Slashdot Mirror
Microsoft Wants P2P Avalanche to Crush BitTorrent
pacopico writes "Microsoft seems to think it can be the better Bittorrent. You know faster and more well-behaved. The Register has a story on the P2P work being done by Microsoft's researchers in the UK. Redmond reckons its "Avalanche" technology will be 20 to 30 percent faster than BitTorrent. It's meant for legal downloads only, of course."
Besides BitTorrent might not be the most efficient P2P system any more, but it is one of the most widely used. I guess this is what Microsoft does best, copy other technology, add a little to it, then destroy it.
I thought if any P2P service is meant for legal downloads only, it'll be more than 20-30% faster?
With the size of Microsoft, it can set up a BittoBot farm, which automatically replicate new content across all hosts (internally, so it's as fast as copy-and-paste), then feed everything through a few fat pipes.
If money is not an issue, and world domination (or just the the Lord of the Data) is the ultimate goal, it should be fairly easy for MS.
Rock that crushes, Paper & Scissors that don't matter.
The ultimate in spyware!!!
I can't wait to get my legal Slackware images 20-30% faster :-)
Stick Men
It's meant for legal downloads only, of course.
Then what's the point?
Palladium anyone?
Will it block access to MP3 files and a big list of other file-types/filename-extensions? Like MSN Messenger 7 does? But, like MSN Messenger, allow .WMA files? And do this under the guise of "security", alleging that MP3 is an "unsafe" format (though unlike WMAs, MP3s can't launch websites or "acquire licenses" and stuff like that)..
SCO employee? Check out the bounty
By "more well-behaved" they, of course, mean "DRM capable"... Innovation is taking everyone else's great ideas and adding "DRM capable" to the name.
(Yes, I know there is a bit more to their proposal.)
What kind of legal torrent would you want to download that goes 20-30% faster?
Any versions of Windows that show up on there will either have adware bundled with them or, when installed, will cause blue screens endlessly with the error "AVALANCHED_J00_F00" on them.
And then there's the whole concept of distributing porn via Avalanche; it gives the term snowball a whole new meaning.
Striking fear in the authors of godawful fanfiction, I am here, appearing in darkness, Tuxedo Jack!
I'll gladly use it. Show me the source code first. How do I know there is no RIAA/MPAA spyware in it.
It's illegal "wink wink nudge nudge" to copy Windows 3.1/98 but it helps spread windows users so that's a good thing.
It's illegal "wink wink nudge nudge" to use our faster service, but it helps support Microsoft so that's a good thing.
(It's not a bad idea, if it gets popular enough they can just roll it into Office and charge huge $$$ for it like their MSN Messenger 8...er... Microsoft Virtual Meeting...)
Sounds familiar. My car is meant for legal speeds only. Which is why the "55" is highlighted in a special color. On my 140mph speedometer.
I guess someone in readman finally read the stories about Kazaa's spyware and said "Hey, we're the kings of viruses and spyware! we need in on this p2p thing!"
I only came here to do two things; kick some ass, and drink some beer...looks like we're almost out of beer.
then no matter how "good" MS makes it then it'll never beat bittorrent.
In other words, it'll never beat bittorrent.
"The article is Slashdotted. Anyone have an Avalanche of it?"
So you will find your way to debian_iso.avalanche, download it, and find that it has transformed into a handy little PDF explaining why linux bites...
Spyware is found in Bit Torrent.
Microsoft Releases competitor to Bit Torrent.
Wow, I'm so glad they were so responsive to that problem. It only took them a couple of hours! That's amazing!
Fantasy remains a human right; we make in our measure and in our derivative mode... -- JRR Tolkien
Used to be that fans of Microsoft and proprietary software used to think of others especially the open source community as copycats. Now Microsoft is promising the better search engine, the better command-line shell, and now the better Bitorrent. Is the shoe on the other foot now?
...You will be assimilated.
Microsoft has always been about the assimilation of the technology of other companies...that in itself is no surprise. But between their music subscription service, their new image editing program, and now this, they've fired warning shots across the bows of three different types of applications, all in the space of a week and a half.
Is this just a momentary flurry, or can we expect this escalation to continue?
____
~ |rip/\/\aster /\/\onkey
What about writing better code than open source rivals.
[My english is better than most other people's Turkish, so please point out mistakes politely. Thank you.]
Ha!
You have two hands and one brain, so always code twice as much as you think!
the ensuing legal fight between M$ and the **AA
I'm glad MS is getting in the game. More people will take BT seriously, and I'm sure the BT folks will continue to improve BT.
Maybe this is exactly what we need. Microsoft technology that allows people to steal stuff.
I'd like to see the RIAA/MPAA sue Microsoft for providing a P2P app.
Mh.. Microsoft..
Will linux-images be declared illegal then, too?
This way we can BitT... oh, oh sorry... MStorrent their new OS directly from them?
"Simplify, simplify, simplify!" Thoreau
Now that they can't beat their 'colleagues' with their OS, they're trying it the P2P way.
...
What do you mean, desperate?
...the knee-jerk reactions that this story will elicit? The original post really doesn't do TFA justice.
This is basically an improvement to the BitTorrent protocol that will overcome scheduling difficulties that really do exist today (I need piece X, but the person who has it is busy uploading piece Y).
What it is NOT:
1.) A Microsoft-proprietary application (at least nor yet).
2.) A production application that only runs on Windows.
3.) In any way (in theory, at least) tied to DRM'ing anything.
4.) A way for Microsoft to track your downloading.
Basically, Microsoft has suggested a way to make BitTorrent-like downloads better. Microsoft! Making P2P downloads of large files easier! Really!
This isn't MS search trying to overtake google, or some such. MS isn't trying to own the P2P market (at least not yet). They're suggesting improvements, and if you read TFA, the improvements make sense.
This is a Good Thing. Yeah, I'm suprised it came from M$ too.
to download my illegal version of Longhorn via Avalanch. Microsoft, now making it easier to steal their own products. Ha!
I found it difficult to figure out what the story was. It begins thusly:
The first sentence would have benefitted from quoting "better Bittorrent" or amending "the" to "a". The second from the writer understanding the difference between colloquial and written English. Aren't the editors responsible for fixing such problems before stories appear on the main page?I just can't wait to download MS Office 2006 with MS's own P2P program ;)
1's and 0's should be free.
Microsoft's #1 flaw is that they want to be better than everyone in everything under the Sun. They need to learn to let go (Luke) and let the world be.
Simpy
The way the Register describes it, it appears that rather than sending out chunks of the actual file, it's sending out something similar to PAR chunks where once you have enough data, you can reconstruct the original file.
Futher, with a few chunks, you can calculate new chunks to send over to others, that way more people have access to more of pieces of the file.
Sounds interesting, I wonder if it'll be incorporated into the next version of BT.
Now, I'm not to swift, but from what I've read, the solution will only transfer files that are "signed" by the author in an effort to stem piracy. Add that to the fact that every file chunk will contain unique identifiers about every other file chunk, it would seem that there's already some serious tracking abilities built right in.
:::: the insomniac's digest
Why are they so eager to announce this? The MPAA has already said they want to crush P2P, especially BitTorrent, and they'll do it by polluting the download pool with invalid content. With this announcement, MS is just inviting the same from their detractors. And they have far more detractors than BitTorrent has.
Given their security record, any MS-created P2P application will be just one more gaping hole in their Swiss-cheese-inspired security implementation.
The scary thing is that if you are a windows user, what's the stop M$ from requiring any updates and patches to come through this new P2P system, thus making it almost mandatory to install it on your system if you ever want to update your OS. Microsoft doesn't want to compete, they want to force.
The Technomancer
"Men of lofty genius when they are doing the least work are most active."-
Actually if you read the article you might find this is actually quite innovative. They reckon they have solved the bottleneck problems that you get towards the end of a large file. I am no expert but it sounds like they are using parity.
The open source community has really been the driving force behind technolgoies like BitTorrent. Sure, obviously other applications have good legitimate uses for BitTorrent-like technologies too, but the technology-savvy crowd are really the people who are using things like BitTorrent... whether it's for slackware images, or anime episodes :) With a closed-source solution from MS, I'd be shocked if it gained a huge following. The momentum from the tech crowd just wouldn't be there.
I store my recipes online (the way nature intended)
Sounds like a combination of BitTorrent mixed with PAR functionality. Being able to generate missing segments from the segments you already have. This actually isn't a bad idea. Probably means more overhead in the download though just like there's overhead in PAR files.
Seriously. They can't even come up with a unique name--we had the torrent already, so Microsoft introduces an "avalanche"? Somewhere, Zeus is crying. Fuck 'em.
Comment removed based on user account deletion
I'm going to laugh my ass off when someone finds a trivial way to defeat whatever DRM MS puts into this to make sure the content is legal, and they get sued for helping distribute copywrited material.
Not laugh because they get sued, but laugh because I can almost guarentee that MS has the money and the lawyers to get off on the "we didn't host it" argument. And in doing so, they are big enough to set precident, and will thus free every other p2p software maker as well.
Of course, how damn amusing would it be if their P2P was used to share...illegal copies of MS products?
Velociraptor = Distiraptor / Timeraptor
This reminds me of when Microsoft wanted to crush MP3, and came out with a highly proprietary format that nobody wanted to use for many reasons, one of which being the ability for the software to curb the usage of copyrighted media. I'm not advocating piracy, but if you're already using a tool that does what you want, and is free, and is... (did I mention it was free?) why switch?
Why should users be expected to dump their already-in-place tools and formats for a probably-proprietary version made by microsoft? Its no secret that MS wants to make money, so if you have a choice of a relatively stable and free version, or a new version by microsoft, which would you pick?
And they said zombies weren't real!
You made it man, you fucking lucky sunnofabitch. Microsoft wants to compete with your work, that's a badge of honor man, you're made now.
How about lawful (non-commercial) downloads? It's only commercial if you are using it for commercial purposes. And accordingly, if the download is influenced by third-party interference, then regulation is all-together not applicable.
I use an OS for lawful purposes. Consider non-commercial commercial software. The same can be said about freedom of speach opposing commercial speach, a justice of the peace opposing a Policy Officer, and a man that catches fish to eat (survive) opposing a commercial fisherman.
Commerce is not lawful, it is legal! Don't get in the law merchandise trap: someone selling you laws, because they aren't among We the People and have no standing in the Constitution!
without prejudice
It's meant for legal downloads only, of course.
I wonder how long it'll take to seed my Win2Ksrc.zip file.
Hehe.
"It's not your information. It's information about you" - John Ford, Vice President, Equifax
Even though I can't understand it fully. Can anyone explain it to me a little further how does it exactly work?
On the other hand, I guess a few projects will spring using that same technology.
But I have heard that a similar file format has already been developed, so I guess the real key in this whole endeavour is integration. Can anyone tell me specifically how that (unknown to me) file format works and if it's dissimilar (in any way) to the one from Microsoft?
The nice thing about this technology that it will only be available for Windows. Then again, Apple could create a similar application and call it Tsunami. Gotta love these names.
I'm trying to think if there is anything "Legal" I want to download.
"Simplify, simplify, simplify!" Thoreau
And it will interface with MS Anti Spyware, to avoid all that spyware that comes with BitTorrent, which is hacker software :)
"What's that, Bill?"
"Well, I've been downloading copies of Napoleon Dynamite off of this site called Suprnova.org using this awesome new technology called Bittorrent!"
