of from the main server.

I may be incorrect, but couldn't this scenario be remedied by FreeNet? Doesn't FreeNet distribute and cache popular content all over the place?

If the Slashdotters won't support FreeNet, nobody will.

Re:Will not help the slashdot effect

[Com2Kid]

This is a drop in replacement for HTTP though, browsers can easily be made to support it with minimal work.

Re:Will not help the slashdot effect

[CaseyB]

Remember site like slashdot are database backed and update very quickly. Sure slashdot caches pages, but many things like user preferences and comments are updated way to quickly for a P2P network too distribute it.

Do you even know what the slashdot effect is?

This has nothing to do with changing slashdot itself, it's about the possibility of using software to help distribute the load that slashdot dumps on third party web sites when their home page becomes the subject of a hot story. Slashdot readers could become temporary mirrors for the links.

Re:Will not help the slashdot effect

[Anonymous Coward]

"Web developer" heh, yeah I am sure web developers know about scaling network topologies. Go back to hacking out your asp/perl/php and leave the real coding to the big boys!

Re:Will not help the slashdot effect

[MrEfficient]

I know it has problems, but I still think this is a great idea and is feasible if done right. I've been thinking of something like this ever since slashdot introduced subscriptions.

I'd really like to see some research on something like this, but unfortunately I don't have the programming skills to even attempt it. I think you would not only need a specialized client for this purpose, but you would need to design your content around it. There would probably be some things you would have to do differently than you'd do them on the web. This idea seems to share a few traits with Freenet, although you would'nt need to worry about anonymity of sender and receiver. You'd only need to worry about protecting the content as you mentioned.

P2P is gaining a lot of momentum right now, and what's important is that its in areas other than mp3 trading. There was an article in Scientific American about distributed computing a month or two ago. I think it's on their site in you want to read it. I truely think that we will see some websites built like this in the future, but I'm not sure they will be technically "web" sites at that point.

Re:Will not help the slashdot effect

[Xenopax]

I realize that we are talking about slashdotted sites, but I've heard so many things already about "distributed slashdot" that I felt the need to debunk that. Besides, many of the sites slash dot links to wouldn't be able to be distributed anyways. Only a static site could be served in this manner, and even then how do you get the browser to pull content from those nodes? And even if we could hack our way through it, we'd just be adding another level of complexity to a system (the web) that is broken to begin with.

Re:Will not help the slashdot effect

[Com2Kid]

Most sites that have grown large enough to NEED a dynamic page delivery system (and if you don't need it, FOR CRYING OUT LOUD DO NOT USE IT) are able to easily (or at least without too much trouble) withstand the /. effect.

And the web is not broken, it is just littered with porn and blogs. . . . .

(used to be porn and geocities pages with out of focus pictures of kittens on them. Now it is blogs with out of focus pictures of kittens on them. My how we have progressed. . . . )

BitTorrent

[Eloquence]

I'm sure Bram will notice his server being slashdotted soon enough, but let me say a few words in defense of BT anyway. What makes BT different from Morpheus and BearShare is that files are sent by users to each other, while they are still downloading. This way, the downloaders themselves act as backup. It's not simple multi-source downloading, but targeted towards content-providers who want to reduce load on a central server. In its advantages and disadvantages it's similar to Multicast. Good for high load for specific files at specific times. Kernel.org should use it.

eDonkey has the same feature (with some differences in the publishing process), but is really an application of its own, very file sharing oriented, closed-source and banner-supported. Not exactly what a content provider would want users to download before they can access his files. Still, ed2k has the advantage of a large user base, and also supports ed2k:// URIs that can be used on webpages.

SwarmCast is interesting, but the company behind it mostly died, and now it is somewhat in limbo. Its Java base has made it problematic as a desktop application. The only real alternative to BT is Mojo Nation, which is currently being reworked as "MNet".

If you want to know what CodeCon is all about, check the Feature box on infoAnarchy, we had some detailed coverage.

Re:BitTorrent

[mark_lybarger]

Kernel.org should use it.

all the reports i see during a /. effect on kernel.org show the site is handling the traffic very well. i don't believe i've ever seen the FTP servers at 100%, 90's maybe, but not 100. IIRC, they have some _serious_ bandwidth on those those servers. if there were an issue, i think the kernel.org folks would probably ask the /. folks to _always_ post a link to the mirror page instead of to the tar.gz file itself. maybe kernel.org actually welcomes the /. effect. you know something like:

sysadmin: "don't tell me, we've got a new stable kernel release being posted in an hour?"

Marcelo: "yep"

sysadmin: "/. effect coming 3 seconds after its posted?"

Marcelo: "most likely"

sysadmin: "bring it on!"

Re:BitTorrent

[Eloquence]

Quite possible. But then I think about all the poor starving developers who could put food on their families if kernel.org gave their bandwidth budget to them and used BitTorrent instead ..

Re:BitTorrent

[mark_lybarger]

i see the big "operated by transmeta corporation" right on their front page. this doesn't exactly scream that their paying the bandwith bill, but then again it lends itself to being construed as such.

as a wild guess, say their fat pipe is costing 5k per month. how many starving kernel developers is that going help out? are there really people who are large contributers to the kernel that also have troubles finding a day job?

Re:BitTorrent

[Eloquence]

as a wild guess, say their fat pipe is costing 5k per month. how many starving kernel developers is that going help out?

Heh. You don't really want to know the answer to that question :-)

Re:BitTorrent

[mark_lybarger]

i do wanna know (maybe my estimate on the bandwidth was way off who knows). but as i see it if someone is a major kernel developer and they have a family and have troubles putting food on the table (i'm assuming they're not living at cowboyNeal's place), i gotta think they got some prioritizing to do.

Re:BitTorrent

[Phork]

Last i heard, the bandwidth for kernel.org was being provided by ISC, and they figured the donation was worth about $8500 a month. thta was from a post on LKM

Re:BitTorrent

[Anonymous Coward]

5k? Hell consumptionjunction.com says its bandwith bill is 45k a month.

What about Gnunet?

[Anonymous Coward]

I'm very surprised at the little ammount of attention that GNUnet has gotten in the P2P arena. GNUnet is anonymous, distributed, encrypted, reputation based, has accounting, allows for distributed queries, and uses dynamic routing. While GNUnet is still beta software, I think it's a great anti-censorship tool. What all this means in non-buzzword speak, is that you have a tool that combines a lot of the great qualities from other similar networks (FreeNet, mojo nation, etc) and doesn't have all of the short comings. Give it a shot.

