Slashdot Mirror
Bram Cohen's Response to Microsoft's Avalanche
An anonymous reader writes "Bram Cohen has reduced Microsoft's proposed file-sharing application--codenamed Avalanche--to vaporware, dubbing its paper on the subject as "complete garbage". "I'd like to clarify that Avalanche is vapourware," Cohen said. "It isn't a product which you can use or test with, it's a bunch of proposed algorithms. There isn't even a fleshed-out network protocol. The 'experiments' they've done are simulations.""
Sounds like it, and the first salvos have gone back and forth... having read both, I have to give the points in the first round to Bram. Microsoft won't find him so easy to push around, methinks.
I guess microsoft is just doing research, so they can patent their inventions. Those patents can than be used to make (other) fileswapping/p2p programs illegal due to patent infringement.
They are marketing their IDEA, not the actual software at this point. Sure, what they have done is research and simulations, which is obviously just one part of their software development cycle.
Microsoft has a huge amount of resources that they can and probably will pour into the p2p projects they are working on. It is foolish to mouth off and bash their development procedure, treating it as something other than it is. Microsoft has a strong track record of eliminating its competition by integrating products into its OS. Dont be too suprised if you see Avalanche as part of Longhorn.
"If we hit that bullseye, the rest of the dominos will fall like a house of cards. Checkmate." -Zapp Brannigan
BT is relatively new, I am sure within a few years some serious inadequacies will be found which will make this research from Microsoft more significant.
BT is NOT relatively new - in fact, it's relatively old, and there HAVE been a few years for any "serious inadequacies" to surface. What has happened in those years is that users of other P2P networks have flocked to BT by the millions, simply because it works much better at delivering maximal bandwidth for highly sought-after files.
The illegal we do immediately. The unconstitutional takes a little longer.
--Henry Kissinger
Most software isn't spouted off as the best thing since sliced bread when it's at this stage. Then again, this IS Microsoft we're talking about (not a dig at Microsoft's coding abilities, but they're PR department).
Microsoft, on numerous occasions, has indicated that they have a(n) [ insert competitors product ] killer... with a cool code name and features that look very appealing. We find out months/years later that their product either doesn't materialize or doesn't deliver on the original specifications. Sure the 2.0 or 3.0 version might, but my point is, they fend off competition by using vaporware.
I'm tired of it... I'm moving back to my TRS-80
Longhorn is just one prime example. I wonder how many people didn't consider switching over to Linux/?nix/OS X/etc. etc. because of the overly hyped features of Longhorn... which now are disappearing left and right.
It takes years to make something like bittorrent, but it takes days for a marketing team to come up with a flashy code name and feature list.
Good. The mud has officially been slung. We are in for a hell of a fight, it seems.
The "Avalance is vaporware" vibe is a true one, but let's give Microsoft a chance for a real-world test before we cast our lots. Not completely dismissing the paper demonstrates Cohen in a more rational and less infuriated moment, and is fortunate that he did so, as industry leaders who dismiss competition get burned all too often. This is not to defend the test model in the slightest, which is junk and atypical of typical Bittorent usage as Cohen rightly points out.
The Avalanche paper is a start. Microsoft will need to finish, refine, and check their facts about the product with which they are competing. The idea of building a file without all the pieces reeks of difficult implementation, for example.... that's one protocol I would love to see come into reality. Bittorent will need to flex and build upon the established track record of the protocol, and innovate on top of that. Decentralized trackers were a good step.
The Crimson Dragon
So you think he's bashing them? Having read Bram's comments, what he seems to be responding to is the way (he says) they misunderstood and misrepresented BT; which strikes me as a quite legitimate response.
Research papers that are released, even by MS, usually aren't intended to get the attention of the broader market. They're intended to put forth ideas and let a few select people know what's going on inside the heads of MS engineers. That, and to raise the ire of slashdotters who don't read them and choose to automatically assume that MS persues every single research idea as a major project.
Look at it this way - MS can't afford not to be looking into the area of filesharing, because it's obviously something that their customers really, really want. There hasn't been any announcement of any product, there's just a whitepaper with no details. The not-so-sinister truth is that this research paper is just evidence that they are starting to think about the problem, not a representation of an imminent product offering.