"What does it do?"
"Well, guys, you won't believe it: It decentralizes the process completely -- no ugly middleman file-sharing program to get in the way! Think we can harness it?"
"Well, Bill, here we go, we created this technology called Avalanche, it's just like WMA! 30% better and it uses the patented Microsoft DRM!"
"Damn bitches, watch this shit take over the market!"
"What the hell, guys? Why is Apple having so much more luck with their video-distribution technology?"
"Umm...... because we completely missed the point of what BitTorrent's all about........ we added a degree of middleman DRM to make sure that nobody was stealing the new Counting Crows album."
"Well, here's what I'll do, guys. I'll mass-market the fuck out of this, and we'll still be winners anyway! I mean, I've been doing that since 1990, why can't I succeed at it now?"
"What do you mean that I'm losing money on this grand experiment? If anyone should be able to push through the status quo, I should! Goddamn it, what the hell happened to that PDF killer I made that wasn't cross-compatible with everything?"
"Well, Bill, your mom always told me that you couldn't keep your hands out of the cookie jar."
"Fuck you guys... I'm going to play in the ball room in my mansion."
* Bill_Gates has left #bad_ideas
NO CARRIER
ShortFormBlog: Writing a little. Saying a lot.
The lead-time required to "authorize" any file I want to put up (how are they going to circumvent basic encrypted files named .jpg?) will nullify any technical advances that will make it 10%+ faster much less 20 or 30 over the same network, namely the internet at large.
Why would I use this again?
Oh yeah, MAYBE POSSIBLY because small underused networks are sometimes easier to search or seed. Avalanche should fit this model nicely since the developers at MS haven't come up with any earth-shatteringly ingenious ways to overcome problems that OSS has been dealing with for years. I don't believe MS has anything to add with Avalanche.
Often wrong but never in doubt.
I am Jack9.
Everyone knows me.
that these fuckers won't get into? okay, video game comsoles. lotsa money there. P2P? how are they going to make a dime? competition is great, but they seem to do things just to do them, when other established technologies, protocols, or systems are in place. and it isn't to enter a market, it's to destroy a market. yeah, i know why they do it, but it is amazing really.
My problem? I was perfectly gruntled, until some numbnuts came by and dissed me.
The best way for Microsoft to keep Google from inovating in yet another area is to get there first.
The best way for Microsoft to prevent Apple from distributing [THE REMAINDER OF THIS COMMENT HAS BEEN LOST DURING THE MICROSOFT TAKE-OVER OF SLASHDOT. YOU MAY NOW RETURN TO YOUR MS-WINDOWS DESKTOP]
"It's the height of ridiculousness to say for those 9 lines you get hundreds of millions."
This is really just getting ridiculous now. Is there any type of software Microsoft won't touch? They need to be getting back to the basics of their buisness - Operating Systems. Maybe then they'd release Longhorn, With New Features! GASP!
And that, my liege, is how we know the Earth to be bannana-shaped.
From TFA:
I guess that's what you get when you have unlimited funds to throw at something....
From the description (and the paper), it seems to be similar to distributing a torrent of PAR2 blocks that are generated automatically from context. I guess they are using redundancy to overcome BitTorrent's statistical distribution problems.
I hope it works, and that they don't patent it, so that the technique will become widespread.
BitTorrent vs Avalance.
This is an academic research paper, and one of the authors in in MS Research UK. He gets paid to come up with researchy ideas, not to build products or do anything related to MS's business. No P2P product will ever be release by MS based on this technology. Why in the world would they want to make a ton of enemies without making any money?
This would save quite a lot of $ in servers hardware for distributing windows updates.
Another solution would be to make less security holes, of course.
Speaking of which, I wonder how many of them will be in this little "innovation"...
1's and 0's should be free.
This sounds like Microsoft re-invented Reed Solomon Error Correction (sometimes called "Forward Error Correction") and simply applied it to BitTorrent.
So, essentially what they're doing is bundling in a PAR like system. This will add quite a bit of overhead as you need to "recover" the entire file using the PAR files, rather than just copying them into the correct spot. I don't see this as a particularly effective idea, as BitTorrent for example will share the slices of a file in a random order, thus the last 2% doesn't always have to be slow because you've run out of people with that last 2%. Instead of falling to NIH syndrome, perhaps they could just have a BitTorrent client that is more likely to pick slices of a file to share that aren't widely shared yet, this would seem to be much more effective and less processor intensive.
For those of you who doesn't know, PAR is a system similar to RAID 5 and is in wide use on Usenet and other potentially unreliable mediums. Files are usually distributed in a spanned archive, and if you're just missing a few of those, you can grab enough of the PAR files to fill in. For example, if you were missing 50 megs of rar files, you could have 50 megs of PAR files instead, which could be used to reconstruct the rar files.
Game! - Where the stick is mightier than the sword!
pirated versions of MS Office became the most popular shared files.
Vote for Pedro
Is that the sound of 'substantial, non-infringing uses?' for p2p? NIce!
TFA does not make any reference to competing with Bittorrent, nor does it indicate that m$ will be releasing their own P2P client. TFA does link to a white paper in which the researchers discuss how they solved the following problem:
Nothing particularly evil about that. No mention of wanting to embrace, extend, and extinguish P2P. In fact, their solution is rather clever. Certainly there was no need for the editors to go trolling...
(Queue "you must be new here" comments)
*** Where are we going? And what's with this handbasket?
It sounds like it performs the same basic function as par files do when included with other files being downloaded. Any missing segment can be recreated from the successfully downloaded segments if a sufficient quantity of par (parity) segments are available.
This would be a natural evalution for Bit Torrent to follow.
I read it, and it doesn't make any sense to me.
But that could just be me.
Anybody who understands it willing and able to describe it in a way that makes sense to any idiot (or at leas this one)?
.mp3 is dead? No one told me. Or any of the other thousands of people who still use it. We need a much better notification system about these things.
maybe this is a hit from the obvious bong, but it seems like microsoft has been really quick(or just plain adamant?) to address all software techonology "successes" (ie itunes, page ranked search, portal, tivo, now bittorrent) with their own MS version. Why is it that average MS loyalist is totally oblivous to the statement that their "new" MS technology, as soon as they download/remove-the-plastic-wrap, is already years removed from the bleeding edge?
Is there no innovation left in that company? Is every "new" MS product just a response to the latest software/net phonomena? Did I just reveal that I've been living in a cave for the last 15 years?
and now back to the fallout shelter...
The sum total of windows machines online already forms a distributed storage network, that is for those of us who can find unpatched, unfirewalled Windows box?
I want to delete my account but Slashdot doesn't allow it.
In other news, Microsoft announces that they believe they can do everything better. Whether it's copulation, corpulence, cognitive dissonance, making cars, driving cars, smashing cars, going into space, falling down, drinking, smoking odd chemicals, giving birth, open heart surgery, baking cakes, frying eggs, urinating, indigestion, folding napkins, digging holes, filling holes, etc., Microsoft has announced that they intend to do it all. There's nothing that Microsoft can't do, and won't do. Soon Microsoft will even be able to do you.
The world's burning. Moped Jesus spotted on I50. Details at 11.
You may be interested in the other comment I posted in this discussion.
Game! - Where the stick is mightier than the sword!
Before people jump up and down on the "legal only" comment, go read the white paper linked in the article. It's actually got some interesting ideas, and (from what I've skimmed through so far) doesn't state anything about dealing with "legal" versus "illegal" downloads - it even mentions Bittorrent favorably in its use to retrieve Linux distributions. The writers just want to tweak out some weaknesses - something I noticed the other day when I was at 99.9% of a recent Bittorrent download, and that 1% had to wait for the one guy's slow pipe to get me the one last piece of the file I needed.
With this technology, my understanding is that my system could have figured out what the last piece looked like on its own rather than waiting four days (!) to get it. Granted, this is just me skimming it for about 2 minutes, but I'll print it out and read at my leisure later.
Now, we can all guess that the technology will be Windows only, blah, blah, blah - but it's still an interesting one, and I recommend you read the PDF document (yes - it's a PDF, not a DOC!) before you pass judgement.
52 Weeks, 52 Religions with John Hummel
MS recognized the existence of Linux. Watch out.
The GeekNights podcast is going strong. Listen!
This is actually a really intelligent thing to say. I have seen discussions on that kind of thing.
If you don't vote, you don't matter, so don't waste your time telling me your opinion
In a recent talk at Stanford Bram Cohen recent commented on people trying to be "clever" in his words. The fact of the matter is that Bram's protocol works because it assumes that every client is out for it's own best interest. The moment you start adding features that require a client give "truethful" information concerning order of packets or whatsoever you open the door for people cheating and trying to leach more than give. MS's idea is just another "clever" idea which forgets this.
This is a first. I did a quick search and there were none of these jokes posted so far. Or maybe they were modded down. On a more serious note, I'm all about this if it works better. Bittorent is a good program/protocol/piece of software. If there is competition to make it better, then that's good. I like the idea of all file parts being of the same importance. It reminds me of what newsgroups used to do with par and par2 files. Is this the same thing?
While BitTorrent made the technology widely available, I don't think it was the first implementation either. Kontiki was using peer-based delivery in the late 1990's, albeit in a highly controlled environment.
Blizard uses bit torrent to distribute updates and i must say it is the slowest piece of junk ever. The first day the patch comes out, you can expect at least 2 hours to download a 45 meg update over broadband. I usually goto download.com and download it there in a quarter of the time.
Have you ever been to a turkish prison?
P2P downloads YOU!
1) 20-30% is not enough to make a noticeable difference.
2) BitTorrent is already the good enough solution.
3) We want an open solution, we already have one.
Now, it doesn't mean it isn't worth it to Microsoft to develop such software. It may save them mucho dollaros when the next major update is released. But it will be used behind the scenes. No way it will displace BitTorrent.
Okay, I read the horrible Register article (I can't stand that site) and followed their link to the original pdf from Microsoft. (The press release mentioned by the Register was not linked to.)
In the pdf, they explain on page 10:
The main advantage of using network coding for distributing
large files is that the scheduling of the content propagation in
the overlay network is much easier. Deciding on the correct
block of information to transmit to another node is difficult
without global information; the transmitted packet is useful
to the receiving node, but, may not be useful to other downstream
nodes. With network coding, each generated block is a
combination of all the blocks available to the transmitter and
thus, if any of them is useful downstream, then the generated
block will also be useful.
Um, so if I have blocks AB and CD, I might send down the pipe block BC? Okay, I see how that helps someone who needs block AB because now they have half of it (B) but they still have to wait on the other half (A) to be sent. If they already had block CD and needed AB, then if I send them BC and later send them AD, then they have in effect had to download block CD twice because they got it once as block CD and again as subblocks BC and CD - am I missing something here?
The attempt to make the files more useful to downstream clients is noble, but while this may help when originally seeding the file from a single server, I would like to see what happens once there are 25 seeds. At that point, I presume the performance would be the same as Bit Torrent.
Your thoughts?
I only came here to do two things; kick some ass, and drink some beer...looks like we're almost out of beer.
Plus, how cool is it going to be to download Windows Server 2006 (or whatever it is) off a P2P network they created.
Find coupons in Greeley
Microsoft's attempt to jump on the Peer2Peer bandwagon is a very good thing. P2P has had the stigma of being merely a playground for hackers, pirates, and open source zealots for too long. Like it or not, Microsoft's involvement in this area will give P2P instant legitimacy among many who have been slow to embrace the technology. Consequently, the MPAA, RIAA, and other entertainment industry cronies won't be able to push through draconian legislation against P2P.
