Microsoft Reinvents Bittorrent

Anon E. Muss writes "Microsoft has a new Secure Content Downloader tool that sounds an awful lot like a Bittorrent clone. It's described as a 'peer-assisted technology' where '[e]ach client downloads content by exchanging parts of the file they're interested in with other clients, in addition to downloading parts from the server.' Right now MSCD is just a time-limited preview, intended to support downloads of select Microsoft beta releases (e.g. Visual Studio 2008). If this test goes well, Microsoft will probably start using MSCD for all their large downloads. How do you feel about subsidizing Microsoft's bandwidth costs?"

It's not Bittorrent. It's better.

[SilentChris]

MS didn't reinvent Bittorrent. They built something better: Avalanche [microsoft.com]. It's more efficient and (I know this phrase is weird to use around MS, but...) more secure. Read the research papers (they touch on BT, its advantages and disadvantages). I imagine most of this stuff is on its way into standard BT, if it hasn't been worked in already.

"How do you feel about subsidizing Microsoft's bandwidth costs?"

Frankly I don't give 2 shits as long as they don't patent the hell out of it (and sue existing P2P solutions). But this came out of MS Research, so I doubt that'll happen (one of the only decent groups at MS).

By the way, MS has been messing around with P2P for years. How do you think Xbox Live works? Every time a game is played multiplayer, at least one Xbox/Xbox 360 is hosting. Not a single MS server hosts a game. Question this all you want (why pay $60 a year then?) but the fact of the matter is that from a technological standpoint, it works well.

Re:bllizard, wow patcher

[JimDaGeek]

How is it a double standard if someone doesn't want to support Microsoft while wanting support a company they like, such as Blizzard? If MS were a better company with better practices, supported standards better and didn't abuse their monopoly position, I am sure there would be a lot more supporters on the side of Microsoft.

Me personally, I won't give any of my bandwidth to Microsoft. Let them pay for it. Now if Microsoft wanted to pay me to use my bandwidth, I would consider that option.

Re:Good for them

[Meccanica]

Exactly what I thought- therefore, why not just rebrand bittorrent clients as MSCD clients? Everyone wins!

Although, the RIAA/MPAA will still claim to be losing.

A Brilliant Plan occurs to me!

1. If all or most of current bittorrent networks could be 'changed' into 'MSCD' networks

2. Upload a bunch of fake 'torrents' using the 'old' technology as a trap (a reversal of the very same technique that the RIAA types have tried using).

3. Hammer them with legal action and bad PR over attempting to obtain + distribute child pornography or something horrible like that

4. ?????

5. Profit

Far be it for me to disagree with Microsoft.

[khasim]

From that article that you linked:

Peers do not need to find specific pieces in the system to complete, any subset encoded piece will suffice.

Huh?

Also, no peer becomes a bottleneck, since no block is more important than another.

In bittorrent, no block is more important than any other.

And the only bottleneck in bittorrent is when a specific block is only available from a single seed with limited bandwidth. The moment that block is uploaded to another machine the bandwidth expands.

Finally, network bandwidth is efficiently utilized since the same information does not travel multiple times over bottleneck links.

I'm not understanding that either. You need updates as to who has what. This will be changing constantly as different peers download different blocks.

One possible solution is to use a heuristic that prioritizes exchanges of "locally rarest" pieces. But such local-rarest policies often fail to identify the "globally rarest" piece when peers have a limited view of the network.

Why would you need to? All the client has to do is connect to as many peers as necessary to find each block a minimum number of times. The only time there is a problem with this is when there is only one seed with limited bandwidth.

There is no way that a "globally rarest" will appear more often in your peer group than it does globally. This seems more of a seeder issue than a swarm issue. And it has been solved with the "super-seeder" enhancements. The seeder feeds more blocks to the guy who seems to share them the fastest.

Re:Better download integrity, yes please.

[Dr_Barnowl]

Downloaded it onto three different drives (one of them a flash drive) mounted in two different machines, all of which are showing no signs of disk wear. Also downloaded it across two networks, one belonging to a national government infrastructure, one of them being my ISP at home.