Re:What about Gnunet?

[Jagunco]

Technically Gnunet is the most complete p2p kid on the block (even with the 1k filesize block thing). It seems its main problem is the lack of a pretty user interface, specially for windows. (it may hurt to say, but no p2p applications out there are likely to succeed without nice windows interfaces). Maybe a port to .net (mono?) could help it :)

Re:What about Gnunet?

[Anonymous Coward]

I agree. From reading the technical documentation on GNUnet I think it's way ahead of anything else out there in the p2p arena. A win32 port is on the way. A primative gtk gui has been written and is sure to improve in the upcoming months. I do agree that these are some of the 'downfalls' of GNUnet though. I find it amusing that in order to succeed, an application needs a win32 port. I agree, but I think it's discouraging none the less.

Kazaa/Morpheous/eDonkey compairisons

[goofy183]

There seems to be a lot of people who really haven't read the site or understand how the technology works. Yes all those P2P filesharing utilities allow you to download the same file from multiple people at once, it's not all that impressive and many of the problems such as validating matching files and such have been worked out.

This solution is different in a few very large aspects. It allows a company to keep track of who is currently downloading a file from their webserver. This information is then sent to the clients who can start the P2P poriton of the process and download segments of the file from other users, releaving the load on the companies server. In contrast to those other P2P FILE SHARING programs which share all your files not just ones you are currently downloading. A system like this makes the file server not only the original source for that file but the P2P server to find other people to download that ONE file from.

I can see where people may not want their upload bandwidth being used by others. For this reason any site implementing this feature would probably end up having to provide the file for normal download. The selling point would be a possibly faster download for users of the technology.

I would personally love to see huge sites like FilePlanet put this to use. Granted it would only be truely usefull for sites that have a constant stream of concurrent downloads for a file at any point in time but it would be much better than having to wait 2 hours in line to download a file :-P

Re:Kazaa/Morpheous/eDonkey compairisons

[Paradise Pete]

any site implementing this feature would probably end up having to provide the file for normal download.

Well yeah. Otherwise how would it get started?

Re:Kazaa/Morpheous/eDonkey compairisons

[Glass of Water]

Yes Yes Yes Exactly!

Now, where would one find a large file (say an ISO) that lots of people wish to simultaneously download every once in a while?? THAT's what this is perfect for.

Comment removed

[account_deleted]

Comment removed based on user account deletion

Unreasonable requirement, unwanted feature

[tftp]

Several people commented that this thing allows to redistribute files before you finish downloading them. But this is not a big deal simply because most of the time file is not being downloaded, it just sits on the HDD 99.9999% of its life. The gain from the early upload would be next to nothing.

Re:Unreasonable requirement, unwanted feature

[Account 10]

it just sits on the HDD 99.9999% of its life

However its popularity decreases rapidly throughout its life.

The demand for a file may be incredibly high during the first hour of its life, stay high for a day and then start decreasing rapidly as it becomes 'old news' and more widely available.

For example, a trailer for the new Star Wars film may take an hour to download and be in huge demand. The next day, the demand is less concentrated since it has been on television, all the hardcore fans downloaded it the second it was available, etc.
Being able to start the upload before it has downloaded takes enormous pressure off of the sites that have the complete file.

It may be an unreasonable requirement, but I'm sure it is not an unwanted feature.

Red Swoosh

[metalhed77]

Red Swoosh is a cool technology specifically aimed at distributign the load for things such as images on a website. The client download for IE just involves clicking install and DLing a client that's a few 100kb. After which you mirror a portion of the site. www.deviantart.com uses this, and to good effect. I'm not sure if you can mirror large files on it. It is of course centralized.

Gnutella clients already do this

[alanjstr]

The latest BearShare and LimeWire both allow you to "swarm" gnutella downloads.

Re:Gnutella clients already do this

[anandsr]

Swarming a gnutella download is different than
what BT does because, BT is site initiated.
BT is used by the site putting up a download file.
So that anybody with the appropriate plugin getting
the file from the site, automatically gets the file
from other site. Its not like searching for a file,
and then you are given a site that enables the same
Torrent effect.

ToDo list: handle more than a dozen downloads

[specialized_sworks]

From the BitTorret to do list....

better scaling/performance
BitTorrent currently scales well to a dozen or so simultaneous downloads. With further modifications, it can be made to scale to thousands.


A dozen or so simulataneous downloads? Dont think that is going to help prevent the slashdor effect. Though I guess that is getting tested right now!

Multiple different people

[Anonymous Coward]

Thank goodness it's not multiple same people or single different people!

multiple identical different same distinct

[Dr. Awktagon]

BitTorrent allows users to download a file from multiple different people.

Or if you're downloading the latest boy band single: multiple identical people.

Re:multiple identical different same distinct

[metacell]

Or if you're downloading both Britney Spears and Black Sabbath: Multiple personalites.

Not good for asymmetrical connections?

[Malc]

"As soon as I started downloading (cable modem) at around 300k I got a request for the file and began uploading at 40k."

I have a 1184/160kbs asymmetric (DSL) connection. This seems like a common ratio with many ISPs these days. A full speed download consumes at least a fith of my upstream bandwidth. Presumably that's due to things like TCP ACKs. Any kind of serious upstream activity squeezes things and can quickly reduce a download to half speed. I can't find the concept described very useful, especially if I'm in a rush to get something. Is there a way to throttle upstream bandwidth consumption?

Re:Not good for asymmetrical connections?

[Anonymous Coward]

No.

See, after so many packets are sent over a TCP/IP with a valid checksum, the MSS grows -- which means less TCP ACKs.

So, it automatically happens that way.

Re:Not good for asymmetrical connections?

[Anonymous Coward]

The MSS is never going to be larger than your MTU, which is probably 1500 bytes. The window grows, but that doesn't matter since every other packet still needs to be acked.

Re:Not good for asymmetrical connections?

[Malc]

No, you're wrong. I've sat there when I've been downloading big files such as ISO images. The TX throughput on the WAN port of my router does not decrease with time. Anyway, isn't the TCP receive window size of more relevance than MSS?

Re:Not good for asymmetrical connections?

[Jeffv323]

Yeah, I have the same problem with my cable. I am happily downloading at 200KB/sec in some P2P program (used to be closer to 600KB/sec until @home went away) and then someone starts downloading from me and maxes out my upload. All of a sudden my download speed drops to like 40KB/sec, response times on web sites are painfully slow, and I end up cancelling the upload... Anybody know of a way to fix this?