While thinking philosophically, we see problems in places where there are none. -Wittgenstein
Given that Microsoft purchased Groove, I wouldn't be surprised if they came out with a decent P2P product.
And if history is to repeat itself, the MS P2P product won't be appreciated during 1.0 or 2.0 and will start gaining converts around 3.0. Around 5.0 or 6.0 it will have more marketshare than competiting technologies.
Hrm, it looks like the sarcasm bus missed a few people... :-P
So far the vast majority of comments here add up to a tiresome MS-bashing snidefest.
:)
I havent read the PDF yet, but this reminds me of the use of PAR files to fill in for missing chunks when downloading usenet binaries. It's a really cool technique that seems almost magical.
This looks clever - stop being such a bunch of whiny bitches!
it all comes falling down...
I'm only paranoid because everyone is against me...
No, .mp3 is dying, not dead, its right there with apple, BSD, and our civil rights.. (only 1 seems to be true..)
What are we going to do tonight Brain?
methinks one of us forgot to turn on our sarcastron 2000. I'm not sure which, though.
Anyone wanna take a gander at what percentage of bittorrent traffic really is legal? I guarantee no one will be switching to MS over bittorrent if avalanche is controlled in the legality sense...
Here's where I call BS: "20-30% faster."
I don't know. I wouldn't underestimate the MS marketing beast. They've done better before.
Let's say, they tell their users it will be "faster". Everybody knows MS users are idiots. With the new firewall in SP2, there's no way more than 20% of them know how to open a port for bittorrent anyways. Of that, I'd bet even less are motivated to do it all the time. So, bittorrent is either worthless or slow for 80% of Microsoft users.
Bam! In comes the Microsoft "solution": integrate bittorrent into their "OS". The client automatically gets a port opened up whenever it's used. Hordes of idiots go running around saying "it's faster". Add in a few more integration techniques, and it may very well be faster (ie. bittorrent is crippled).
Oh, and also, the whole thing is funded by the RIAA. MS bittorrent checks all the shared files for piracy and/or requires DRM. Step, umm, five? Profit.
"I assumed blithely that there were no elves out there in the darkness"
MS actually have more background in this than you might think. Windows Update uses some clever file transfer technology which is very robust and resumes well (I think it's called BITS or something). They also have some components called specifically designed to enable P2P apps. Their experimental app "3 degrees" used these components in some sort of P2P fashion for music distribution. Probably have done other stuff I'm not aware of.
So, there is a pedigree there.
Read reviews of shopping cart software
I read that in Singapore, the world capital for techo-fascist innovation, trucks would have flashing lights attached to poles on the side of the cab. When a sensor on the engine detected that the truck's speed ever went above 35MPH, the light would start blinking. Then the first police car to see it would issue them a speeding ticket.
If only half the things that I've heard about Singapore are remotely true, then this is one seriously weird place that reasonable people would be wise to avoid.
Obviously no one has told you about "sarcasm" either.
for the last time people, I am "frodo from middle eaRTH", not "middle eaST".
Meanwhile Apple is developing Tsunami and it will be better than both.
mbbac
If I recall correctly Bittorrent will find out if a peer trying to cheat you and gives a corrupt piece. I don't think this is possible with their encoding scheme as it's entirely based on linear combinations. You won't be able to find out who cheated you and the entire file will be useless. That would be an easy way to disrupt people's downloads.
Yep. Apple's a goner. :-P
Work sucked, until it became unemployment, when it became slightly more tolerable. -Tet
Weapons of Mass Distraction?
It's all fun and games until someone loses the key to the handcuffs.
Nice spin on the article title to make it sound like this is some huge MS scheme to take out BT. Anyone with a clue would realize this is nothing more than RESEARCH. Its not being marketted or forced onto anyone. Its just something The REgister got wind of and reported on and in troll fashion its spun into the daily anti MS swill.
Its not Microsoft. It's Microsoft Reasearch. Its nothing more than scientists, engineers and think tankers funded by Microsoft. Nothing more.
Azureus opens my port for me. Yay UPnP.
I've been reading slashdot for a while now and I must say, it's a great news site.
That said, I don't know WHY I bother to read the comments on stories now. Microsoft is trying to improve an awesome technology and overcome the existing problems said technology has. Therefore, using Slashdot User Logic(tm), it is an evil scheme to make money and monopolize the whole market.
They looked at Bittorrent. They saw problems. They figured out a way to improve upon the idea of Bittorrent and ended up creating a new protocol/program/whatever. But, of course, it's Microsoft and everything they do is inherently evil.
As for Linux support, I'm sure the protocol will be easy enough for some Linux developer to implement so don't worry about not having Linux support.
Lighten up, guys.
-Eric Smith
Like Bittorrent, but only legal files. Hmmm... I've heard this before. Its like the Korean movie theatre owner who was unhappy with "The Sound of Music". It was too long, and he couldn't get enough showings each evening.
So he cut out all the songs.
Porn.
I agree, to a point. MS involving themselves with P2P opens a brand-new field of pros and cons. The good news: Any MS-bred P2P app will help legitimize the technology. But there's also a lot of bad news: what if we see M$ P2P apps that somehow set themselves up as "defaults" a la IE, and/or hijack functionality / block other programs?
And what about MS' "corporate responsibility" to stop copyright infringement? Let's say you download a torrent and try to open it; but instead of BT or BitLord or whichever flavor you're using, an "Avalanche" window opens, because it has associated itself with .torrents. Then, a message box opens: CSI_s3ep24_tvrip_xvid.avi is a copyright protected file. This download has been aborted; your IP address has been logged. If you have any questions, please go to [address of anti-piracy site here].
This is a little far-fetched, but then again, it might not be so far off.
Bottom line? MS might do a good job of popularizing and, to an extent, legitimizing the underlying principles of P2P. But I don't trust 'em as far as I could throw 'em, as far as their role as a content distributor is concerned.
Would I want to use the bandwith I paid for so other Windoze lusers can leech off the copy of Longhorn Service Pack 3 that I downloaded? Microsoft wants me to take part in some damn hippy-dippy bandwith commune? While they're world renowned for not playing nice with others?!!! Get the fuck out!!!! You can't have it both ways Microsoft!!!!
the preceding comment is my own and in no way reflects the opinion of the Joint Chiefs of Staff
Is what it sounds like to me. Basically they create a lot of parity blocks as well that can be used to recreate missing blocks. One can do this today w/ bittorrent and a smart client, such that the torrent would contain AVI file of say 700 blocks (1 per MB) and another 50 of parity blocks. So the torrent would describe 750MB of data. However, one would only have to download any 700 of those blocks as the rest could be simply recreated.
I gave this some thought a while ago, unsure if it really solves a pressing problem.
This paper is from some researchers who have nothing to do with Microsoft's products. MS may not ever use this technology in any product. And if MS does use Avalanche for something, it will probably be buried away inside some other application (like Windows Update) instead of a standalone app.
Actually, if you read the actual research paper, you can see WHY it's faster. Basically, it combines two technologies. A bittorrent like protocol, and file parity generation (such as PAR). This allows you to generate additional pieces you didn't download and reduce the amount of code you need to download by about 20-30%.
This also solves "the last block" problem where everyone is waiting for the last block, since if you have 99% of the blocks you can generate what's left.
It's an interesting approach.
If you need web hosting, you could do worse than here
er, what's wrong with this cheap watercolor photocopy?
by 'better behaved' perhaps microsoft means that all avalanche clients will now open partially-documented 'administration ports' on ports 135-139 which will allow (cough) 'global centralised management' and (cough) 'lower TCO'.
MS obtains press bragging rights for one business day, and the following month somehow 21.8 million PC are strangely infected with 'fonehome.reportallOSSapp.licenseviolations'.
MS does not deserve to utter the words 'better behaved' in the context of the development of its own products. Not in this decade at least.
-k0ward
Look who's talking. ;)
Distribute randomly constructed packets, and, at some point in time, you will get the whole file(s) :)
I have discovered a truly marvelous proof of killer sig, which this margin is too narrow to contain.
This is addressed briefly at the end of the paper.
...will be Longhorn betas. Funny how that works out. :)
Laughter is the Spackle of the Soul.
It seems Microsoft is falling into the Google trap by expanding their product line (e.g., this and Acrylic). That now has become a boring avenue.
Google was avoiding the fight, but Microsoft failed to see the advantage.
So they are doomed.
Just kidding.
http://stephan.sugarmotor.org
...is how long till u see those Avalanche Longhorn ISOs floating around? I 3 Irony.
hey now we got a legal mircosoft supported way to share our corperate versions of mircosoft office an d that free copy of windows xp that microsoft doesnt have enough bandwidth to offer us ;)
there free right
...that M$ went to great lengths to hide the history trails of IE. What's to stop M$ from hiding your torrent download habits then selling the back door to the MPAA so they can sue you?
Eternity: will that be smoking, or non-smoking? I Corinthians 6:9-10
They've been trying to stifle innovation for long enough with the old "The VCR will destroy Hollywood!" BS. And it took Sony, with the deep pockets they have, to protect the rights of the public, so that today I can use my Video Cassette Recorder without worrying about being taken to jail.
Now if Micro$oft will come to bat for us....
Imagine the possible irony if MS steps up to protect the rights of the public against the special interests...
I must be dreaming.
-ex
Swarm downloads with FEC coded pieces was implemented by Justin Chapweske's Swarmcast in the summer/fall of 2000 at the latest. A year before Bram Cohen even started working on BitTorrent. Bram was using FEC in the Mojo Nation system at the time. Despite this, both Justin and Bram decided not to use FEC in their next systems, WebRAID and BitTorrent.
For their next trick, Microsoft Research Cambridge is investigating how to do one-click shopping by utilizing a database of customer information and browser cookies.
Congestion ... "bad weather" on the Internet or (more likely) at your ISP can make the available bandwith vary by orders of magnitude day to day. A new botnet turns up, your neighbor ticks off someone on IRC, zap, you're down by 50% or more. Nobody's going to worry about a factor of 0.30 difference in download speeds next to that, it's lost in the noise.
"The researchers claim download times are between 20-30 per cent faster"
:-)
I would love to know how they can get 20-30 percent more information down my connection than an ordinary BitTorrent connection (which can, and almost always does, fully flood my connection) in the same time. This is some AMAZING compression or transmission technology there... I bet you can then compress it even more by compressing the compressed data.
I'm assuming they mean that they can improve download speeds on low-usage "avalanches" (nice naming, by the way) because each peer doesn't need the full file necessarily. However this must mean that, at some point, extra information is added to each "chunk" of the file, to be able to replicate the information from other, missing chunks. Hence, people are just effectively sending out larger chunks of information into the void.
It might be a more efficient way of doing things, yes, and it might well save bandwidth for the initial peer but you're still going to run into the exact same problems as BitTorrent (no download without "enough" peers with "enough" info, for example), it won't become a standard without a massive end-user difference from BitTorrent and global uptake.
Almost any torrent I use is peered enough to download at top-whack for my broadband connection. What's in it for me to run ANOTHER protocol, with another set of servers, another non-standard standard when I can't download any faster at all?
