Re:Gnutella2 - The real story!
on
Gnutella2?
·
· Score: 1
The point is that Gnutella is based on open standards. Open standards, however, rely on those standards going through a "standardization process." The w3c exists for a very good reason. The GDF exists for a very good reason.
This is not a personal issue. I like Mike and think Mike has done a great job with Shareaza, and I applaud him for it.
Re:Gnutella2 - The real story!
on
Gnutella2?
·
· Score: 3, Interesting
I agree with you to some degree -- the key is coming out with the best technology for everyone in the end. It's just that open standards are typically what make this possible. This is as much of a social issue as anything else -- the Gnutella world is a pretty tightly-knit group that relies on trust to a large degree, but maybe it needed a little shake-up.
As far as Gnutella vs. other networks, this is really the crux of the issue. Gnutella has lagged behind in some ways precisely because it is open -- it just takes forever for people to agree. I would argue, though, that it's worth it. Why? Because you come up with better standards in the end. Gnutella is the one network that has a public set of RFC-style specifications precisely outlining how the various protocols should work. This takes time, but it allows interoperability.
eDonkey (and Overnet) are great counter-examples. They work very well, but they are proprietary standards for the most part. This has meant things like the eDonkey web URIs not really being standards compliant. But, then again, they're everywhere, so does it matter? Maybe not.
It's my belief that open standards win in the end because they allow unforeseen innovation and creativity to be built on top of them.
Re:Gnutella2 - The real story!
on
Gnutella2?
·
· Score: 1
Entering the inevitably losing battle of discussing bundled sofware, LimeWire bundles one piece of software on Windows - TopMoxie. This is far less than Kazaa, Morpheus, Bearshare, etc. The install is optional during LimeWire installation. That's about the last I'll say on the topic. Believe me, the technology itself is just so much more interesting. Is there any chance we can focus on that?
Re:Gnutella2 - The real story!
on
Gnutella2?
·
· Score: 4, Interesting
I think you're right -- our toes were stepped on a little =) -- ours and everyone else's in the Gnutella world who put in most of the work to come up with "Gnutella 2," many of them open source programmers who donated their time to the effort.
Mike has done a hell of a job on his client and is a very nice guy, but he simply is not the originator of the vast majority of the standards being branded as "Gnutella 2."
The key word in your last paragraph is "unfortunately." Yes, it was unfortunate that IE created it's own standards and bypassed the w3c. Are you truly advocating proprietary standards over open standards? Am I misinterpretting you?
Re:Crossing fingers
on
Gnutella2?
·
· Score: 2, Informative
Almost all of these protocols are existing standards that have been developed by the rest of the Gnutella community. In face, not one of the primary protocols has actually been developed by Shareaza -- they're basically just repackaging the existing collection of newer protocols and calling them "Gnutella 2."
The only new additions proposed here are the binary tree structure for packets and the node addressing system. Otherwise, all of these protocol have been developed by other members of the Gnutella community. "Gnutella 2" is a marketting gimmick aimed at doing things like getting on Slashdot.
That said, it's not all bad. The perception of Gnutella should change, as the network is continuing to develop rapidly, with powerful protocols including the Hash/URN Gnutella Extension (HUGE), the Gnutella UDP Extension for Scalable Searches (GUESS), and the Ultrapeer proposal.
Perhaps the more important issue at hand, however, is whether or not Gnutella will remain an open, interoperable protocol, or whether it will disintegrate into proprietary schemes. As yet, none of the new parts of Gnutella 2 have been posted in public specifications. This is really a first for Gnutella -- the Gnutella community works because standards are published publicly and go through a review process among all Gnutella developers. Gnutella 2 may bode ill for the future of Gnutella as an open network, but I really I hope not. I hope that Mike (Shareaza) will quickly publish any new specifications that he has to alleviate the fears of myself and everyone else in the Gnutella world!
Re:UDP and firewalls
on
Gnutella2?