Re:Not good for asymmetrical connections?

[anonymous cupboard]

I have my packet sizes maxed out for ADSL and this seems to help. Also, I not that uploads don't take a lot away because they can share physical packets with the d/l ACKs.

Re:Not good for asymmetrical connections?

[Jeffv323]

Well shit I fixed it! A simple firmware upgrade to my lniksys router was all it took. I read some forums before that and I guess it has something to do with the MTU of the broadband connection. Mine changed from 1472 (or something) to 1500 and I got an amazing difference! I just wish I had done that back when I had more bandwidth...

Re:Not good for asymmetrical connections?

[Anonymous Coward]

ACKs can be piggyback'd within a TCP connection, not across several connection (not even to the same hosts, etc).

Re:Not good for asymmetrical connections?

[Anonymous Coward]

Furthur allows users to limit how much upstream the program uses.

Check it out: http://furthurNET.com

edonkey?

[2MuchC0ffeeMan]

nothing new, edonkey2000 has been doing this for months now.

linky linky

Nice BUT....

[Dr. Weird]

Gnutella/Limewire already does this... (I said nice butt... hehehehe) ~Dr. Weird

Re:Nice BUT....

[Anonymous Coward]

The gnutella spec specifies the use of SHA, *NOT* CRC32 or MD5, as some others have recommended. Both of the latter two can be exploited to pass garbage by a check (with CRC32, you have some control over the content, even).

MD5 is *not* suitable for ensuring that two files are identical when a malicious user is involved. It *is* suitable for ensuring that a malicious user may not hand you anything that passes but pure garbage (given what we know about MD5 today).

CRC32 is totally unsuitable for any environments that could involve malicious users.

SHA is the only common hash appropriate for this sort of problem.

Re:Nice BUT....

[Anonymous Coward]

I find this odd that the post is marked up +2 when the author makes baseless attacks on md5. No where does the author explain how it can be exploited, and no were doesn he cite any reference or said exploits. I guess if you make wild claims about a much used algorithm the masses will just assume you know what you are talking about.

Re:Nice BUT....

[Anonymous Coward]

...and this is based on??? Why is "MD5 is *not* suitable for ensuring that two files are identical when a malicious user is involved"? Please back-up your statements with fact or at least a few URL's to make you sound intelligent.

Re:Nice BUT....

[treat]

MD5 is *not* suitable for ensuring that two files are identical when a malicious user is involved. It *is* suitable for ensuring that a malicious user may not hand you anything that passes but pure garbage (given what we know about MD5 today).

I challenge you to find me any two sets of data with the same md5.

Re:Nice BUT....

[Dahan]

I find this odd that the post is marked up +2 when the author makes baseless attacks on md5.

*shrugs*... the truth is truth whether or not you provide references. But if you want references, check out RSA's own FAQ. MD4 is definitely broken, and MD5 might have some significant weaknesses. It's likely to be brute forceable with reasonable resources.

Last form of fail-safe antipiracy crumbles

[Dunhausen]

The weighty download time of acquiring 2 CD's worth of executable has literally been the only deterrent keeping a lot of Morpheus users from pirating the entire contents of their harddrive.

This represents a key-step in issuing in the new era of "freeware."

It's not a IE plugin

[icepick]

it's a helper app. You can build it for any browser that knows how to open a helper app for certain files.

Toronto area radio station doing this (sorta)

[xenyz]

Mojo Radio, a Toronto area radio station ('talk radio for guys') uses something similar to do streaming audio. They use technology from ChainCast Networks to distribute the streaming of Windows Media broadcasts. It installs a little app in your Windows machine and runs whenever you listen to the stream.

My experience with...

[Anonymous Coward]

"My experience with" is a code used by trolls to signal to each other that a post is a troll. Surely you remember the series of "My experience with linux" trolls.

DNA Lounge

[apirkle]

Interesting tidbit, CodeCon was hosted at the DNA Lounge, the nightclub in San Francisco that was recently bought, remodeled, and reopened by JWZ (also known as Jamie Zawinski), the "retired" programmer of Netscape fame.

Pics from the con are also available on the club's website.

Re:My experience with BitTorrent

[icepick]

As near as I can tell, they arn't using BitTorrent, which is a shame because it's perfect for just this.

They'll need to release Netscape plug-ins

[yerricde]

A dozen or so simulataneous downloads? Dont think that is going to help prevent the slashdor effect.

Slashdot linked sites that use BitTorrent technology will respond an order of magnitude faster, and that's never a bad thing. However, it might not exactly ease slashdottings until they release a Netscape plug-in for all major operating systems, as a larger than average proportion of Slashdot readers use Mozilla and non-Windows desktop environments.

Re:They'll need to release Netscape plug-ins

[bramcohen]

BitTorrent actually should be working under Netscape currently - I seem to have it setting all the appropriate things in the registry, but I must be missing something, because it doesn't seem to work.

If anyone can help troubleshoot this problem I would much appreciate it! Setting mimetypes in the registry seems to be undocumented mystic voodoo - I got it working under IE by configuring manually and then searching through the registry to see what it had done...

I'd like to hear peoples's experiences

[Anonymous Coward]

I'm Bram Cohen, the author of BitTorrent. This little slashdotting seems to be going well so far from my end (over 40 downloads currently, and still going smoothly) but I'd like to hear about peoples's experiences doing the download. Here are some questions -

• Are you getting pauses where no download is happening? If you are, please be patient, it should kick up again (or start in the first place) after a while.
  
• Are you behind NAT? People behind NAT may be getting worse performance, it's a complicated issue.
  
• How's your upload/download ratio? There are enough people now that you may see the phenomenon of getting about the same download rate as your upload rate - Cutting off your uploads wouldn't help with this, your peers would just get pissed off at you and stop uploading (I'm not kidding, it has tit-for-tat leech resistance.)
  
• Did you run into any technical glitches? It's still fairly young software, so there may be a few little things to iron out.
  

So far, this looks like it's going pretty well. Any and all feedback is much appreciated, and will hopefully help make BitTorrent an even better product. Please mail me about your experiences.

Re:I'd like to hear peoples's experiences

[XRayX]

1. Sometimes I get massive speed-drops (around 5 k/s), but no freezes until now.
2. I'm behind a NAT. May it be possible to configure an incoming-connections port?
3. Very Various. At The Moment it's 30 K/s down (max 90 k/s) and 7 k/s up (max 14 k/s).
4. No problems! Plugged in pefectly into IE.
Very good work so far. I'll try to set up set up some files later.
X

The questions are...