Maybe servers would like this, but then for initial releases any popular torrent only need seed one, maybe two COMPLETE copies of a file to let the peers take over and finish the job off. With avalanche, surely the server would still have to send out roughly this same amount (or more) of information, the client would have to recieve roughly the same amount (or more) of information and might even end up sending out a lot more info?
Where's the advantage? Is it better if you are constantly seeding forever, is it better if you are seeding LOTS of different stuff? I don't see how the maths adds up for any non-trivial "avalanche".
"No one peer can become a bottle neck"
"overall network traffic is lower"
So this is advantageous only to the TCP/IP middlemen? Does that mean I have to change all my protocols and see next to no benefit for the sake of that transatlantic pipeline to the US? I don't see how anybody else really benefits.
...on a world and global scale.
(puts on Satan hat)
You would think Unix environmentalists would want DRM, so to help Microsoft develop a better Unix that would better manage file permitions than a combination of read, write, and execute for a single person and a group.
As much as DRM could hurt people, with the little overlord in me, DRM could be used to hurt Microsoft just as well because 1)I bought Microsoft Windows DRM edition, 2)There is no support contract with Microsoft that is enforcable or legitimately active, 3)I own the software and am a shareholder, 4)Microsoft doesn't own the software anymore because they have sold it to me, 5)I do not give Microsoft permition to administer the data on this harddrive.
(removes Satan hat)
All that has previously been held as valid in a court. DRM: whoever gets their hands on userID 0 is in control of the King's Banc (bench); count your blessings, dispel the curses.
without prejudice
And doesn't Spirent Communication own the trademark word "Avalanche" for their high-speed HTTP packet accelerator?
regretably, they do get some
No one seems to be approaching this from the technical viewpoint, so I will try.
First the term used most often in the paper: network coding.
It looks like they're basically taking a shortcut with the bittorrent method. The neat thing is that this looks easily implemented in bittorrent. They use the term linear combination a lot which, as near as I can tell, means that they'll be doing something like this:
There will be your packets floating around just like in bittorrent. Suppose packets A, B, C, and D make up your file. The Avalanche method would work exactly like bittorrent except at some point there will also be mixed packets created. These mixed packets, the linear combination thing, would be like a packet that's actually AxorB (xor doesn't appear to be the specific function they use since they keep saying "linear combination", but it would work). The idea is that if you can get packet A, B, and C then you'll be able to get packet D by getting either D itself or AxD, BxD, or CxD. This would make the packet that you need to finish 4 times more common, and at the same time it looks like the system in this case would be handling only 10 different kinds of packets (A,B,C,D,AxB,AxC,AxD,BxC,BxD,CxD). This makes it almost twice as fast as bittorrent for finding you that last packet. The 20%-30% number seems pretty accurate.
They're likely using something more clever than a simple xor. They probably have something in their little "linear combination" algorithm that would let you get your whole file even if all you had were mixed packets.
Direct away from face when opening.
"You know faster and more well-behaved."
The above quote is a classic example of a sentence fragment. Please return to elementary school to complete the second grade, Taco.
All the Linux folks need to do is find the security holes in Avalanche, and share them with spammers, spyware vendors, etc. Meanwhile, of course, filters to keep BitTorrent clean would be put into place....
When you use an app that wants to accept inbound traffic, Windows Firewall asks if that's okay and automatically opens the port if it is, so you don't need to know how to open a port.
Just RAR it up and stick a few PARs in there before making the torrent? Oh wait, I forgot, bittorrent users only like sexy new stuff.
I am trolling
Yeah,they seem to think this about a lot of things.
How is it going to be "faster". Will the software increase the theoretical maximum speed of my cable modem service? Will I magically go from 7mbit/sec to 10mbit second with with this P2P software?
"Now with 20-30% more vapor!!!!"
the preceding comment is my own and in no way reflects the opinion of the Joint Chiefs of Staff
They could want the moon to be made of marzipan, too, for all the good it would do them.
Their efforts to crush things in the last few years have been rather unsuccessful. Most of their stuff hangs around, but when I hear of them trying to crush something I just have to think back to Blackbird, C#, etc etc. Bittorrent ain't goin nowhere.
Intolerance for ambiguity is the mark of the authoritarian personality.
Pure genious and innovation.
not to nitpick but Kademlia came out of NYU Secure Systems Group, not Microsoft...
Umm.. no. BitTorrent works on the block level, not the file level. You end with with only parts of files. A partial rar is usless. You need the parity to be on the block level, not the file level.
If you need web hosting, you could do worse than here
That there seems to be no consideration these days of whether a headline or article summary comprises anything even close to a gramatically well-formed thought before it's posted?
And in this case, by creating a BitTorrent work-alike, they can draw up patent specs that INCLUDE BitTorrent's features, and then use that patent to shut down the servers. Time to start informing the Patent Offices!
Also, folks, make a note of the DATE of that paper describing Avalanche. One PTO rule that seems to me gets violated often is that there is supposed to be (or used to be) a one-year limit between the public release of an invention's description and the patent application. After more than a year, it's too late to apply. How many existing dubious patents were applied-for too late and could be overturned on those grounds?
TANSTAAFL. If the file is already well compressed, generating blocks from parity information won't make it faster - because there's no more redundancy to be squeezed out, you have to transfer more data than the raw file to be able to have the extra information to generate these blocks. Sure, you may be able to get it slightly faster, perhaps 0.5-1% - but certainly not 20-30%, unless the file is uncompressed or only lightly compressed.
Most files that go out via BT are very well compressed.
Oolite: Elite-like game. For Mac, Linux and Windows
Either you're exagerating the part about maxing out your broadband connection, or your bandwidth is considerably smaller than most people's and therefore it does not take much to reach the limit.
I have a 1 MB/sec downstream connection (which I believe is very common for home cable modem users). I have never once encountered a single BitTorrent download where my download speed went over 250 KB/sec, and I can count on one hand the number of times it even got up that high. The vast majority of my BitTorrent downloads max out at approximately 100-150 KB/sec. Therefore, I'm assuming you have a download limit of something like 128 KB/sec. Is that correct? Or, perhaps, something close to one of the slower DSL hookups?
If I'm wrong, please enlighten me on how it is possible for you to get torrent download speeds so much faster than mine. To me, downloading a torrent is only worthwhile when the file I want is on a server under a massive amount of stress or if it's just not available through any other means. Direct downloads, for me, are still MUCH faster than torrents. Even non-torrent P2P clients, like Shareaza and Ares, are considerably faster than torrents.
UNIX: A computer user is defined as a programmer. WINDOWS: A computer user is defined as a consumer.
I'm sure that once M$ maps out how their program works, other distributions like BT++ and Azureus will be all over it, implementing it in their own programs. If anything, M$ will have the advantage for about a month after it's beta release.
Just what exactly is microsoft planning on distributing other than patches to their crummy software? MS isn't open source and open source seems to be about the only legitimate product to be shared among P2P networks. With a closed software solution they'll just fail with P2P, a freedom technology.
Besides BitTorrent might not be the most efficient P2P system any more, but it is one of the most widely used.
.mp3s with their .wma files and banished .mp3s into oblivion? Er, scratch that. Recall how MS won the browser wars once and for all by building IE into the OS? Um, ok skip that one too...
No way man, just because bittorrent came first, doesn't mean that MS can't improve on it and release it as their own. Remember how MS improved on
Microsoft - We make software so lame that even grandma won't be threatened by it...
HA! I just wasted some of your bandwidth with a frivolous sig!
Of course, how damn amusing would it be if their P2P was used to share...illegal copies of MS products?
What do you mean, "if"?
The enemies of Democracy are
Why is it that when ever Microsoft enters a market they must 'crush' all competitors. Can't they just get along? Is it the 'in order to win, everyone else must lose' school of thinking?
Why can't we all just get along?!??!!??!?!!
I don't think i was ever asked to open a port with my torrent client. And i do have XP-SP2. If anything it did the popup window asking me if it was ok to access the internet.
The good think about their firewall is it's very user friendly so you don't have to manually go open ports up.
Nothing really major to see here, basically the entire paper and news article are the "discovery" that Reed-Solomon error correction can be applied to a peer-to-peer network.
o rrection
See: http://en.wikipedia.org/wiki/Reed-Solomon_error_c
And honestly, how long do you think before Microsoft patents the use of distributed generation of forward error correction packets in a peer-to-peer network?
Yay! Now the two titans we all hate can collide and kill each other off! MPAA vs. Microsoft!!! And the winners? Bittorrent users, because the MPAA will see more money from going after MS, rather than poor kids that use Bittorrent!
I'm reminded of two words from the Microsoft glossary:
'Embrace'
and
'Extend'
Together, when used by MS, they are enough to cause many developers to suffer from severe gastro-intestinal distress.
My office has been taken over by iPod people.
Let's see what percentage of illicit file sharers end up getting arrested on anonymous complaints that are later found to hail from Redmond, WA, USA.
While the parity system is nice, we need distributed parity systems with encryption and without connection logging. Microsoft not log, not trace, not interfere?
I can't laugh hard enough at that concept.
If my grammar and spelling are off, I am [distracted/tired/careless] (take your pick)
BitTorrent is a protocol. Not an application.
It isn't meant for anything, legal or illegal. Think about it.
And even if you write a protocol for illegal things, I do not think that should proclude it from being available.
It's this half assed bad thinking in society that just begs for dumb laws. When we question the legitimacy of something tat is UNQUESTIONABLY legitimate, how will the other people, who are not tech savvy look at something that they see as a tool to remove money from their pocket?
If one more person talks about legitimising bit-torrent again... I mean, who the hell was the first person to say it, and everyone latched on.
Rather than fight the notion that an idea (and code is a well defined idea) needs to be validated in case it can be used for bad purposes, we give CREDIT to the implausible, and try and defend against it, when it does not exist, and then, one day, without realising, it will exist, because we were fighting the wrong battle.
to the asshole who has written the "To confirm you're not a script" logo, go fuck yourself, just because something can be done, doesn't mean it should be done. PS: The Y letter is cropped, and next to a V letter, so they look almost indistinguishable. DO SOMETHING ELSE WITH ALL YOUR FREE TIME. Jackass.
#hostfile 0.0.0.0 primidi.com 0.0.0.0 www.primidi.com 0.0.0.0 radio.weblogs.com
Comment removed based on user account deletion
This also solves "the last block" problem where everyone is waiting for the last block, since if you have 99% of the blocks you can generate what's left.
.PARs don't require any less data to be downloaded, it's just that you can substitute parity data for the original data, then do whatever transformation on that to get the original data back. If the file you're trying to get it 1GB, you're still going to need to download 1GB, whether it's 100% original data, 80/20, or anywhere in between.
Not really, it just (possibly) changes the nature of the last block.
The only thing this really helps is if clients prioritize the parity data and then all seeds disappear, although it's of very limited use there as well, since the data shared between the remaining peers still needs to total 100% of the file size.
https://www.eff.org/https-everywhere
It seems that the only real benefit of a distributed PAR system would be that more bits of the file are out there for you to pull from. That could be easily duplicated in BT by forcing everyone to have a share ratio of 1.25 or higher before they can disconnect.
Probably too late to add PARs to the bittorrent design, I expect Microsoft has now patented the technique.
a,e,i,o,u and sometimes w and y (at be if of up cwm by)
personally im getting sick and godamn tired of people wanting my bandwidth for nothing, half the reason i REFUSE to play world of warcraft, im REQUIRED to pay for it AND host patchs for the company, - you want my bandwidth you give me the fucking game, as for p2p for piracy grow the hell up and get pds (private dump site) for the luser group.
i guess if you let the lowest common denominator play it drags the whole net down.
ahhhh... i got ya
Get your Unix fortune now!