·
· Score: 5, Interesting
Firewalls block most incoming UDP traffic in the same way that they block incoming TCP traffic -- there's really no difference. Incoming traffic is generally denied except for specific ports.
So, with both UDP and TCP, only outgoing data will not be blocked as a general rule. With TCP, this poses less of a challenge because once you've established a connection, data can be passed both ways. With UDP, you cannot establish a connection in the same way. That said, most firewally will allow incoming UDP from a specific endpoint if you've sent outgoing data to that endpoit "recently." In this way, a quasi-connection can be established.
All that aside, though, the short answer is that non-firewalled hosts, and specifically "Ultrapeers" on Gnutella, act as proxies for firewalled hosts, allowing firewalled hosts to behave on the network almost exactly like hosts without firewalls.
Re:Gnutella2 - The real story!
on
Gnutella2?
·
· Score: 5, Interesting
Seconding Susheel's comments, "Gnutella 2" appears to be primarily a marketting gimmick. Gnutella 2 is really just a collection of protocols, most of which have been in use on Gnutella for some time. The one apparently new protocol is a version of the Gnutella UDP Extension for Scalable Searches (GUESS) open standard, that was proposed by LimeWire some time ago, as Susheel mentioned, and that is in experimental stages.
That said, perhaps "Gnutella 2" makes some sense as a name, as the computing community seems to be out of touch with how rapidly developments are happening on Gnutella. The collection of protocols used on Gnutella today make it a vastly different network than what people typically think of as Gnutella. If Gnutella 2 changes that perception, then it's great. Just keep in mind that "Gnutella 2" has little to nothing to do with Shareaza -- they primarily contributed the name. The new protocols in use on Gnutella are the result of countless hours of work from many Gnutella developers around the world.
"Weaselmancer," or maybe I should just call you "wheeze"? Yeah, I like that better. I was not selling my software. I was answering the question as simply and as straighforwardly as you possibly can. Glad to know you would have modded me as a troll too. Good to know slashdot has such impartial and responsible mods.
On the whole "trojan" idea, TopMoxie is by no means a trojan, first. Second, its aim is not to take revenue from webmasters -- we've been trying for many months to get them to fix the affiliate link issues, but they have yet to do so.
Nice work on the whole douchebag thing again -- that's spectacular. Again, thanks for coming.
Like I mentioned, I don't fully understand the whole moding blah blah blah system -- I don't read slashdot particularly frequently because it so often degenerates into your "douchebag" threads like this one. I just don't appreciate being marked as a troll for a completely innocuous post that in fact was helpful and answered the question in a very straightforward manner.
Then again, I'm sure you'll just reply with another mindless "douchebag" response, so this isn't really worth my time. Thanks for coming.
Wow. Perhaps a review of this whole modding system is in order, eh? The importance of communicating articulatly should not be underestimated. I'm not at all sure where "charities" came into the picture. TopMoxie is primarily used by schools and charities, in fact, to earn revenue for them. Anyway, this is clearly a losing battle in the face of mindless insults, so carry on.
We put TopMoxie in the free Windows version of LimeWire. There's a page included in the installer notifying the user of this, in addition to the license agreement.
OK, to be honest there "Weaselmancer," I don't read Slashdot particularly frequently, so I'm not really familiar with moding and trolling and everything else. It just doesn't make sense to mark my post as a troll when it has nothing to do with trolling. In fact, I was simply and clearly answering the first posters question. I'm astounded that you even are granted the authority to mark posts as trolls when you call people "douchebag" in your posts. That's the type of post that makes me not read Slashdot -- it dilutes otherwise interesting dialogue with mindless drivel and name calling.
Thanks for the little troll mark and for diverting the conversation from what is an important topic.
The key that you're missing is that Gnutella is an open network. Kazaa is a closed system targetted primarily at downloading media files, most of which are copyright infringing. Gnutella actually is an open network that is becoming a platform for doing far more than simple file-sharing. Gnutella is starting to be used for distributed processing and for cheap, CDN-style broadcasting.
