Spammers Win: Frostwire Drops Gnutella

[skinvista]

http://torrentfreak.com/frostwire-kills-gnutella-to-go-all-bittorrent-110627/

Frostwire 5 is apparently dropping Gnutella to become BitTorrent-only.

The reason given is the obvious spam problem designed to moot the network. Hence the spammers finlly score a big PR/practical win, instead of a pyhric victory after Limewire. And strange, since its BT support is(was?) only a tack-on Azureus solution.

While Shareaza's support may be dated and redundant, there needs to be comitment. Philosophically and strategically, ground should not be surrendered. Both historically, and as a hedge against attacks on G2 etc., a viable Gnutella -and any non-bittorrent- is in everyone's interest. Much less expanding on BitTorrent in a balanced way, such as integrated search.

As Limewire/Frostwire share gradually fades, there are questions to address. Is there a new standard bearer for LimeWire code? Will old clients be required for any effective utrapeers? What other active G1 clients are there to engage? What standards-compliant -or simply dominant- client may lead any ongoing development? What is our role/reaction moving forward? G2 network's? How do we better address the spam landscape?

On the issue of the current search spam itself, there needs to be an analysis or distributed update. Last week I spent less than an hour idly searching Gnuttella until I had a thousand unique IPs -many residential range- that act in the same coordinated pattern. With no benefit from the already unwieldy filters.

Not only would a solution potentially save gnutella from the corporate-paid disruptors, but G2 would need to be better secured from any such attack before attracting further attention. At a time when Shareaza has not been updated in a month, the longest span in its history.


Simply a sad turn, pragmatic or not.

[borsti67]

Well, I also disabled Gnutella a long time ago, because the spam rendered it useless anyway. I would be more concerned if we were talking about G2...

But I can't see the benefit of their decision. It sounds like Frostwire is going to be just-another-bittorrent-client? Why should anyone choose it?
Bittorrent may be nice for actual things, but only as long as enough people are interested and seed. There is no kind of "archive" like you have in G2 or ED2K.

[ailurophobe]

What is the "coordinated pattern" you saw?

Speaking of protocol improvements, wouldn't simply throttling keyword queries you forward to a single leaf or neighbour reduce spam? The key feature of a spammer is that they match almost any query term, after all. Something like counting the time connected and queries matched and not actually forwarding the query if the ratio is bad. Would save leaves with high file counts some "ignoring query in excess of query limits" as well.

No opinion on Gnutella, and IMHO we don't need one either. Let's just see what happens...

[cyko_01]

[old_death]

[aaron_walkhouse]

Interest in gnutella will soon surge once the next wave of harassment tactics hit american users
of bittorrent and they realize that gnutella, G2 and other methods don't broadcast your position
to MAFIAA copyright trolls and their contractors through forced sharing.

[aaron_walkhouse]

Get ready for a lot of new Shareaza users too. Bittorrent is about to start bleeding and
plenty of gnutella users will be looking for something similar in the short term if FrostWire
disappoints them with this move.

[old_death]

[biggestnoob]

This could be good for Shareaza. The problem is how many people know about shareaza and not the fake shareaza 7.0?

Judging by Alexa many people are fooled by shareaza.com website. Hopefully some hackers can take down that website and put a virus to destroy all their data.

[taxiboy]

So this begets the question will Shareaza also drop G1 for Kad?

[ailurophobe]

How? The two are not related in anyway that I know of.

[old_death]

Probably we're going to just keep gnutella support and add KAD anyway after some internal structural updates...

[taxiboy]

[bichodosol]

[aaron_walkhouse]

You're that willing to let the spammers and anti-P2P bunch try to kill a network?

I'm not, and have already made FrostWire spamproof for those who will stay on gnutella.
If necessary I will fork FrostWire again because a lot of people are sticking with it as-is.

It's attitudes like that which make you the next target. Apathy attracts that sort of thing.
The spammers didn't win. Don't declare surrender for others. You don't have the right.

[aaron_walkhouse]

[taxiboy]

[kevogod]

[ailurophobe]

[aaron_walkhouse]

For proof go get FrostWire 4.21.8 and follow the instructions I published in their forums.
Since I built and installed those two tiny filters for IP and keywords the number of spambot-
spoofed search hits abruptly shrunk to just a few hits at most and for the past few days, totally
stopped. I have been blocking spam since long before P2P existed and it's pretty safe to say
that I am starting to get pretty good at it. It only took me a week to clean up FrostWire's search
results to the point where you must try hard to find any automatically generated spam at all.

I'll see if I can translate those rules into Shareaza's format next because it will make gnutella
one of your favourite networks again and those spammers might just try to target G2 next.
Don't forget that gnutella still has more users than G2 and eMule combined so ignoring it
is giving up on a resource at least as valuable as the others.