[Anonymous Coward]

Does the software report the size already downloaded or the full size, and does it handle buffer underruns, like when you uploaded the song to the point where the upload is trying to get more than you have yet to download.

eDonkey hashes files

[bonch]

You download file chunks from multiple people, and files can even have a completely different filename. All files are given a hash value to compare to.

Speaking of good things about eDonkey, there is also forced uploads, meaning no losers cutting your downloads on you.

bootlegs : recordings :: crackers : hackers

[Wakko Warner]

Please, call them "legal live concert recordings", not "bootlegs". That's like saying "legal pirated MP3s".

- A.P.

Re:bootlegs : recordings :: crackers : hackers

[mark_lybarger]

just to clairify.

bootleg:

1. To make, sell, or transport (alcoholic liquor) for sale illegally.

2. To produce, distribute, or sell without permission or illegally: a clandestine outfit that bootlegs compact discs and tapes.

it was very hard to find someone using the term bootleg to not mean anything more than a live recording though. lots of people call even a bands released live album a bootleg. here's another definition i found, sorry no link, i could only get it on google cache:

"When someone tapes a show, that is called a live recording. When a company releases an unauthorized copy of that show, that is called a "bootleg". Bootlegs are usually found in compact disc form. However, a CD can only hold approximately 78 m inutes of recording time, forcing the bootlegger to cut songs out of long shows. In essence, a live recording will maintain the original, unadulterated full show while a bootleg version will have songs missing. In fact, they may even be out of order.

When individuals trade live recordings, no money is transferred or involved. However, when someone buys a bootleg, someone is making money--and usually a lot of money--off music that someone else wrote and performed."

Still needs Work

[Mick D.]

Being that I am tired and needing sleep I am not going to kill myself trying to get this to work, but just FYI, it imediately crashe IE on my win2K machine.

I'm not saying this thing is busted just that it certainly seems that this guys request for more money to work on it is obviously nessesary. Oh, well, I was hoping I could get one program to work tonight even it it wasn't one I have been slaving over for the past week. I guess my computer karma is kinda low right now.

brains?

[mark_lybarger]

and no brawn? back to the drawing board folks.

BitTorrent meets Mandrake 8.2, please?

[ediron2]

This is seriously cool stuff. Two minutes for the exe plugin (ick, IE... grab *nix version next), a restart, and I'm seeing 90%-plus of my pathetic CableOne allotment coming in from the demo mp3.

Meanwhile, I'm in day two of pulling ISO files from Mandrake 8.2. I'm at 7% on the first image after an hour. Admittedly, it's not literally 24+ hours. I'm just in my second day of intermittent efforts to get an ftp connection, being told mirrors are overloaded, being told that the 8.2 cd's aren't for sale (talk about poor marketing!!! WTF do these companies THINK we're going to do if we can't buy it before we can download it for free!).

So, anyone want to get into the gritty details and make a 2nd BitTorrent demo page: three Mandrake iso files come immediately to mind. I'd delve in but, well, I'm sort of missing the iso files.

... and I'm still at 7% downloaded on the one ISO I've managed to start a download on. Seriously, Mandrake, at any modest price (say, under $30) I'd have paid for a guaranteed download or a quick-shipped CD. Heck, talk to cdbaby.com or someone else and let them do fulfillment. I won't donate to a publicly traded company, but I'll buy the hell out of convenience at a modest price.

For now, I'm off to see if cheapbytes or anyone else has the cd's. Ooh... time spent editing has me at 8%.

Re:BitTorrent meets Mandrake 8.2, please?

[khuber]

Why do you have to have Mandrake 8.2 now? In a month the servers won't be overloaded, and they'll have released bug fixes.

Obviously the above is rhetorical, but really there aren't any files that I can't live a perfectly happy existence without for a while or forever. </soapbox&gt

Forward error correction distributed file transfer has been around a while now, yet it hasn't found widespread adoption. Sure if Mandrake ISOs are on it that's great, but what about the many files that aren't? FEC needs to become a core part of HTTP or some other basic net service.

-Kevin

Scales Best for small number of large files.

[billstewart]

While it can help out those slashdotted sites with cool casemods, one of the cool things about BitTorrent is that it doesn't just work on whole files at a time, like Freenet and the Napster followons - it's made to handle individual chunks of files (e.g. separate megabytes of CD-sized files), so once you've downloaded a chunk that other people want, you can start uploading that to other people while you're downloading the next chunk. This means that you're able to do useful work before you've finished downloading the whole file, and instead of everybody downloading Meg 1 from the server, then Meg 2, different people get different chunks to download, and can share them with each other.

Also, because you're typically downloading a few tens to hundreds of megabyte chunks, you're a useful server for 90-99% of the time you're downloading, rather than the Freenet model where you're only useful *after* you've finished downloading the stuff you want. So instead of a long-term persistent set of users who always want stuff, BitTorrent is designed for temporary communities of people who want stuff Right Now, and it doesn't depend on them hanging around being useful after they've got what they want. (So you can download the latest release of a Debian ISO and then go install it without feeling like you're depriving the community by taking your machine offline.)

BitTorrent might be able to manage larger numbers of smaller files, e.g. a Slashdot event, but I haven't looked lately, and it's more interesting for the bigger things. (Of course, some slashdotting problems aren't file retrievals, but server interactions, like that one-IC web server powered by a potato battery, and it doesn't have anything to offer for that :-)

Re:Scales Best for small number of large files.

[Com2Kid]

The main site says that it is designed as a drop in replacement for HTTP and even has the potential to match HTTPS services if somebody ever gets around to paying him $$$ to develop that feature. :)

I do not know how well it scales (hell I have gotten 10 karma points today just for spending 2 f*cking minutes at the site, if even that long. The site has all of 5 or so small pages to it. . . . LOL! ), or how the connections are handled.

It does NOT appear help with keeping users anon though, naturally not I would think that it shares IPs like mad, LOL!

"Dumb" file resuming methods (like the one Direct Connect) uses does not need for the person serving the file to have the entire thing, often times I will download the first 300 megs or so of a file from one user and then wait a few weeks until somebody comes along who has the complete file.

By coincidence I have about 15GB of 1/2 done *COUGH* files *COUGH* on my HD. . . . (no, not MP3s, bleh! :P :)

But it does work a bit better then those stupid systems that keep record of where files are to leave off and what not.