The program will run fine, return mediocre results, and download at 80% of the speed of bit torrent but since it comes pre-installed with every windows-longhorn it will kill off bit-torrent.
"There are more things in heaven and earth, Horatio, than are dreamt of in your philosophy."
the bitTorrent creator did not patent it.
Microsoft is just copying the idea.
or at the very least.. the RIAA will sue MS and they'll both bankrupt eachother in legal fees, and then the lawyers win... wait this started off as a good thing
That only leaves Linux ISOs, doesn't it?
to be easily usable. The technique of using parity to add some redundancy is pretty old, hardly patentable (unless you live in the US).
Recently Azureus and Bittorrent proper extended the protocol with distributed trackers. Who says they cannot extend it another time to include a parity bottleneck remover? The extra info you need would not be very hard math (it has been done before) and is mostly above the transport layer. You just need to know how much you need to download and when you can stop to do the magic trick with the parity reconstitution.
But inherent to the protocol is some sort of bottleneck at some percentage. Say if they add 10% redundancy, that means a tracker/seeder that stopped before 90% (of a download that is now at least 10 % bigger) still will stall all downloaders. This technology will mostly work for trackers with a small number of peers. If the seeder removes right after seeding one copy, a tracker with this technology will have more chance of not getting deadlocked.
But caveat lector: I did not read that MS article and never studied the math.
This space is intentionally staring blankly at you
If you stop and think for a minute, you'll realize that if Microsoft does anything with this technology, it will probably only be used for very specific applications (such as distributing Windows Updates), and that it will not be released as a end-user application that is designed for general file sharing.
Arguing about vi versus Emacs is like arguing whether it's better to make fire by rubbing sticks or banging rocks.
This is a bunch of systems researchers at a Microsoft-funded facility who've been fiddling with new methods. Kudos to them. Actually, if this is the facility I've heard about (I study in the area) they've got a significant portion of their computers on Linux, and don't tell Bill I told you that.
For the love of God, please learn to spell "ridiculous"!!!
"since if you have 99% of the blocks you can generate what's left".
Ok.
How did you get those 99% already ? How long did you wait for those ?
I mean you're moving the actual value at which you get problems, but it's not magic.
It's like having an amplifier that has Volume values between 1 and 9 because your ears are sensitive...
Comcast offers residential users up to 6mbps down for $65/month. If you have their base service for $55/month, they offer the bump up to 6mbps (from 4mbps) for $10 more. It'll also bump your upstream to 768k.
Their business package can far exceed 8 megabits.
Imaginary hypothetical hallway conversation in redmond...
Bill, the Architect: so what's up with you folks down here in the basement, I haven't been down here for a while...
Ms. Ava, the Research Lab Tech: after we finshed rediscovering electricity and patented it, we heard of this crazy fast distributed downloading system where you rely on people installing spyware to zombify their machines instead of buying servers and reinvented it, but we can't find a legitimate use for it yet other than download videos of cspan and subpona-ed legal documents for SCO...
Mr. Wu, the Product Support Tech: you think that's bad, we just got the lastest round of bug fixing in windows, but the hotfix is 2Gbytes and we figured out we'd have to spend $100M on new servers to upgrade our website to release it and we don't have the budget for it...
Bill the Architect: Mr. Wu, meet Ms. Ava. That's enough for today, I'm tired now, back to my shiny Medina castle to count my money...
They want. Hm... I just noticed that in recent years somehow less of what they want happens. They'll have to come to terms with that. Kind of hard after the ride to the top Darth Gates and his apprentices enjoyed...
You're an idiot. PAR data isn't free. You don't download less... you just get some redundancy that you can use to rebuild lost/corrupted parts.
I have a 4Mbit(down) cable connection, and I routinely get 420+ kBps when I connect to a busy torrent, or one that has a few pretty quick-uploading peers. With protocol overhead accounted for, that's quite near maxed.
What could be causing your problems is that your download is being choked by ACKs. That is, packets are acknowledged when you receive them. On a non-symmetrical link, if your uploads have your entire upload capacity filled, your computer can't send many acknowledgements, and the peer(s) have to wait to get them--so they know when to send more data... Or something along those lines.
You might try lowering the upload cap on your torrent client just a bit to see if it helps. If it does, that might have been the problem.
It's hard to say, though. Perhaps your ISP has implemented some QOS rules that really slow down torrent traffic, or your firewalls/routers are incorrectly configured. It could be a dozen things. But, in my experience, Bittorrent works damn well, and usually much faster than a plain old connection to a very high in demand file.
Constitutional rights may be respected, repealed, or modified; but they must never be ignored.
It's them trying to protect themselves - but when p2p hubs put up notes that users are liable for their own use the Fed's storm down their doors anyway.... but the p2p program writers seem to be protected so far... unless they host the files on a centralized network.... then their screwed.
Given the problem, the solution is obvious. It's exactly the same as error correction when you know all the positions that have errors. A parity check block is the simplest case.
More generally, you have k values position 1..k, you run an k-1 degree polynomial through them, you also send the values for that polynomial at positions k+1..n, then whichever k values you get first (provided you know their positions) you can derive the polynomial and from that all the other values. When the values are whole blocks, it's easy to include the positions.
I can't say the whole thing is obvious because I wasn't aware of the problem.
Heh, no, I'm not impatient. I'm just downloading at 1/10th my capacity over a protocol which people claim is the fastest way to download.
And yeah, I do have over 8 megabit (1 megaBYTE) down and about 1 megabit (128 KB) up. It's "Optimum Online" in New York, it's their standard speed package. In fact, I'm pretty sure it's Comcast's standard speed, too, and virtually every other major cable provider I've ever heard of. Unless you have DSL (which you do), there's no reason why your provider would offer you anything less than 8 megabits downstream.
I doubt my ISP is interfering in any way, especially since BitTorrent is the one and only protocol through which it's impossible for me to download at faster speeds than what I mentioned previously. Bottom line is that, to high-speed cable broadband users like me, BitTorrent is far from the most efficient download protocol. If Microsoft's method would give me 30% improvement, I'm all for that.
UNIX: A computer user is defined as a programmer. WINDOWS: A computer user is defined as a consumer.
See, and I thought the 55 mph was highlighted in cars from the 1980's due to the fuel shortage.
My car is a 2004 model, nothing is highlighted (speedo is digital), but it does have a nice red bar at 9200 rpm for redline.
SearchIRC - Now with live chat directory!
Every time someone asks you for a block, you send them a new block, which is a random linear combination of all the blocks you have. This new block will almost always be useful to them. As soon as you get n blocks, where n is the number of blocks in the original file, you can reconstruct the original file. So bandwidth is never wasted sending a block the long way when the short way would do - you squeeze the maximum work from every hop.
The really interesting bit is right at the end, almost as an aside:
"In Avalance we use special sets of secure hash functions that survive network coding operations and consume very little computational resources"
So even though each block is novel, they have a way for the receiver to ensure that it's a real piece of the puzzle. That's a hard problem indeed! So why isn't the solution part of the paper? Are they holding off from publishing that until the patent comes through?
Xenu loves you!
+1 Funny...
:o) // so Im not moded down =oP
Man, thnks for the laugh before sleep.
Really, thank you!
p.s. I know I will be moded down
Ubuntu is an African word meaning 'I can't configure Debian'
Use PAR2 then. A partial rar is not useless with par2.
Opening the source wouldn't help as much as opening the protocol itself... that way people could make new clients without needing to refer to the source, and you wouldn't have to worry so much about the protocol being changed underneath you.
Karma: It's all a bunch of tree-huggin' hippy crap!
Listen. This is very important. A 1Mbps cable connection is very common for cable users, a 1MBps connection is not. And it is never sold under that term in consumer broadband. A 1MBps connection would be sold as an 8Mbps connection. Because of these discrepancies I'm going to assume you meant 1Mbps. This would explain you never having a torrent go over 250KBps as your connection couldn't even handle it (1Mbps is 128KBps, exactly what you seem to think his download limit is and what you think torrents magically limit out to... in fact is probably your own download limit). If you have a connection which isn't rediculously assymetrical you can max just about anything out on many, many torrents. If you can upload at even 30KB/s you can max out a 10mbps pipe on anything with a lot of users on it.
--
WHO ATE MY BREAKFAST PANTS?
Actually, in a lot of countries residental 10megabit and up is getting more and more common. Where I come from you can get up to 50megabit, and in korea, only old people use 50megabit. There's also adsl2 and various other fiber solutions. Right now I am at college, and i got a shared 100mbit connection. The best my neighbour has done on BT-dls was over 8000k/sec...
Rest in peace Malin "looxn" Kristiansen. We miss you...
If BSD dying is counted, then two are true. Civil Rights are dying as well.
you'd probably get caned once a week
Most files that go out via BT are very well compressed.
Many ISOs are not compressed. I guess that means that the only thing 20-30% faster will be Linux ISOs!
> I personally think there's no place in the world better to sample many different cultures' foods at once.
Toronto.
TANSTAAFL
QFT. You have to download enough bits to have the file. Parity wont decrease the number, it only helps if you're missing pieces. HOWEVER, there are practical scenarios, especially dealing with users on slow links in the very beginning, or maybe just mayst a not very popular, torrent. Its more noticable when someone has a very large difference in up and down speeds.
Example: 5 peers, 1 seed. After the seed has sent out enough data that there is another complete copy of the file amongst the five peers. If one or more of the peers is very slow upstream, their blocks would be the hardest to get (and if the seed disconnects, you're SOL), and it might help to have some general purpose parity files spread out on all the peers, so that the fastest upstream links can provide these pairty blocks to "fill in" gaps where some blocks are not evenly distributed and are on slow links.
The Doormat
If you're not outraged, then you're not paying attention.
Microsoft doesn't notice that the reason most people use filesharing software is to obtain illegal software
Windows Update is sometimes inexplicably slow about installing the patches that it downloads, but in my experience, the download itself has generally zipped right along. I snagged 8MB of patches in under a minute this morning. (And no, I'm not really thrilled that my simple e-mailing and web-reading desktop needs 8MB of patches every couple of months.)
I'd guess that on the scale that Microsoft buys it, bandwidth is dirt cheap and getting cheaper. In that case, what compelling business case is there to resort to a peer-to-peer strategy to distribute patches? They can probably afford to budget, say, 5GB of transfer for patch downloads for every OS license sold. How much would that be, US$0.50? Less? Why bother with all the security headaches of a peer-to-peer system when a client-server system offers better protection, control, and is so cheap?
I had figured that Valve's Steam platform would use some sort of peer-to-peer boost, especially when they hired Mr. Cohen, but I guess Valve decided the same way on the security-headaches vs. bandwith-is-cheap issue.
Heh, umm, no. When I said 1 MB/sec, I wasn't using my ISP's terminology, I was just saying that because that's a much simpler term to understand for most people than using the term "megabit". I do in fact max out at ONE MEGABYTE PER SECOND for downloads and at approximately 128 KILOBYTES PER SECOND for uploads. And you're wrong about 1Mbps being common for cable. That's DSL's standard speed, and DSL is usually offered at much slower rates than cable.