Each file showed corruption throughout the file, each file had a different, incorrect, MD5 hash - I actually went so far as to write a "chunkhash" util to hash chunks of the file to see if I could construct a single "good" file from the 9 corrupt ones. After reviewing the output I decided it was hopeless - there just weren't enough blocks where the hashes matched on more than one copy of the file to stick it together.

Plus the actual confirmation that there was a problem through a mutual friend at MS kinda gave it away.

Re:bllizard, wow patcher

[NeilTheStupidHead]

0... ^>^ I actually have a huge problem with Blizzard's distribution system for patches. My ISP shapes their traffic and it can take hours for a small four megabyte patch to download. If I go directly to their site and download as a standalone file: about a minute. A distributed download system is a good idea both for Blizzard as it saves them bandwidth and for most of their customers as they get their patches faster (especially when it comes to large patches), but the standard download model has to be available for those who cannot use this type of system.

not a "troll" at all

[oohshiny]

Microsoft is charging a lot of money for their software; there is absolutely no reason anybody but Microsoft should pay for the bandwidth related to their software updates.

From a practical point of view, no matter how "secure" the protocol may be, if this thing is running on a host as part of a P2P network, it is essentially broadcasting to the world that (1) the host is running Windows, and (2) that it's not up to date with its patches. That's not a smart thing to broadcast.

Re:bllizard, wow patcher

[the not-troll]

As I am not playing WoW, and thus don't patch it, I have no double standard. Thus, let me say: As I use Linux, I won't use this program, so I won't subsidize Microsofts bandwith. So the question whether I'd have a problem with subsidizing Microsofts bandwidth doesn't make any sense in the first place. But if I were using it, I doubt I'd have a problem with that - I'd have bigger ones. After all, I'd be using Microsoft software.

Just wanting to steer against any perceptions of the unsuspecting reader that the /.-community might be homogeneous in any way. Oh, wait...

Re:bllizard, wow patcher

[mrsteveman1]

The real problem is that MS intends to turn this into a system totally encumbered by DRM.

Another expected build in, Microsoft will probably implement a way for "content owners" to remotely delete the metafile and all data if they so choose, regardless of how valid their claim is. I also fully expect traffic shaping to ignore this new protocol while throttling bittorrent.

Re:bllizard, wow patcher

[_Sprocket_]

Microsoft does bad things: 5% of Slashdot articles
Microsoft does perfectly innocent things, but Slashdot declares them bad: 95% of Slashdot articles.

Made up figures don't mean anything. Having said that... it always seems to be a judgment call. There are plenty of screwy things that Microsoft does and always a contingent claiming double-standards, religious zeal, reality of business, ignorant-basement-dweller, and whatever other non-argument they can to justify or detract from the issue.

I don't buy your figures. But I do agree that there are certainly times when articles or comments are beyond the pale. Microsoft does occasionally get skewered over non-issues. I'm 100% behind calling those out. They detract from the real issues.

Which issues are "real" is probably the point where we would disagree.

Of course Microsoft does "bad things." The problem here is that, on Slashdot, the term "bad things" is basically defined as "Microsoft does it." It's self-fulfilling. Hell, when Microsoft gave a free 3-year warranty on Xbox 360s, somehow that was construed as a "bad thing" on Slashdot...

Great example. You say "free 3-year warranty on Xbox 360." The critics noted design flaws, a history of denying said flaws, and said "damage control." Is this one of your 95%?

The bashing here is entirely out of control. It makes the Steve Jobs Reality Distortion Field look tiny in comparison. All you need to do is type "embrace and extend" or "FUD" and you get an instant +5 insightful.

Sometimes. The bashing does need some sanity checking. However, it's not as simplistic as you claim.

By the way - cute use of colorful terminology while decrying other's over-use of catch phrases. Reality distortion field indeed.

Re:bllizard, wow patcher

[Ph33r th3 g(O)at]

Good, then P2P networks using that protocol can spring up and not be blocked by ISPs.