What's more, the open nature of Gnutella will soon push it far beyond the functionality of other peer to peer systems. Why? Because Gnutella developers from around the world cooperate to create interoperable standards that are generic and powerful. Don't pigeonhole Gnutella -- it's a far more dynamic, creative, and evolving community than you're giving it credit for.
As I've mentioned on a couple of other threads, no Linux versions of LimeWire have any bundled software. Only the free Windows version has TopMoxie bundled.
LimeWire Pro contains no ads or bundled software. All non-Windows versions (free or pro) contain no bundled software. So, there's no software bundled with any Linux version of LimeWire.
The most advanced Gnutella clients are almost identical to the FastTrack system. FastTrack was coded off of the Gnutella model. It's basically just a pumped up old Gnutella, and that's exactly what the newer Gnutella clients are.
Once again, folks, the Gnutella protocol is not a static entity. At this point it's really an evolving set of the original protocol and extensions.
In short, Gnutella is where the action is. It's an open protocol with a great deal of development among the client developers as well as the academic community. Many of the clients are also open source, providing a rich overall platform for innovation. The original protocol with all of the extensions has also become quite complex. The design, I assure you, no longer "sucks."
If you're interested, check out the center for active Gnutella development at:
http://groups.yahoo.com/group/the_gdf/
There's nothing that makes it inherently any slower than FastTrack. LimeWire's UltraPeers use compressed messages to communicate with leaves, meaning that 56K users probably devote less bandwidth to messaging traffic than they do on FastTrack, i.e. they have more bandwidth left over for downloads and uploads. Bearshare will also soon have this feature.
As usual, there's a great deal of misinformation floating around regarding the file sharing world. Saying that Gnutella can't scale is really ridiculous at this point. Gnutella is a protocol, and now really a set of protocols, that is evolving at an extremely rapid rate. It's scalability is continually improving, and the most advanced Gnutella clients are able to do just about everything the FastTrack clients do and more. Where FastTrack wins out now, the openness of the Gnutella standard will likely leave FastTrack behind fairly shortly. The two networks are extremely similar at this point. In short, if Gnutella doesn't scale, niether does FastTrack. This, of course, is to some extent true. No system can ever be fully searchable if it reaches, say, a trillion hosts.
I assume that by design, you are referring to the underlying architecture of these systems? AudioGalaxy is basically Napster. It's a centralized system that avoids lawsuits by occasionally filtering their songs when things are looking ominous and with some people at its head that have connections with the record industry.
As far as security goes, it is true that the Gnutella protocol does not include encryption, whereas FastTrack encrypts many of its packets. Just running netstat will show you everyone's IP on either network.
I strongly encourage everyone to support Gnutella as an open standard. As Lawrence Lessig has eloquently argued in "The Future of Ideas," open standards are vital to encouraging innovation and for creating resources that can be enjoyed by all, not the few.
I think your research (and dissertation) is a bit out of date. The architecture of LimeWire's "UltraPeers" is very similar to the architecture used on the FastTrack network. In fact, UltraPeer's use of intelligent query routing makes them almost assuredly more efficient than the similarly designed FastTrack implementation, although it's really difficult to know because FastTrack is not open source and its messages are highly encrypted.
All that aside, FastTrack has not changed at all over the past 8 months, whereas Gnutella has made huge strides. What's more, many of the clients (like LimeWire and Gnucleus) are open source, and Gnutella is an open protocol. We need open standards like this to keep the Internet free (as in free speech).
LimeWire currently implements a variation of this -- what we call "UltraPeers." UltraPeers establish a significantly greater horizon on the network, and there are other distributed protocols that do this in other creative ways, such as Chord out of MIT, which can be found at:
http://www.pdos.lcs.mit.edu/chord/
That aside, there is significant evidence to show that a distributed network can scale far better than any centralized network. Remember that Napster had serious scaling problems as well -- you could only see the files from the hosts on whichever server you happened to be logged in to. The only solution to that problem is the brain-dead purchase of a yet faster multi-million dollar server. I would not call that scaling.