Of course one big disadvantage of not having any sort of file integrity checking is that when a SINGLE mistake is made in a file transfer and the rest of the transfer is completed, that mistake can spread throughout the network expontentialy.

There is actualy a corrupted copy of the Yu Yu Hakusho movie that was going around p2p programs for 2 years, and very well MAY BE STILL going around on the networks, that I started passing out on VNN2000 (any other VNN2000 fans out there? I know there are, I first got linked to it from a persons /. sig, LOL! Hia, this is Shiloh, need help with your VNN2k client? :) ) after a tranfer of that file from a different user got corrupted.

On VNN2000 I was one of the few (at times only) user with that file, so. . . . by the time I realized my mistake I:

A: Decided screw it I may get around to fixing it some day

B: Technicaly it is just a few bytes so anybody who REALLY wanted to get the data could boot up a hex editor and fix things. . . . if they had insansly large amounts of hex editing skills. :) )

The error was actualy caused by a 'known but WTF is causing it????" bug in VNN2000 that resulted in minor file corruptions at times when resumed files were appended to each other. (I forget the exact details, the bug popped up a number of times, once I do believe it added some of the network code to the end of the file, oops. No, 'close connection'[1] isn't a valid part of an MPEG4 encoded AVI frame, just ask WMP. :) )

{caveman ugh footnotes}
[1] Me no bothering to look up WTF this would have really been, you worry, you get bin code, {/end caveman ugh}

Comment removed

[account_deleted]

Comment removed based on user account deletion

You damn liar!

[vena]

There seems to be a lot of people who really haven't read the site or understand how the technology works.

that NEVER happens here! take it back!

Actually reasonable requirement, useful feature

[billstewart]

Some environments, like the Gnutella/Napster/Freenet things, have communities that hang around connected for a long time even if they're not downloading anything. But others, like distributing a new release of RedHat/Debian/Mandrake CDs, or even just Mozilla, have a lot of users who want to show up, download stuff, and leave. This feature makes it possible for them to be a temporary community providing services to each other without requiring longterm committment. If you download a CD using BitTorrent, you're useful for 95-99% of the time you're on line, rather than being consumer-only for the first 100% of the download time and having to hang around for another 100% of the time to be any use to anybody, so the community scales much more cleanly even if the first thing you want to do after downloading the latest Linux release is install it. (Software's a much different usage pattern than music here.)

Additionally, it makes it very efficient for the first set of people who are downloading the file. Instead of having to download the whole thing from one source, which is probably overloaded, you're able to download pieces from lots of different people. The server takes advantage of this - instead of giving Alice chunks 1, 2, 3, ..N in order, and giving Bob the same things, it spreads around the load, so Alice is downloading chunk 1 while Bob downloads chunk 2, and when they're done, Alice starts downloading Chunk 3 from the server and Chunk 2 from Bob, and other chunks from Dave, Eve, and Freddie if they've gotten them.

This also reduces the latency required for later people in the process to get their material - instead of waiting for the entire 600MB CD to be copied N times in a row, the downloading gets pipelined.

"chaining" is DIFFERENT than "swarming"

[rubybroom]

I'm not completely versed in morpheus/kazaa/bearshare/whatever, but I understand they allow you to download a file from more than one other person simultaneously, known as "swarming" the download (btw, this is called "anteloping" on furthurnet). It is my further understanding that you can only do this from people who have the *complete* file.

What bitTorrent (I think) and furthurnet (I know) are doing is different than this. If 5 people are downloading a file from the one person who is sharing it, those 5 people can be the beginning of 5 chains of people, relaying each packet down the chain as they get it, regardless of whether or not anyone has the complete file.

Furthurnet uses a protocol called PCP (Packet Chain Protocol) to do this, and it automatically arranges the chains so that those with faster upload speeds are toward the top, with the dialup users toward the bottom.

If the main host goes offline, even if no one on the chain has the entire file, everyone on the chain can still continue downloading everything that the topmost person on the chain has already saved.

A good example: say a dialup user has large file that is in high demand. A T1 user comes along and spends a long time downloading it off of the dialup users horrible upload speed, and gets about 80% of it before anyone else comes to download. Then you show up with your cable connection and instead of being at the mercy of the upload speed of the dialup guy, you have access to 80% of the file from the plentiful upload speed from the T1 guy. And of course Furthur knows to hook you up to the fastest open slot available when you come along.

The result of this is that the underlying host and network shape becomes transparent, and you just see a list of shows to download, you start downloading one, and all this stuff happens in the background. The longer everyone stays connected to the network, the more efficient it comes because it has more time to structure it with the faster folks in the "middle", and the slower ones on the "outside".

Over at furthurnet, the current record is having 71 people on a downloading chain. Combine PCP with the Anteloping and you can have some serious improvement over "dumb" p2p.

I wont even go into the benefits of the md5 checking furthur does...

Re:"chaining" is DIFFERENT than "swarming"

[Anonymous Coward]

Combine PCP with the Anteloping and you can have some serious improvement over "dumb" p2p.

I would think giving PCP to Antelope would result in some SERIOUS horniness issues...

Maybe it's just me...

Re:"chaining" is DIFFERENT than "swarming"

[bramcohen]

BitTorrent does in fact behave generally the way you describe although it's a bit more sophisticated than 'anteloping'. The way furthurnet works, distribution always follows a tree, and the leaf nodes don't do any uploading. BitTorrent is much more mesh-like, with pieces pretty much flowing every which way across the network.

BitTorrent also makes extensive use of checksums, in what I'm guessing is the same way furthurnet does.

It's actually not too surprising that BitTorrent and furthurnet have a lot of similar features - they were both designed with etree in mind as a primary customer.

Re:"chaining" is DIFFERENT than "swarming"

[Anonymous Coward]

Ah, but only one was designed *by* the guys from etree wioth input from the lossless audio trading community.

Re:"chaining" is DIFFERENT than "swarming"

[Hast]

nown as "swarming" the download [...] It is my further understanding that you can only do this from people who have the *complete* file.

Not really. There is nothing stopping you from sharing the file befire you have all of it. It might be that it's not implemented in the different P2P programs, but that's just a coding issue. (Most likely to avoid problems with different files and such. You want to be sure that the file you have is correct before you begin sharing.)

For a WWW site this is not adequate however. As has been commented before the basis for Napster/Kazaa/etal is that the users are logged in a long time. Then it makes sense to wait until the download is completed before you start uploading. On a website you generally get the file and then go browse somewhere else. In other words, you don't gain anything.