I can download 5 different torrents SIMULTANEOUSLY, each one going at about 150 KB/sec. Or I can download 1 torrent going at about 150 KB/sec. Very rarely does this speed ever go any higher than that.
And, just to repeat, whenever I say kilobytes and megabytes, that's exactly what I mean. I know perfectly well the difference between a bit and a byte.
UNIX: A computer user is defined as a programmer. WINDOWS: A computer user is defined as a consumer.
Last time I looked at Comcast's Terms of Service, they still had extremely fascist policies about what you could and couldn't do with your service, e.g. not running web servers or mail servers, and it was pretty obvious that they considered Napster and newer P2P applications as Evil. While I do have their television service in my house, it's not worth dealing with an "Internet Service Provider" that doesn't like the end-to-end principle that makes Internet Services work, or in general with anybody whose policies are noticably more restrictive than Sonic.net (whom I use) or Speakeasy.
Bill Stewart
New Fast-Compression-only CPR http://preview.tinyurl.com/dy575ks
Not with SP1
Would you like to inform us all of which Bittorrent program comes with spyware? AFAICT There's not one single BT frontend or software that comes with spyware.
Still waiting on Serviscope_minor to wake up to fucking reality and realize that Jessica Price isn't going to fuck him.
You have a 8 mbps connection? What city, what company, and how much??? I have only a 5mmbps connection, and I get about 615 KB/S Downstream... Please imform me, so I might upgrade, the extra bandwidth my ISP gave me (3 to 5 mbps) cut off about 25% of my gentoo installation and updates, another 3 mbps would be a great help.
sorry for the OT, but I thought many users might be interested in a faster service.
Watch for Penguins, they eat Apples and throw rocks at Windows.
Every torrent I've ever downloaded with over about 50 peers maxed out my broadband connection, up and down
Sounds like you should upgrade your 'broadband' connection then. I am currently downloading a mainstream torrent file from a mainstream tracker site. UL = 36 kbytes/sec. DL = 65 kbytes/sec. The only thing that is unusual about this is that the upload is kind of slow.
My download bandwidth is 10 mbit/sec theoretical and around 500-750 kbyte/sec in reality (maybe partly due to my modem). The fastest I have ever downloaded anything with bittorrent is around 120-150 kbytes/sec and that is often with thousands of peers and seeds. Typically I get speeds of around 50-75 kbytes/sec.
An important limitation to download speed in BT is your upload speed. I never get much more than a 1:2 UL/DL ratio. Usually if I can increase my UL I can increase my DL as well. This is the main reason IMO why BT is so much faster than Emule etc. It actively encourages people to devote as much UL bandwidth as possible to the torrent.
I have read the article and this tech sounds pretty cool to me. People have been talking for years about ways of incorporating error correction technologies into P2P clients. It works great for usenet releases. The biggest problem with BT IMO is not speed, but chunk availability. Torrents die off too quickly due to lack of seeds. Sounds like this tech could extend the life of a seedless torrent.
In my view, the biggest limitation on torrent speed is the archaic assymetrical limitations placed on most broadband users these days. I understand that ADSL is an inherently assymetrical technology, but cable modems are not. ISPs need to realize that people are not just browsing and downloading from websites anymore.
Quite an experience to live in fear, isn't it? That's what it is to be a slave.
Harbinger
There's been a lot of comments about how M$ will try to compete with BitTorrent and how the whole idea sucks just because it originates from M$.
BUT. There is nothing stopping Bram Cohen from incorporating these new ideas into a future version of BitTorrent. The ideas behind Avalanche ARE good and would improve BitTorrent just as they claim. Just because the idea came from M$ doesn't mean that we cannot benefit from it.
You mean like the difference between Xerox and Xerox PARC?
Optimum Online
UNIX: A computer user is defined as a programmer. WINDOWS: A computer user is defined as a consumer.
Let's say that you have a bunch of people using BitTorrent. The only people who have segment 499 are behind slow modems. But lots of people want those.
If there's a rare part, you only need one downloader with a decent upstream to break the bottleneck. By the rarest-first scheduling algorithm used in both BitTorrent and eMule, the rarity of segment 499 would have long ago prompted some user with broadband to go get segment 499 from the dial-up user and then start seeding it out to other downloaders, quickly remedying the situation. Besides, with the "penis size varies directly with share ratio" mentality in many BT communities, there will still be quite a few complete seeds once demand for the file builds up.
Well, that is a stupid idea. Like they don't know the outcome of this.
[%] Cingular Ringtones
My point:
Quit beating on Microsoft software for being unreliable. It reliable for those that know how to look after their machines. If you still think Windows crashes daily, you've obviously not used Windows in a very very long time, making you unqualified to comment on it. Just like any vehicle can be made to crash by the pilot, end-users can do the exact same thing to any OS if they abuse it, neglect to install OS updates and repeatedly click "yes" to install spyware (because spyware does not install itself if you are proactive about it).
Right now my Bash prompt says:I last restarted the machine before playing the Battlefield 2 demo that came out last week (needed to update my video drivers). I'm currently running (and have been for most of the week) Visual Studio, Perforce, Thunderbird, Firefox, FileZilla, Octave, iTunes, three different IM programs (I use the MSM video chat which is why I'm not running Trillian), a Cygwin Bash window, the Apache web server, a virus scanner and the application I'm developing - on a 3GHz HT P4 with buckets of RAM, all across a lovely dual 19" monitor setup. And it works great!
A full anti-virus system scan runs automatically at midnight (Symantec AntiVirus), a full anti-spyware scan runs automatically at 2am (the free Microsoft AntiSpyware) - then my hard disk is automatically defragmented at 4am (via the built-in Defrag tool). OS updates are downloaded automatically and I click "yes" on a taskbar icon to install them when I'm ready.
However, these are mostly preventative measures: I've not had a virus since the DOS days, but it's my work's policy to have a virus scanner because I connect to the corporate network. I've never had spyware either, but it should be caught if I ever get any.
The only crash I've had this week was my app while debugging it.
The last bluescreen I had was caused by hardware failure about six months ago.
Microsoft products, Windows XP in particular, can be very reliable if you look after them.
and why not.
bandwith is expensive.
computing only costs the recieving end time and power, i realt do not mind having to wait 30 minutes for a 10 minute download to becomme usefull.
on the advent of multicore cpus who would
This distributed par2 idea seems really obvious now after all.. :)
:) It'll be FTBT transfering.
Wait until they figure out how to send just qubits where you can apply a function and collapse it into the file you wanted.
>>If anything it did the popup window asking me if it was ok to access the internet.
that opens the port for the program
This is funny.
P2P is not about speed and "better behaviour".
The P2P "corporate culture" is just not compatible with Microsoft.
Welcome to the future!
Welcome to the place i'm going to drain my lizard!
VIVA! LOS BIO-DOME!
Instead of seeding the actual file(s), seed an archive that's been broken up into a multiple-files by par2 as the torrent, just like in newsgroups.
Then when you're x > 7x% done downloading all the subpar sections, just run par2 to let repair/reassemble any of the sections you don't have yet.
So what you said above, just not automatic.
Done!
I thought this company had finally grown up; I was wrong.
1.) Range of use.
2.) Compatibility
3.) Extensibility
4.) KISS
5.) Efficiency
6.)...
.
.
.
.
.
10e100.) Fast.
Ok, a bit of an exaggeration; but you get the idea.
Hey, it could have been BitAvalanche. Or Torrent Explorer. Or Microsoft Downloading Suite 2005. Or Large File Download Wizard......
20-30% faster seems easily achievable. Bittorrent is an appalling protocol when it comes to overhead, all they have to do is rip out some of the disgusting overhead in bittorrent that can easily add 20-30% to download size.
So now you don't have to wait for your torrent to complete before getting infected with Spyware......watch out for the Avalanche!
Democracy Now! - uncensored, anti-establishment news
After RTFA it sounds allot like something I designed a year or so ago. That design and my source was stolen from me by a media company who then went under. I never even got paid for my work. I really hope when it comes out it is not found MS bought this from that company, as then I would have the legal duty to sue, win, then open source my code at there expense. I'm sure even if its a derivative work I could get something out nice of them.
- d
And I, for one, welcome our new peer-to-peer overlords. May Micro$uck reign supreme over all our illegitimate downloads and guide us to the holy grail that is DRM.
I don't know if its BS. I actually read this paper last week as network coding is an area related to my field of study and it seems pretty legitimate. The paper actually claims much larger increases compared to uncoded transfers in cases such as networks made predominately of slow nodes with infrequent well-connected nodes.
The technique is actually pretty neat. They form a set of linear equations of the form:
ax_5 + bx_4 + cx_3 + dx_2 + ex_1 = g
where a,b,c,d,e,f are chosen randomly and x_n represents the data to transmit. They then send the coefficients and result a,b,c,d,e,f,g to other nodes.
With a block size of n, you typically need n sets of such coefficients (assuming they're linearly independent) to recover the original data.
This basically makes the rarest block problems of bittorrent irrelevant assuming the server has seeded a little more than n data blocks.
Like what overhead? My copy of Azureus reports less than 6% protocol overhead by bytes, although maybe you're talking about some other kind of overhead.
No, .mp3 is dying, not dead, its right there with apple, BSD, and our civil rights.. (only 1 seems to be true..)
Aaahh Hahh, I knew it. BSD *is* dead.
this nation, under God, shall have a new birth of freedom. -- Lincoln, Gettysburg Address
How can you forget that you can have file extensions larger than 3 characters in FAT32 and NTFS?
I imagine that MS wouldn't dream of releasing a technology that could upset the (RI|MP)AA. They will probably have a way to ensure that each and every download is legal*.
* Legal in the sense that you have to pay money to d/l it and MS gets a cut.
Don't buy windows!!
Really!!
MS tries to jump onto yet another bandwagon, 3 years late. (the other one being the internet :D)
smash.
I run: Windows, OS X, Linux, FreeBSD. Just because you have a hammer, doesn't mean everything is a nail.
And BTW, I believe the parent said that 1Mbps is common for cables because you said "(which I believe is very common for home cable modem users)", when in fact 1MBps is not common for home cable modem users, unless either i) they have a crappy ISP that offers only ultra-high-speed, ultra-high-cost connections, or ii) they have extra bucks hanging around that they just have to kill on higher-than-necessary connections.
Also... when you say that you know you max out at 1 megabytes/sec, I really have to wonder if you know what you are talking about. If all my downloads max out at 2000 KB/s when I am on my school's T1 backbone (granted, shared among like 1,000 students on the LAN at the same time with me), how likely is it that a home user's connection will max out at 1000 KB/s? Remember that most speed tests will also list in units of bits/sec, not bytes/sec.
Comment removed based on user account deletion
Patches for your illegal copy of longhorn?
The subject of this post is about as logical as this article's subject "Microsoft Wants P2P Avalanche to Crush BitTorrent". Come on... some researchs working for Microsoft published a paper... A PAPER!! Ridiculous.
Of course it will.
It will also keep other programs from running. It will be a trusted program, like WMA is a "trusted" format. How else will something so suck crush something so much better?
My question is where does this Reg author buy his crack? You have to smoke some really good stuff to believe the "facts" about this vaporware. Let me quote and disassemble the nonsensical notion that a program that uses more cpu and transmits more bits is somehow faster than one that does not:
Microsoft Research's approach gets around this [a missing last piece problem I've never seen getting a torrent] by re-encoding all the pieces, so that each one that is shared is actually a linear combination of all the pieces, fed into a particular function.