As everyone else has pointed out, this discussion began and ended in the Gnutella community about a year ago.
Apparently you have not looked at the LimeWire source. If you had, you would know that LimeWire was coded by a team, and that the LimeWire source is quite well engineered and well documented. I would recommend giving it a look at www.limewire.org.
If you find specific performance bottlenecks, please feel free to e-mail me anytime at afisk@limewire.org. (think profiler). Thanks.
Adam Fisk
LimeWire Team
Slashdot Mirror
The point is that Gnutella is based on open standards. Open standards, however, rely on those standards going through a "standardization process." The w3c exists for a very good reason. The GDF exists for a very good reason.
This is not a personal issue. I like Mike and think Mike has done a great job with Shareaza, and I applaud him for it.
I agree with you to some degree -- the key is coming out with the best technology for everyone in the end. It's just that open standards are typically what make this possible. This is as much of a social issue as anything else -- the Gnutella world is a pretty tightly-knit group that relies on trust to a large degree, but maybe it needed a little shake-up.
As far as Gnutella vs. other networks, this is really the crux of the issue. Gnutella has lagged behind in some ways precisely because it is open -- it just takes forever for people to agree. I would argue, though, that it's worth it. Why? Because you come up with better standards in the end. Gnutella is the one network that has a public set of RFC-style specifications precisely outlining how the various protocols should work. This takes time, but it allows interoperability.
eDonkey (and Overnet) are great counter-examples. They work very well, but they are proprietary standards for the most part. This has meant things like the eDonkey web URIs not really being standards compliant. But, then again, they're everywhere, so does it matter? Maybe not.
It's my belief that open standards win in the end because they allow unforeseen innovation and creativity to be built on top of them.
Entering the inevitably losing battle of discussing bundled sofware, LimeWire bundles one piece of software on Windows - TopMoxie. This is far less than Kazaa, Morpheus, Bearshare, etc. The install is optional during LimeWire installation. That's about the last I'll say on the topic. Believe me, the technology itself is just so much more interesting. Is there any chance we can focus on that?
I think you're right -- our toes were stepped on a little =) -- ours and everyone else's in the Gnutella world who put in most of the work to come up with "Gnutella 2," many of them open source programmers who donated their time to the effort.
Mike has done a hell of a job on his client and is a very nice guy, but he simply is not the originator of the vast majority of the standards being branded as "Gnutella 2."
The key word in your last paragraph is "unfortunately." Yes, it was unfortunate that IE created it's own standards and bypassed the w3c. Are you truly advocating proprietary standards over open standards? Am I misinterpretting you?
Almost all of these protocols are existing standards that have been developed by the rest of the Gnutella community. In face, not one of the primary protocols has actually been developed by Shareaza -- they're basically just repackaging the existing collection of newer protocols and calling them "Gnutella 2."
The only new additions proposed here are the binary tree structure for packets and the node addressing system. Otherwise, all of these protocol have been developed by other members of the Gnutella community. "Gnutella 2" is a marketting gimmick aimed at doing things like getting on Slashdot.
That said, it's not all bad. The perception of Gnutella should change, as the network is continuing to develop rapidly, with powerful protocols including the Hash/URN Gnutella Extension (HUGE), the Gnutella UDP Extension for Scalable Searches (GUESS), and the Ultrapeer proposal.
Perhaps the more important issue at hand, however, is whether or not Gnutella will remain an open, interoperable protocol, or whether it will disintegrate into proprietary schemes. As yet, none of the new parts of Gnutella 2 have been posted in public specifications. This is really a first for Gnutella -- the Gnutella community works because standards are published publicly and go through a review process among all Gnutella developers. Gnutella 2 may bode ill for the future of Gnutella as an open network, but I really I hope not. I hope that Mike (Shareaza) will quickly publish any new specifications that he has to alleviate the fears of myself and everyone else in the Gnutella world!
Firewalls block most incoming UDP traffic in the same way that they block incoming TCP traffic -- there's really no difference. Incoming traffic is generally denied except for specific ports.