Programs like Swarmcast use more advanced algorithms to distribute the files. They chop it up and then make redundant packages, so you increase the odds of getting valid ones. This also mean that you don't need to bother with administrating chains and other similar things. If a user is slow you will just not get all that many packets from him/her.

I'd recommend people that are interested in these more 'intelligent' P2P features to read the papers the Swarmcast guys have made. (www.opencola.com I think.) Lot's of interesting info there.

Hopefully we'll soon see these more advanced downloading schemes being used in normal P2P programs as well.

Lossless - big files - Bittorrent helps a lot

[billstewart]

If you're shipping around small files, like MP3s, there are lots of transfer systems that can do the job. But the Lossless Compression movement for music means that a concert tape is typically a few hundred megabytes large, maybe 1/3 the size of the uncompressed original, so it takes much longer to download, just as ISOs for Linux distributions are large. In that environment, you can't always depend on connections being up for a long enough time, so you need to be able to download parts of files, and swarming systems like BitTorrent help a lot.

Re:Lossless - big files - Bittorrent helps a lot

[mbourgon]

>1/3 the size of the uncompressed original

so, about 320kbps. Sounds about right for the postings I've seen lately.

how is this informative?

[Anonymous Coward]

Talk about biting a troll's bait! The guy wrote "yourmom" in his post for Christ's sake!

Let's see Mandrake 8.2 ISOs...

[Anonymous Coward]

made available. I'd do it, but I don't have them :(
I think it would be best way to demo this technology!

aarrgh!

[Anonymous Coward]

It's written in python!

What the hell?

I'm flabbergasted!

And disgusted!

And impressed!

All at the same time!

I want to port it to Perl or Ruby!

Comment removed

[account_deleted]

Comment removed based on user account deletion

beware

[Anonymous Coward]

BitTorrent is...suspect. Even worse, its creator is a bit much. Please see, on the same site, http://bitconjurer.org/a_torturers_account.html and http://bitconjurer.org/it_happened.html. This Bram Cohen is One Sick Fuck.

Re:beware

[cerenyx]

No shit, man. Whats up with this?

Re:beware

[bramcohen]

fiction -

1. a literary work based on the imagination and not necessarily on fact
2. a deliberately false or improbable account [syn: fabrication, fable]

Re:beware

[Anonymous Coward]

I dunno, you gotta respect someone who will put the word "devil" in their resume.

Re:beware

[Anonymous Coward]

Oh BOO HOO. Are you really such a wussy slashdotter that a few long words on a swirly background make you paranoid and superstitious?

Keeeeeyrighst. Grow a brain.

Re:beware

[Anonymous Coward]

Yup.

Actually I thought 'Purple' was kind of interesting. But the two linked above make it look just a teensy bit like you need to work on your attitude towards women. Unless of course you just like creating extremely unlovable characters.

Particularly this passage;

"You fucking bitch!" I yelled. "You think your pretty fucking face makes you not have to do shit! Fuck you! You haven't done shit! You like your fucking face, huh? You fucking fat whore! You're a fucking whore!" [followed by murder scene].

Sure it's a free planet, write what you want, but as a piece of free career advice, it's probably best not to site the words ' I'm the best, so I can fuck up any captured bitch I want. I pick 'em out of a list and do whatever I want to 'em.' two clicks away from your CV.

Re:beware

[matrix29]

BitTorrent is...suspect. Even worse, its creator is a bit much. Please see, on the same site, http://bitconjurer.org/a_torturers_account.html [bitconjurer.org] and http://bitconjurer.org/it_happened.html [bitconjurer.org]. This Bram Cohen is One Sick Fuck [bitconjurer.org].

YIKES!
I hate to agree on this one, but even though BitTorrent might be the next big wave to reinvent the concept of FTP, Bram needs to reexamine his fantasy life (which seems to come straight out of a woman-despising serial-killer profile sheet) before he gets the cash investments to play those fantasies out in reality.

The only thing that really differs the wealthy from the poor is the money to act upon their sickest whims without consequence.

Python, of course ;)

[NeuroMorphus]

Yes! Once again Python shows its fangs! ;)

My NAT configuration.

[edunbar93]

Heh. Get this, I have a box running as a NAT-router only (ie, no firewall) with zonealarm on my desktop.
(the reason I'm doing this is mostly because all I need the NAT box for is to share a single IP, and having a real firewall on that got to be too much of a hassle with things like Starcraft and Quake) The NAT box is a P100 running FreeBSD 4.3-Release with natd and ipfw. More interesting is that my NAT box is currently behind *another* NAT box that acts as the gateway router for my ADSL service, also running FreeBSD. (I work for my ISP, which is why I know this :)

When my download started from the site, it was at ~150Kbps. (pretty much the max for my 1.5M/640K ADSL) It slowed down a little as the upstream bandwidth went up, but that was fine, as it consistently stayed at over 100Kbps.

I have a question though. How the hell is it that my upload is working at all? I'm on a network so private that it's scary.

Re:My NAT configuration.

[bramcohen]

How the hell is it that my upload is working at all? I'm on a network so private that it's scary.

BitTorrent connections are bilateral, so you're able to both upload and download on all connections you have to your peers, regardless of which side initiated them. Your uber-NATing keeps anyone else from connecting in, but once you establish a connection out it can send in either direction.

Xolox (gnutella) been doing that for a year now!

[Jagasian]

Xolox, a gnutella for Windows, has been out for a long time now, and it allow for a client to serve parts of a *incompletely* downloaded file. Never underestimate the abilities of GNUtella.

Re:Perhaps misinformed

[anonymous cupboard]

Unless this is very-very new (even newer than Gnucleus-Morphues), Morpheus never allowed another download to start until the first in the chain had completed. Under Fast-Track, the issue was the MD5 checksum which was only calculated on a completed file, never on fragments.

After their move to Gnucleus, I had never seen them do anything that Gnucleus couldn't. However, as the modified source code is included, it should be easy to check.

not ogg complaint

[Anonymous Coward]

Now, who is going to be the first to complain about the use of mp3s instead of oggs?"

Remember, if it's not open source, it's CRAP.

now that's annoying

[Freaky-Monkey]

what the hell is up with that damn HP ad?

jeez... thought i was on Business 2.0's site for a second. how annoying

They know about /. effect...

[AVee]

From the CodeCon website: Welcome, Slashdot visitors! The CodeCon site seems to be holding up just fine, though we've removed our graphics as a precaution. The CodeCon mp3s are also holding up well due to BitTorrent. Please report any client-side problems you encounter.