So it will require active CPU processing for the error correction every byte transmitted where bit torrent does not, but maxes out my download speed anyway. Brilliant! But wait, there's more:
Once you have downloaded a few of these, you can generate new combinations from the ones you have, and send those out to your peers.
I'm doing this error correction on my uploads as well. So what will this take, four times the CPU time as simply sending the files I receive?
This means no one peer can become a bottle neck, since no piece is more important than any other. It also means overall network traffic is lower, since the same information doesn't have to travel back and forth multiple times.
Say what? When has redundant error correction ever produced less size? What's this "multiple travel" problem? When I get six people sending me the same packets, that's six times the likely hood that I get it and this is what makes bittorrent work, no?
Only Microsoft can spin 20 year old error correction concepts, and announce the "best p2p application ever" before they have a working demonstration. I can hardly wait till they put it out. The same dumb asses that verbatim repeated M$ marketing, "XP is stable because it's based on NT which is like, solid," will be clogging the networks trying to share top 40 WMA files bloated to six times their natural size.
Why oh why does anyone pay any attention to what Microsoft promisses?
DMCA, Hollings, Palladium. What might have sounded like paranoia is now common sense.
You know, Microsoft isn't the only entity capable of propogating FUD -- they're just one of the most notorious. I've seen this ridiculous complaint about MS Antispyware more times than I can count, and it's not getting any less ridiculous with repetition. THERE IS NO AUTOMATIC REMOVAL OF VNC. It simply gets flagged as a potential backdoor, and the user is asked if they want it removed. This is perfectly reasonable behavior: if you don't know what VNC is or what it's doing on your system, it probably shouldn't be there. If you're running it intentionally on the other hand, you simply tell Antispyware to ignore it, and it never bothers you again.
SIERRA TANGO FOXTROT UNIFORM
Coral relies heavily on a modified DHT for indexing and clustering.
It currently serves >1TB per day from systems all over the planet
I must say that server based p2p is considerably faster. In the past this meant 421ing ftps or paying for usenet, but now with the magic of high speed high volume email....
Often when dling a file I notice five sources have 80% of the file and one source has the whole file.
BUT BT just treats the whole source as an ordinary source and just leechs data the others also have.
What it SHOULD do is get the data only that one has. Get the rare pieces first, you see.
This will prevent waiting ages for that last little bit.
Mumia Abu-Jamal is *laughably guilty*. Check the evidence.
Yeah, it's worth looking back at the proclamation of bittorrent as the source of all evil that ails M$ platforms:
"This is the marketing campaign to end all marketing campaigns," said Boyd, the Microsoft Security MVP (most valuable professional) known throughout the security industry by the "Paperghost" moniker.
How fitting! He might as well have been called "barkto". There's more vapor flying out of M$'s ass now than ever.
DMCA, Hollings, Palladium. What might have sounded like paranoia is now common sense.
Microsoft is just doing this at the request of the RIAA and MPAA so they can exploit and track use. They don't really want to make a better product. Of course I don't know what I'm talking about and just airing the thoughts that a conspiracy nut might think.
Heroscape, it's like legos combined with anachronistic wargames.
I guess that means that the only thing 20-30% faster will be Linux ISOs!
Hey, what makes you think you'll be allowed to download linux ISOs with it? Notice that:
It's meant for legal downloads only, of course.
I take this to mean that there's DRM involved. Once MS's P2P is widely accepted, we can assume that communistic things like linux won't pass the DRM test.
But we can always just use BT (which always makes me think Bacillus thuringensis).
Those who do study history are doomed to stand helplessly by while everyone else repeats it.
You have a 1 Mb per second connection, not a 1 MB per second connection. The difference is a factor of eight. You should never see down speeds faster than 125 KB/sec (which is the same as 1 Mb/sec). 1 byte is eight bits, connection speeds are measured in bits. A one megaBYTE connection is extremely fast, and, I guarantee you, you do not have one in your home.
If the file is already well compressed, generating blocks from parity information won't make it faster
I don't think you understand the smart trick here. It's got nothing to do with compression; the point is that Microsoft's technique increases the number of choices you get for candidate blocks when downloading a message.
Here's an example: suppose you want to download a file composed of 100 blocks. The blocks are distributed randomly between a number of servers, some blocks may be missing, some duplicated. Your node needs to collect all 100 packets to reconstitute the original file. Now, let's suppose 70% of the blocks may be obtained quickly and easily. 20% more are on a slow servers, and the last 10% are missing altogether. At this point, your machine can't complete the message until some server carrying the missing blocks comes online.
With Microsoft's technology, the original file is also split in 100 blocks, but extra blocks are computed as a linear combination of the original blocks. Both the original and the extra blocks are distributed. The nice trick is that your machine can reconstitute all the original message by getting any 100 of the original or extra blocks. Any 100 different packets will do! Your node will have to solve a system of linear equations (as many equations as the number of original blocks you've missed)and it will obtain the whole original message
Let's assume 100 extra blocks were created (for a total of 200 blocks) and let's also assume we have the same server quality distribution as above; 70% of the 200 blocks(140 different blocks) are available quickly and easily. At this point you don't need to wait for the slow (or missing) servers at all. You can go to the quickest servers because you don't have to wait for a certain missing packet. And you don't need to download extra bits either: you still have to get the same 100 packets and can stop the download once those have been acquired.
Nifty.
Storm
ps. I meant that spelling.
Uh, guess what? If the data is transferred 20-30% more efficiently, then your connection is still maxed out as before, but you receive the file 20-30% faster.
LRC, the best-read libertarian site on the web
let's see, I can use the free, open standard network where I can download anything that someone is willing to give me,
or i can use the closed (probably diabolicaly written) standard network that will be so wrapped up in DRM technology that only the biggest of companies can aford to use it. so, essentialy, it's just a way for large corporations to use our bandwidth tto distribute their product.
Uh, BSD doesn't count, it's dead already, how could a dead one be dying again? Undead?
"The way we can tell it's C# instead of Haskell is because it's nine lines instead of two." -- wadler
Please also realize that Windows 2000 won't be allowed to run it \o/
"The way we can tell it's C# instead of Haskell is because it's nine lines instead of two." -- wadler
Wow, do you always like to show your ignorance and gullibility at believing rumors?
/ 020318oplivingston.html
; en-us;Q316666
http://www.infoworld.com/articles/op/xml/02/03/18
http://support.microsoft.com/default.aspx?scid=kb
If you need web hosting, you could do worse than here
And what about MS' "corporate responsibility" to stop copyright infringement?
Since Microsoft has a (long) history of infringing on other people's IP why should they care unless their IP is involved?
This is basically a variation on forward error correction, which has already been used as part of P2P systems before.
This paper proposes a slightly different way of implementing FEC. Potentially, that might result in modest performance improvements compared to previous FEC methods. Unfortunately, the paper fails to measure and demonstrate any performance improvements relative to previous FEC P2P approaches, so we don't know.
Since Microsoft's technology is likely to be patented, it will not be adopted by open P2P systems. However, incorporating standard FEC methods into BitTorrent might be a good idea.
Of course, a much bigger area of improvement for content distribution would be he deployment of multicast technology across the Internet.
Funny, since I have the same package and it's not 8Mb/s.
I read the paper too! They state 2-3 times speedup over BitTorrent for badly connected networks.
Recovering the original file is tricker than it looks though...
They state that they have to invert a matrix of O(nblocks^2) to recover the original file. This takes O(nblocks^3) operations. Since there is only 1 bit per entry that will fit into memory and won't be a problem. There are plenty of ways to invert matrices faster than O(nblocks^3) too.
They then have to undo the linear equations which is an O(nblocks^2) operation. However each of those blocks is a block of the original file. If you have a 4GB file (say) broken up into 4,000 1MB blocks, you'll need to do 16,000,000 x 1 MB operations, ie 16,000,000,000,000 bytes of operations which takes a while even at L1 cache speed! If you haven't got 4GB of ram, thats going to cause an awful lot of disk IO
I guess you'd allocate the largest buffer you can, and run through the file file_size / buffer_size times. Since file_size / buffer_size probably isn't huge 10 or so (4 GB / 400 MB say), then you'll only have to do 40GB of IO to tidy up at the end. With a 40 MB/s disk that should take 15 mins or so. Not insignificant, but quicker than network IO probably!
Every man for himself, all in favour say "I"
Using a modified client I routinely receive down about 350KB/s on busy torrents and upload basically nothing but overhead.
The reason you receive better transfer speeds with BitTorrent than with eMule is that you're participating in a specialized delivery of a handful of files with a large group of people, instead of tens of thousands of files. More peers devoted exclusively to a smaller set of files for a short period of time. If you download really popular files on eMule, you'll see transfer speeds roughly parallel to that of BitTorrent. Otherwise you're competing for slots and receive poorer performance.
The biggest myth about BitTorrent is that its tit-for-tat implementation prevents leeching. Basically all I ever do is leech, and I receive better results than you do. Maybe if you weren't saturating your upstream by participating you'd receive better downloads, but it's also possible that you just pick torrents without high-bandwidth seeds.
...FEC aka PAR aka recovery blocks (and this is just another variation) are only really useful when you have an unreliable system where only some blocks will arrive. The only problem with bittorrent is usually when the combined downloads are so low, it wouldn't be possible to recreate a complete copy anyway.
Actually, it might be more attractive for trackerless torrents, where you don't have a central tracker to make sure rare parts are redistributed. It is absolutely vital for other decentralized nets such as Freenet.
Kjella
Live today, because you never know what tomorrow brings
Yeah. I actually implemented the basics of it earlier this week (just the linear encoding over a field mod 2).
Forget about the matrix multiply of an n x n matrix by an n-vector, the most significant cost by far in my implementation was the matrix inversion.
Don't suppose you know of any fast algorithms for inverting a matrix of elements in GF(2^n)??
TBH, I don't see why anybody would think MP3 is being replaced with WMA. My MP3 collection still holds >3000 mp3's and none of them will ever be wma. Also, whenever I encounter (through whichever mistake) somebody giving me a wma I convert it to ogg vorbis instantly.
Now, if you'd just do that too...
Wow. What pipe is the rest of the world Sucking on. Cause i live in australia. and for a country tahts supposed to be "early adopters" of technology. The net here SUCKS. sure i can get 256Kbps/64Kbps (Down/Up) broadband for about the same as dial up. But they give me 500Mb per month downloads.
For the real users. They have these fancy "Unlimted" plans where they give you like 10 gig (my isp) to 30 or 40 gig before they cap your speed down to like 64kbps up and down making your broadband expereince HELL. and this is STANDARD PRACTICE ive only found one plan in the whole country that doesnt have these caps for a residential plan. and its nearly $150 a month for 512/128 kbps adsl
Grow up and smell the coffee.
Some damn good places have the worst broadband possible and its not going to change fast. cause theres only 2 ways data gets out of the country and both of them are happy monopolising their market and stuffing over the people here that deserve better.
XML - A clever joke would be here if
Yeah. I actually implemented the basics of it earlier this week (just the linear encoding over a field mod 2).
Forget about the matrix multiply of an n x n matrix by an n-vector, the most significant cost by far in my implementation was the matrix inversion.
Did you do with with data size > memory? Perhaps the matrix inversion will dominate, but I doubt it as in my experience all computation is memory bandwidth limited anyway!