So, with both UDP and TCP, only outgoing data will not be blocked as a general rule. With TCP, this poses less of a challenge because once you've established a connection, data can be passed both ways. With UDP, you cannot establish a connection in the same way. That said, most firewally will allow incoming UDP from a specific endpoint if you've sent outgoing data to that endpoit "recently." In this way, a quasi-connection can be established.
All that aside, though, the short answer is that non-firewalled hosts, and specifically "Ultrapeers" on Gnutella, act as proxies for firewalled hosts, allowing firewalled hosts to behave on the network almost exactly like hosts without firewalls.
Seconding Susheel's comments, "Gnutella 2" appears to be primarily a marketting gimmick. Gnutella 2 is really just a collection of protocols, most of which have been in use on Gnutella for some time. The one apparently new protocol is a version of the Gnutella UDP Extension for Scalable Searches (GUESS) open standard, that was proposed by LimeWire some time ago, as Susheel mentioned, and that is in experimental stages. That said, perhaps "Gnutella 2" makes some sense as a name, as the computing community seems to be out of touch with how rapidly developments are happening on Gnutella. The collection of protocols used on Gnutella today make it a vastly different network than what people typically think of as Gnutella. If Gnutella 2 changes that perception, then it's great. Just keep in mind that "Gnutella 2" has little to nothing to do with Shareaza -- they primarily contributed the name. The new protocols in use on Gnutella are the result of countless hours of work from many Gnutella developers around the world.
"Weaselmancer," or maybe I should just call you "wheeze"? Yeah, I like that better. I was not selling my software. I was answering the question as simply and as straighforwardly as you possibly can. Glad to know you would have modded me as a troll too. Good to know slashdot has such impartial and responsible mods.
On the whole "trojan" idea, TopMoxie is by no means a trojan, first. Second, its aim is not to take revenue from webmasters -- we've been trying for many months to get them to fix the affiliate link issues, but they have yet to do so.
Nice work on the whole douchebag thing again -- that's spectacular. Again, thanks for coming.
Like I mentioned, I don't fully understand the whole moding blah blah blah system -- I don't read slashdot particularly frequently because it so often degenerates into your "douchebag" threads like this one. I just don't appreciate being marked as a troll for a completely innocuous post that in fact was helpful and answered the question in a very straightforward manner.
Then again, I'm sure you'll just reply with another mindless "douchebag" response, so this isn't really worth my time. Thanks for coming.
Wow. Perhaps a review of this whole modding system is in order, eh? The importance of communicating articulatly should not be underestimated. I'm not at all sure where "charities" came into the picture. TopMoxie is primarily used by schools and charities, in fact, to earn revenue for them. Anyway, this is clearly a losing battle in the face of mindless insults, so carry on.
We put TopMoxie in the free Windows version of LimeWire. There's a page included in the installer notifying the user of this, in addition to the license agreement.
Thanks.
Adam Fisk
LimeWire LLC
OK, to be honest there "Weaselmancer," I don't read Slashdot particularly frequently, so I'm not really familiar with moding and trolling and everything else. It just doesn't make sense to mark my post as a troll when it has nothing to do with trolling. In fact, I was simply and clearly answering the first posters question. I'm astounded that you even are granted the authority to mark posts as trolls when you call people "douchebag" in your posts. That's the type of post that makes me not read Slashdot -- it dilutes otherwise interesting dialogue with mindless drivel and name calling.
Thanks for the little troll mark and for diverting the conversation from what is an important topic.
The key that you're missing is that Gnutella is an open network. Kazaa is a closed system targetted primarily at downloading media files, most of which are copyright infringing. Gnutella actually is an open network that is becoming a platform for doing far more than simple file-sharing. Gnutella is starting to be used for distributed processing and for cheap, CDN-style broadcasting.
What's more, the open nature of Gnutella will soon push it far beyond the functionality of other peer to peer systems. Why? Because Gnutella developers from around the world cooperate to create interoperable standards that are generic and powerful. Don't pigeonhole Gnutella -- it's a far more dynamic, creative, and evolving community than you're giving it credit for.