I just love this, especially on a site thats about how to handle bandwith ;-)

BitTorrent does do integrity checking

[bramcohen]

BitTorrent checks the integrity of each piece of the file individually, based on hashes it gets as part of the original http request. I should probably explain that clearly on the BitTorrent front page, because a lot of people seem to get the impression that it doesn't do any integrity checking at all, when in fact it's integrity checking is extraordinarily good.

Re:BitTorrent does do integrity checking

[Com2Kid]

Well darnit, that just takes all the fun out of it!

You mean no more downloading 600MB-1.2GB files only to find out that 1 or 2 bytes have been corrupted and that the entire download was useless?

Oh man! NOW what I am I going to bitch about? :)

Seriously though, Thank God you have come to the thread! I got here within the first 5 minutes carrying the "It is NOT a p2p application, it is for web servers darnit!" banner, but it seemed to be a losing battle for most of the evening, LOL!

I think that there were just to many links posted in the stories main body, people lost interest partway through and never got to your actual Homepage. . . .

On the plus side though, if that isn't the case and they did get to your homepage, congrats on surviving the /.'ing. :) Though with text only you had good chance for survival from the get go :) . On the flip side seeing as how you are doing network engineering it defiantly wouldn't look good if you didn't survive the /.'ing. ;)

Oh yes, and your page needs more information! Even doing things like saying how it directs users around (does the main server keep a list of users and how far along they are in the clients download process or does it just give the client a list of other clients and say have at it?)

If the information is there, then shove it all into an FAQ and indent to hell!

(see my page for an example of indenting everything until it is blitheringly obvious. :) )

Haha, there's still a bottleneck =)

[Andorion]

The link to the IE plugin is slashdotted =)

-Ando

Upload/Download ratios and ADSL

[osolemirnix]

How will tit-for-tat leech resistance work if someone has an Asynchronous DSL connection? If my download bandwith is 768 kbps but my upload bandwidth is technically limited to 128 kbps (as is common with many DSL offers for private home users), will the leech resistance feature think I'm guilty?

Re:Upload/Download ratios and ADSL

[bramcohen]

Yes, there is a general tendency for people to get about the same download rate as the upload rate they provide, due to the tit-for-tat algorithms. That's just a general tendency though, practical download and upload rates are dependent on many factors.

As to whether being on ADSL makes you 'guilty' I don't know, it's very non-judgemental software :-)

Re:Upload/Download ratios and ADSL

[kaiidth]

Well, that's somewhat less than optimal for those of us with lousy slow upload.

Additionally, maxing out my upload kills downloads entirely, all the way to timeouts (cable connection) - turns out that if I cap uploads at about 5/6 max upload speed, I get normal looking download speed. But another 2k upload and downloads die completely. Looking at the comments further up this page, I can see that other people have had this problem and some have found solutions, so I'll take a look at some of those. But perhaps it wouldn't be an entirely bad idea to consider allowing people to cap uploads at something less than the absolute maximum speed, since otherwise, at least in my case, this software is about as much use as a DOS attack.

Cheers.

Re:Upload/Download ratios and ADSL

[Saeger]

maxing out my upload kills downloads entirely

Same problem here.

I've got RR cable (2Mbps down / 384kbps up) - and as I approach my ~50K/s upload cap, my dl rate suffers a lot, since the ACK's are bottlenecked...

bearshare, edonkey, and other apps are aware of this problem, and allow you to 'cap your cap' by restricting the ul rate; BitTorrent should do the same.

--

Re:Upload/Download ratios and ADSL

[Saeger]

Thought I would provide the average rate I got during the codecon dl demo: 80K down / 48K up.

(If I could cap that 50K to 30K, my dl speed would jump to 240K - allowing others to grab a greater selection of file chunks from me (ala edonkey2k))

--

Re:Upload/Download ratios and ADSL

[bramcohen]

I understand now - maxing out upload causes on some connections for downlad ACKs to get dropped, resulting in lousy download performance. I'll look into capping for that reason. I really wish peoples's ISPs configured their net connections right though.

mp3s are better than oggs for me...

[Anonymous Coward]

I'm using a P166, and while I can play MP3's (using about 90% of my CPU), when I try to play OGG files (that I previously had no trouble with on my 600mhz box) I get skiping.

I'm all for open formats, but when MP3's deliver working music, and OGG's choke up my wimpy box, I stick with MP3's.

BTW- I own all the CD's from which these tracks were riped, and don't support copyright infridgement, so go erase all those illegal MP3's you own....

Re:mp3s are better than oggs for me...

[Enahs]

Are you sure you're using the latest Ogg and Vorbis libs? Older releases used loads of resources. They're slowly being made more processor-efficient.

Of course, they're also still in development, so Vorbis has that against it. But don't give up on it yet. :-)

wouldn't an Apache plugin for this be great?

[elliotj]

If there emerged a distributed downloading standard that was generally agreeable and became better and better the more servers that participated, I would think it would be great to see a plug-in for popular web serving software to support it.

Think about something like this: if you were running a site under Apache and had the option of installing a plug-in that would participate in the file sharing network as a server node. The plug-in would let you allocate a defined amount of disk storage and a defined amount of bandwidth. Then sysadmins who felt this was a good thing could just turn on their participation.

Sure it wouldn't be much at first, but you might get a very large base of servers with good connectivity all playing a role in the system. I think it would help it scale.

Just a thought. I wonder if anyone has considered a scheme like this.

eDonkey

[Anonymous Coward]

But I think a workaround for this would be to have md5 signatures computed for each of these parts and verify them before they are downloaded. I'm not sure if this isn't being done by others already.

A file sharing program called eDonkey does this.

It's fantastic for getting MST3K avis!

---

Check me out!

Re:Xolox (gnutella) been doing that for a year now

[rubybroom]

Ah..so I guess it's incorrect for me to say that the swarming can't occur from people who don't have the complete file.

But I don't think Gnutella allows chaining to occur... By chaining I mean the ability for each user to pass on to the next person the packets of the file they have already saved, while they are downloading other packets of this file at the same time.

Future of FTP

[WyldOne]

IMHO P2P (or something like it) will eventually be implemented on all the anonymouse ftp sites as well. Just need a open- ource standard version and away it would go.

Only thing it would need is a 'auto share partners' system to replace mirroring.

Re:Future of FTP

[Hast]

For that to happen I'd imagine that first we need to see a separation between dowloading the file, browsing a site and searching multiple sites.