Don't suppose you know of any fast algorithms for inverting a matrix of elements in GF(2^n)??
Any of the fast algorithms that work over the complex field should work over GF(p) or GF(p^n). You are using GF(2) which should make it easier still. There may be specific algorithms for GF(2) I don't know!
Every man for himself, all in favour say "I"
At first glance this looked a lot crappier than it is. I'm starting to believe their 20-30% line.
If you need blocks 'a', 'b', 'c', 'd', 'e', 'f'.
You can be sent any of 2^6-1 different packets to get *some* information. Once you have any information, so long as the as a person doesn't already have that exact combined packet or a complete set of non-overlapping subsets (for example a^b^c is worthless to a person with a^b^c | a^b & c | a & b & c | a & b|c, but not to a person with a^b^d et al.) The the very first packet you have. You have information needed by a likely majority of people assuming that that packet is unique enough (w/ 2^N-1 possible packets it should be). So right out of the gate your upload speed should be maxed. With properly random xor'd packets of any number of packets combined you should be able to send that data to anybody who has not derived a perfect subset of that data.
If you get a^b^c and a^b^d you can derive packet c^d, then if you get e^f^c^d you can derive packets e^f, a^b^e^f^d, a^b^e^f^c. You now have a^b^c, a^b^d, e^f^c^d, -> c^d, a^b^e^f^d, a^b^e^f^c, e^f. So anybody can still send any of the (2^N-1 - 2^X-1; N total packets, X packets received) remaining possible packets for information. So for the last packet there are ((2^N-1) - (2^(N-1)-1) = 2^(N-1)) different possibilities. Which is vastly more than (N-(N-1) = 1) in BT proper.
You still must receive N packets, but rather than N possible candidate there are 2^N-1 possible candidates, which only gets 2^X-1 fewer for each packet.
As for the original question: In theory, you could have every every packet, save one, and still have no actual data. Eg, a^b, b^c, c^d, d^e, e^f. Or, at the upper bound every packet save one of real data. But, this is the upper bound. 90% of data is *at most* 90% of real data and at the very least 0% of real data.
As for the claim of 20-30% speed increase, I will go ahead and agree that it is feasible. This scheme adds a lot of computational overhead but it gets rid of scarcity and last packet problems. As soon as anybody has a packet they can send that packet to *almost* everybody, as not only scarcity is gone but also redundancy. Not only that, but if two people have 60% of the file done, odds are actually good that they have the entire file between them (assuming random pieces, which in theory man not be the case if they have overlap of physical data due to a person sending out the same piece to a number of people, which would occur if it is the only piece they have).
However AP2P folks could start dropping faulty data on you and, there's not a very good scheme to discover the villainy until you have N packets and can solve for the file. This would also destroy any packet you derived by using this faulty packet. You can't realistically store 2^N-1 hashes. You can't ask a seed as they might be AP2P and lie, although you could ask them something you already knew to figure it out, and processing that much data on demand would be hard for any computer (although the packet doesn't have to be sent out or derived from until verified, so there is plenty of time). So one corrupt packet in this scheme could corrupt any number of derived packets until a faulty packet is used to derive a real data packet to be hashed. And even then it wouldn't know which parent packet did it until you have N packets.
It would also be possible to withhold a specific packet (having specially crafted all other packets), without which, nobody could derive any real data. Which would might waste a good chunk of time for people caught in a AP2P honey-pot.
When they say that this is only for legal trading, they might be right. It raises a lot of red flags when it comes to catching the introduction of faulty data. Although, it might just take a scheme a little more advanced than hashing.
It is no longer uncommon to be uncommon.
The first Microsoft virus spreading via the Avalanch system shall hereby be named "Snowflake".
-
- - You can't take something off the Internet! That's like trying to take pee out of a swimming pool.
I hadn't heard that word before. Love it. I'll have to find ways of using it now.
Government of the people, by corporate executives, for corporate profits.
when there's nothing to download on it because it's all "legal"...
Richard Steven Hack - This sig is TOO GODDAMN SHORT TO DO ANYTHING USEFUL WITH! MORONS!
Ok, I just had to make sure. Even still, 1MB/s is not common amongst cable users. 1MB/s might be (and honestly this is a stretch) commonly available to cable users, but it is not a common thing. I'd say cable speeds there and above are maybe the 90th percentile of cable users.
--
WHO ATE MY BREAKFAST PANTS?
Every torrent I've ever downloaded with over about 50 peers maxed out my broadband connection, up and down, if I let it.
And any I've tried has always been slow as an evil year - probably because their system relies on fast upload as well, if you have fast download but slow upload you can stick it. Getting the new patch for World of warcraft (which they distributed via bittorrent technology) is an agonizing wait.
If Google really cared they would fix Android Chrome to reflow text, instead of discriminating
You make it a directory. Split the file up into 70 rars and make maybe 10 pars, just like you do when you post it to a newsgroup. Then torrent the directory. Once people have completed any 70 of the 80 files, they can construct the whole thing.
I am trolling
I will grant them a pecayune point in that some sort of Bittorrent-like scheme would allow their customers to get their downloads more quickly. So there is actually some benefit for their customers. But that is a community effect, and has no relation to Microsoft with the exception of them "officially" "allowing" the seeds to be set up. (Didn't some Windows users set up a torrent for XP SP2 only to have Microsoft's lawyers tell them to take it down?)
the preceding comment is my own and in no way reflects the opinion of the Joint Chiefs of Staff
Maybe not, but they're often packed with compressed files, like the Linux ISOs being filled with rpm or gz files.
I hope you aren't referring to BitTorrent, because they do NOT copy BT. The novel portion of the whitepaper describing the system is the use of network encoding, not tit-for-tat.
Get the facts... =)
-dave at limewire-
The pig browse. With Google. Sigh is to the chicken. Chicken is fool. Giggle. The DailyWTF giggle.
Singapore is a democracy only on paper. The people keep voting for the People's Action Party (which has held power continuously since the founding of Singapore as an independent state) because (a) voting is compulsory and (b) electoral districts are drawn to be so small that ones that vote for the opposition can be routinely punished by budgetary allocations. Furthermore, political activism is suppressed under Singaporean laws, ostensibly in the interest of harmony, and opposition politicians who start to look like they could have a chance often end up being sued, and losing the lawsuits.
East Germany was also technically a democracy in the same sense as Singapore; as well as the Socialist Party, there were a number of smaller parties, often sharing the names of West German parties, but deliberately kept far too weak to have any influence.
That's the problem. You still need whole files, and bittorrent blocks can span files. What can easily happend (and often does) is that no single file is complete until you get towards the very end. You have 80 partial files.
If you need web hosting, you could do worse than here
*sigh*
Dude... just... stop, ok? This has already been responded to several times before you posted this. Check elsewhere in this thread. I'm getting annoyed here.
UNIX: A computer user is defined as a programmer. WINDOWS: A computer user is defined as a consumer.
Instead of complaining to me, maybe you should complain to Optimum Online, since apparently you're getting screwed over. Perhaps your ethernet card is crappy? Optimum's website clearly says it offers speeds "up to 10,000 kbps", hence their whole "We're way faster than Verizon DSL" campaign.
UNIX: A computer user is defined as a programmer. WINDOWS: A computer user is defined as a consumer.
In Azureus at least you can prioritise early blocks so the blocks get downloaded more or less in order, so you'll finish the early files before you start the last files.
I am trolling
"For the purpose of this paper we assume symmetric links, where the download capacity is equal to the upload capacity of a node and both capacities are independent. We have experimented with asymmetric access capacities and observed very similar results and thus we omit the details of this case." That's a bit sad according to me, because "this case" as call it, is the most common case out there. Both assumptions are false : - lines are not symmetric - upload and download have an effect on each over (on dsl lines at least). The paper started great, but that's a killer assumption to me, I stopped reading there. So, Avalanche out performs Bittorrent over a LAN? cool ;)
Slurpie also claims to out performs Bittorrent, but again, over symmetrical connections ... check Slurpie if you want to check something, because according to their results, it would also outperforms Avalanche.
I am searching for document MSR-TR-2004-137 from Microsoft Research with title "Cooperative security for network coding file distribution" mentioned in original paper.
It is not accessible from this page anymore.
If someone know in which publication it can found or where it is accessible on the Internet, please let me know.
Maybe it's like that corpse god from the cthulhu mythos, eternally dieing. I forget it's name as it's been a while since I picked up Eternal Darkness.
Well I've wrestled with reality for thirty five years doctor, and I'm happy to say I finally won out over it.
From the sound of it the tech behind it is kind of interesting, I hope it's not patent encumbered so BitTorrent can implement it.
Microsoft not being over-eager patent hounds... I haven't laughed so hard in weeks. Whew. Thanks PaxTech
Your sig(k) has been stolen. There is a puff of smoke!
Anybody who doesn't think so just has to ask Microsoft's legal department. :-)
/whatever/ ...
If Microsoft then turns around and tries to sue you, you can claim prior art.
Enjoy your torrent of
MSBPodcast.com The opinions expressed here are my own. If you don't like 'em... Think up your own stuff.
Finland declared its independence in 1934. The Soviet Union though acquired it in 1924.
So it becomes 20 or 30% quicker because the client overestimates how much data is required.
There is a 100mb file that I want.
Theoretically, my client could end up downloading 130mb of blocks (worst case scenario).
You cannot create data from nothing, and so downloading only 80mb is not enough, you still need to get at least the 100mb.
This might estimate that it takes 130minutes, when the file comes in at 100minutes, the user thinks they have saved 30% of their time.
This only helps because more of the clients have a greater spread of the blocks, and so the faster clients can supply the file better.
But to me, this is a false economy, since everyone will be moving around more physical data, albeit faster.
liqbase
So the bottom line Microsoft created a front document to disinform people about bittorrrent.
Any academic advances are accidental. This is black marketing. This isn't much different from the front organisations that say linux servers cost a lot and are dying anyways.
Microsoft is pure dog-ma. FreeBSD is pure cat-ma.
Have you ever used PAR?
:)
Although the network will have to _store_ more than the original 100mb, you won't have to download more than 100mb. The second you download any combination of unique blocks amounting to 100mb, the original file will be reconstructable. You'll only have to wait for the reconstruction time, but it shouldn't be too bad.
Someone should really really really add this on to bittorrent... let's hope MS hasn't patented this. Please-please-please.
But in Australia, software is considered a product, and a product comes with a 12 month warranty (exceptions for anything more expensive than about $10k+ though). IANAL, but I work retail so I am quite familiar with said laws.
So Microsoft has 3 choices if it provides 'broken' software to Australian consumers, patch it, offer a suitable replacement or refund. Of course, swapping the media won't remove a software bug (and they aren't about to give out linux), so that leaves patch or refund. Tough choice on what they would do then, eh?
Anyway, I'm sure there would be other legal avenues you could travel down. False advertising on claims of Windows being secure perhaps? So I'd argue they are legaly obliged, even if not literally.
And why should you have to pay for them to fix their faulty product?
Not always. It's either too dumb or too smart to work with SafeTP.
That disgusting overhead of which you speak is responsible for making sure you get files that aren't corrupted or tampered-with. What's better, getting a file in 30 minutes that is corrupt and worthless or getting a file in 35 minutes that isn't?
"I assumed blithely that there were no elves out there in the darkness"
Grow the hell up!
If Google really cared they would fix Android Chrome to reflow text, instead of discriminating