Why would the spammers be on gnutella if there weren't a lot of people using it? What would
you do if they migrated to the smaller networks with the same machinery they used on gnutella?
It would be in your best interests to keep them where they are because only gnutella is big and
strong enough to keep them occupied. Having a working gnutella servent at a time when 20-30
million americans are looking for safer alternatives to bittorrent will bring a lot of people in for a
look and Shareaza is now at a point where it will look too good to pass up. Even though it never
had ultrapeer capability there will always be plenty of the existing servents giving it a connection.

Being "good for leeching" is actually very good for a sharing network. The strength of networks
like gnutella and G2 despite being "good for leeching" for twelve and ten years running is ample
proof they work well, that people still use both and that most of them always shared freely.

The myth that most gnutella and G2 users are leeches is not only completely wrong but is quite
unfair to the 75-80% of uses who share freely without being forced to do so. You can easily
verify that percentage by running BearShare and watching it's handy statistics window. Even
though the number of BearShare users is down to only ten percent of the 2005 levels the
percentage of "leeches" has held steady and the total amount of shared files is still over five
million and just under 60 terabytes. That's with BearShare alone. There are many more users
of the other gnutella servents and there's no reason to think they behave any differently.

I use many networks constantly to share over 2000 files and 30 gigs. They all are better than
each other in various ways but none are inferior to the rest. People are not so selfish that they
have to be forced to share and no P2P network has ever needed any extra help. In fact, not
forcing sharing cuts down the glut of popular files and makes it easier to find stuff you really could
never find by other means. Think of gnutella as being like G2 and emule but a lot bigger with
many more files. When you block the spam like I do you would see that for yourself and
leave it turned on all the time.

As for the AP2P contractors' spying: They look for uploads because only that is copyright
infringement, not downloads. Even if they could document your downloads they wouldn't
bother because they couldn't even send a DMCA takedown, much less sue you for copying
and distributing because downloads aren't proof you did any of that. Using protocols that force
sharing make you potentially liable for copyright infringement every time you use them to download
a copyrighted file; so since sharing is the only thing the MAFIAA looks for online it is indeed
sending up a flare that they can not only home in on but use in court against you.

[old_death]

Before starting your work, could you have a look at my security code updates to avoid doing the same work twice? Note however that I am still working on it.

Code update: viewtopic.php?f=13&t=1057&start=40#p7869

[cyko_01]

[ailurophobe]

[cyko_01]

[old_death]

Well, not on G2, but on gnutella they are most probably... If I do use that network for searching for files, usually >60% of the returned hits are spam...

But never mind that, once my security filter udpate has been added to the Shareaza code base, handling filters of that size should be possible without problems...

[ailurophobe]

IP (and hash) filters yes, keyword filters processing might still be an issue, if the number is high enough.

[old_death]

Maybe we could add a map for keyword filters to contain all keywords (included and excluded) to increase the searching speed for rules that do match a content string... This might increase the speed for high numbers of rules, however it might decrease it considerably in case we are dealing with small numbers. Maybe Shareaza could check how many rules there are in regular intervals (or only when they're added or removed, like it is done in my new security code for IP related rules and the miss cache) and use either the current list if there are only a few rules or the keyword map if there are enough of them...

[ailurophobe]

That should work for keywords, regular expressions cyko loves couldn't be helped I think. And it isn't really necessary to optimize performance with low number of rules, it is going to be fast enough anyway, if the performance with large rule sets is acceptable. And the optimization code adds overhead and clutter.

[cyko_01]

[awdrifter]

[borsti67]

[ailurophobe]

Also, back when the upgrade server was lost and people who got used it to offer a fake "update" what saved me from falling for it was the fact that the version number was too large to make sense. So there is a benefit in using a different numbering system than what the ethically challenged use.

[biggestnoob]

Can't you just have the program check for files every time it closes and opens. If the file is created between that time, they will all be automatically blocked. Or am I being noob right now and talking nonsense.

[cyko_01]

[dark146]

Well it looks like G1 is officially dead. At least there still is the G2 network.

Long live G2.

[gs1000]

Where is the story of Gnutella being closed?

[cyko_01]

[gs1000]

Thanks for the answer, I wasn't notified by email. But when you say "officially dead" that has to mean more than just two programs' activity.
I'm not a technician. I assume the digital system is still there for Shareaza to use, etc.
Altho I'm not getting search results like I always used to lately.

[cyko_01]

since gnutella is a decentralized network (no central servers to kill) it is pretty much impossible to really kill it, but with the 80-90% of its users now gone the network is not nearly as big or as useful as it used to be.

[fifi9491]

I'm not sure this is the right place but I do not care any more, my search brings up only limewire and frostwire rubbish, can any one please explain
ever heart of 'Audials' that's my place to go now
BYE BYE SHAREAZA
Fifi