-- Truth goes out the door when rumor comes innuendo. -- Groucho Marx
you know, Cohen was referencing the algorithms that microsoft was using to design their p2p system. If the algorithms are faulty I don't see how the derived software is going to turn out much better. He was also commenting on their misconceptions of how bittorrent operates, and that their idea of how it worked was ignorant at best. As the author of bittorrent I think he has the grounds to say what he said, he wasn't just mouthing off.
twitter.com/gravitronic
> Most software isn't spouted off as the best thing since sliced bread when it's at this stage
Who's the one who spouted off? This was a paper on the MS research website, not an ad on prime-time TV. It's loudmouths like the ones ACTUALLY doing the spouting that will cause MS to just replace that page with a static placeholder and reveal NOTHING to the outside. It's already gutted, it really won't take much more to turn it into a complete facade. Thanks a lot.
I am no longer wasting my time with slashdot
That's exactly the point. The paper is based on _simulations_; Bram hasn't found any good reliable simulations for bittorrent, so he has little faith in this.
Vaporware is technically correct: you can't download and use Avalanche. But you may be able to in a year or two. Hopefully, they'll make it useable by then.
Here's the thing, they're using a 'tit-for-tat' algorithm that was in bittorrent v1, 4 years ago. Which makes me believe that they are currently 4 years behind BT. They do have the advantage of following, so they can catch up faster than Bram's original work, but this is still just ideas.
I must say, I too don't see the point of error correcting codes, I mean, you have to transmit them too. You're substituting data for other data. And instead of tring to calculate all of what you need, Bittorrent will save you the CPU and HD cycles and just wait and find the original, instead of trying to build it. This might work fine if you have 2 processors and 4 gigs of ram, but I'll stick with bittorrent until Avalanche is a proven product. Even then, it will probably still not be cross-platform...
I couple years this may be better than BT(today's) in pure network speed, but then again, BT will likely be faster by then as well. Right now its just academic.
Bram may be right about Microsoft's paper, but he would have had more credibility if he had taken the high road.
Quotes like "The lack of any concrete numbers at all shows the typical academic hand-wavy 'our asymptotic is good, we don't need to worry about reality' approach" certainly don't earn him much respect from academics in system programming research who work very hard, thankyou very much, to ensure that their results are realistic. He has turned a simple observation about the paper (they neglected certain overheads) into a bigoted rant (academics are foolish). Not cool.
Patrick Doyle
I mod down every jackass who puts his moderation policy in his sig. Oh, wait a sec....
a) new features
b) bug patches
Just because they keep releasing new versions doesn't mean it's not 'finished'.
I think most people, including the parent, who say Windows isn't finished are eluding to the fact that it's released in an unstable, insecure, and generally half-assed condition.
If a product is released and a year later a new feature is added to that same product, does it mean the previous product went unfinished for a whole year? Not really. Why do you think they use version names? Mac OSX 10.1 is a finished product - when changes for 10.1 are released, it's under a new version number representing a newer finished product.
Distributions of Linux, and the kernel itself, have updated releases on a much more frequent basis. But that's why there are production (or stable) and testing (or unstable) branches. The production version is a finished product.
Arguably you could still say that all the aforementioned software is never finished, but then the same could be said for a lot of things. Car models are updated on a yearly basis - does that mean the previous year's model was not finshed? No.
At some point a product which is periodically updated must be defined as 'finished' and separated from development leading to the next version of the finished product.
As I mentioned, Microsoft never seems to release a 'finished' version of Windows because it's in a perpetual state of half-assedness. Or like Longhorn, the release date is constantly being pushed back and it appears as though it'll never be finished.
-kidlinux.
MS research is messing with all sorts of interesting ideas. They've hired a number of gurus in computer science research (such as Tony Hoare and Leslie Lamport). They publish lots of papers. How many of these things will turn into real products? Who knows. Mostly they just want to play with ideas so that they stay at the cutting edge of things, rather than missing the boat as they did with the Internet boom.
> "Vaporware is technically correct: you can't download and use Avalanche. But you may be able to in a year or two."
You'll probably be saying that again two years from now. Anybody remember the debacle Microsoft had when Gates said that MS was working on a 64-bit operating system that, according to him, would be available a year after he said that? It was nearly five years before it finally happened.
Avalanche? That's actually what Microsoft will end up buried in, only it won't be snow, it might just be trash in a landfill.