The first two are available with FTP, though they are separated which allows you to do some more advanced things. (Like bouncing data without user intervention.) The latter is available with old progams like Archie and Veronica. (Searching multiple FTP sites that is.)

Most current P2P tools put all of these, and a few more, features in one package. Once you start seeing P2P tools which are sufficiently modularized I bet that more people with use them as dedicated servers. But before that you might just as well use FTP/Samba/NFS.

Fewer and fewer peers near the end of the file

[DavidKirkEvans]

I think this is an interesting way to share the load for distribution of large, popular files. Of course, if the files are not popular, nobody will peer the file, and it degenerates to serving the file from the server. So worst case, you gain nothing.

Best case, you get lots and lots of people peering the file, and share the load. I got about 300 k/s download speeds most of the time, and 300 k/s upload.

But near the end of the file, it dropped down to 8 k/s or so. This makes sense, since there will be fewer people sharing up the tail end of the file (they've all finished, and quit serving it) than parts in the middle. So it looks like the server has to handle the last few blocks.

I think it was cute that the window sat around after the download completed, with a "Finish" button waiting to be clicked. So unless you noticed it was done, you would just sit around and share the file up for people. I guess that's reasonable behavior. :)

Re:Fewer and fewer peers near the end of the file

[bramcohen]

Thanks for your feedback. BitTorrent actually downloads pieces in random order, rather than beginning to end, but there's a tendency for you to finish getting all the pieces the peers you're trading with and have to get the last few from slower ones.

This can be fixed by making peers download pieces fewer of their peers have first, to minimize the chances of stranding their peers. It's on the to do list, along with a bunch of other very tricky improvements.

I don't think so

[Anonymous Coward]

Sorry. MojoNation/MNet are much more feature rich and interesting.

for chrissakes read the site before you post!

[mr_burns]

I've read the site and I was at CodeCon for the presentation.

Having said that, most of these comments are ignorant tripe. Before you post, you might want to take a look at the site and read about what actually goes on in BitTorrent. This will help you avoid looking like an ignoranus.

Re:for chrissakes read the site before you post!

[NightEyez]

hehe.. he spelled ignoramus wrong.. what a ignoramus :P

Re:Lossless - big files - FTP is fine

[Baconator]

Etree uses the Shorten codec, which in my experience will compress hi-fi audio to about 3/5 of its original size. A Phish show is typically about 1 GB compressed.

Personally I'd like to see FLAC getting used... in my most recent tests, FLAC blew Shorten out of the water in both speed and compression ratio. Plus FLAC is free as in speech but Shorten is free only as in beer.

Losing connections is less of a problem than you might think... many siteops run *NIX FTP servers, which as a rule are quite reliable. Some of the Win9x servers are less reliable, but that's when you thank god for FTP Resume.

Furthurnet

[Bklyn]

The killer feature of Furthurnet is its treatment of entire sets of files as a single entity. I am not aware of any other P2P alternative which will allow you do download an entire show as a unit, and not have to hunt down each and every track that constitutes a performance. This is a must have feature if what you are downloading is the tracks of a CD. Who wants Scarlet > without the Fire, or Mike's > H2 > without the Weekapaug?

Re:Furthurnet

[Hast]

I do know that Direct Connect let's you download directories. Which is basically this. (As long as the 'publisher' has the data organized.)

Re:Furthurnet

[Miksa]

in eDonkey one can create a collection for example of all the songs from one artist, or all the episodes in one tv-series.

Re:No ogg?

[Anonymous Coward]

tell lunchbox to get back on the right side of the fence. i mean, sure, he loves the cock, so what. but to prefer closed source to open? maybe that's some kind of holdover from living a life in the closet or something.

MojoNation

[Fencepost]

Sounds like this is similar to what MojoNation is/was trying to do. Their site doesn't seem to be responding right now, but here's the Google Cached version of the technical docs.

Re:MojoNation

[haeger]

Mojonation is suspended from what I've heard.
Got this letter from the creator of it, Jim McCoy
"After more than a year of testing the public prototype for the MojoNation
technology platform we are shutting down the public network. The MojoNation
technology will continue to be available via the soon to be announced Mnnet
project, of which more information will be made available at the CodeCon
conference.

It is expected that MNnet will remove several of the remaining centralized
features of the MojoNation technology and result in a somewhat simplified
version of the current system along with native Uis and other fun features.

More info will be made available over the next couple of weeks as details
are worked out.

-Jim"

However the part that was open source lives on under the name Mnet

.haeger

Re:Xolox (gnutella) been doing that for a year now

[Jagasian]

Correct. "Packet passing" is more realtime swarming, if that makes any sense, than Xolox's "parts sharing".

Direct Connect people!

[attackiko]

All these P2P network suck at one thing: sharing. Noone wants to share his files if it is not absolutely necessary. DC handles this problem very well, you can find much more than using eDonkey. Just type 'neo modus' in Google and off you go!

brains..

[flumps]

Great!! When do I get to plug my brain in?
Will I get an ASDL interface to my skull and how much is one?
Imagine.. a beowulf cluster of brains...mmm brains..

For general approach, see also Plan 9 and Venti

[porttikivi]

Propbably BitTorrent conflicts with GetRight an other similar software, too. Those "multiple-TCP-streams to one HTTP/FTP-server, use mirrors too" programs work fine for me, and I get always very good browser download speeds in Scandinavia, mostly good in (50-150 kilobytes/second) from all over Europe and reasonable speeds from the USA. BitTorrent did not seem to do any better.

And I think building algorithmic ad hoc virtual tree and mesh topologies for distribution is an overkill. Simple strategy like Xolox and EDonkey use is good enough: hash the file blocks, index the file-block-hash-chains by the file names, keep searching for blocks by the hashes in the chain, then swarm download the blocks. This automatically swarms incomplete files too.

Hints for intelligents ways to implement hash based content block distribution can be found in the still unannounced "free-as-in-common-sense source" Plan 9 Venti system:

http://groups.google.com/groups?hl=fi&ie=ISO-885 9- 1&oe=ISO-8859-1&q=venti&meta=group%3Dcomp.os.plan9

The Venti article in PDF

[porttikivi]

http://cm.bell-labs.com/cm/cs/who/seanq/venti-fast .pdf

Re:The Venti article in PDF

[porttikivi]

It had an extra blank,this works: http://cm.bell-labs.com/cm/cs/who/seanq/venti-fast .pdf