Adam Fisk
LimeWire LLC
As I've mentioned on a couple of other threads, no Linux versions of LimeWire have any bundled software. Only the free Windows version has TopMoxie bundled.
Adam Fisk
LimeWire LLC
Thanks for moding me down for "trolling." Shouldn't you actually know what a troll is before being able to mod posts down like that? Rough one...
Adam Fisk
LimeWire LLC
LimeWire Pro contains no ads or bundled software. All non-Windows versions (free or pro) contain no bundled software. So, there's no software bundled with any Linux version of LimeWire.
Adam Fisk
LimeWire LLC
Only the Windows version of LimeWire has any bundled software.
Adam Fisk
LimeWire
LimeWire Pro has no bundled software or ads. In addition, all non-Windows versions of LimeWire have no bundled software.
Thanks.
Adam Fisk
LimeWire
LimeWire on Linux does nothing like this -- there is no bundled software for any non-Windows version of LimeWire.
Adam Fisk
LimeWire
Once again, folks, the Gnutella protocol is not a static entity. At this point it's really an evolving set of the original protocol and extensions.
In short, Gnutella is where the action is. It's an open protocol with a great deal of development among the client developers as well as the academic community. Many of the clients are also open source, providing a rich overall platform for innovation. The original protocol with all of the extensions has also become quite complex. The design, I assure you, no longer "sucks."
If you're interested, check out the center for active Gnutella development at: http://groups.yahoo.com/group/the_gdf/
Thanks.
Adam Fisk
LimeWire LLC
Adam Fisk
LimeWire LLC
I assume that by design, you are referring to the underlying architecture of these systems? AudioGalaxy is basically Napster. It's a centralized system that avoids lawsuits by occasionally filtering their songs when things are looking ominous and with some people at its head that have connections with the record industry.
As far as security goes, it is true that the Gnutella protocol does not include encryption, whereas FastTrack encrypts many of its packets. Just running netstat will show you everyone's IP on either network.
I strongly encourage everyone to support Gnutella as an open standard. As Lawrence Lessig has eloquently argued in "The Future of Ideas," open standards are vital to encouraging innovation and for creating resources that can be enjoyed by all, not the few.
Thanks.
Adam Fisk LimeWire LLC
I think your research (and dissertation) is a bit out of date. The architecture of LimeWire's "UltraPeers" is very similar to the architecture used on the FastTrack network. In fact, UltraPeer's use of intelligent query routing makes them almost assuredly more efficient than the similarly designed FastTrack implementation, although it's really difficult to know because FastTrack is not open source and its messages are highly encrypted.
All that aside, FastTrack has not changed at all over the past 8 months, whereas Gnutella has made huge strides. What's more, many of the clients (like LimeWire and Gnucleus) are open source, and Gnutella is an open protocol. We need open standards like this to keep the Internet free (as in free speech).
LimeWire currently implements a variation of this -- what we call "UltraPeers." UltraPeers establish a significantly greater horizon on the network, and there are other distributed protocols that do this in other creative ways, such as Chord out of MIT, which can be found at: http://www.pdos.lcs.mit.edu/chord/ That aside, there is significant evidence to show that a distributed network can scale far better than any centralized network. Remember that Napster had serious scaling problems as well -- you could only see the files from the hosts on whichever server you happened to be logged in to. The only solution to that problem is the brain-dead purchase of a yet faster multi-million dollar server. I would not call that scaling. As everyone else has pointed out, this discussion began and ended in the Gnutella community about a year ago.
Apparently you have not looked at the LimeWire source. If you had, you would know that LimeWire was coded by a team, and that the LimeWire source is quite well engineered and well documented. I would recommend giving it a look at www.limewire.org. If you find specific performance bottlenecks, please feel free to e-mail me anytime at afisk@limewire.org. (think profiler). Thanks. Adam Fisk LimeWire Team