morpheus ultrapeers
 

I've been reading with interest the discussions dimping on the fact that morpheus (and bearshare)users are increasingly coming up as ultrapeers and not sharing and so reducing the effectiveness of the network for limewire users - I too have noticed that where once I only got limewire ultrapeers, now I get (lots of) morpheus, gnucleus, limewire@version@, bearshare (well once anyway) and even xolox - thing is - with morpheus it looks like I get few hosts and hardly any files available but when I search anyway I get as much as ever - eg - I had ALL morpheus ultrapeers and supposedly only 300 files available but a search of "when doves cry" brought up 200 or so ("when doves cry") files by Patti Smith, Ani D, BNL, Luka Bloom, Reclaimers, Flying Pickets, Shrodingers cat, Tupac, Afghan Wigs, 35 inch Mudder, Depeche Mode - and even - Prince - it all makes me think the 'problem' is partly illusiory - maybe the likely influx of audiogalaxy users won't be sooooo bad either

Morpheus ultrapeers break LimeWire's statistics, so when you are connected to Morpheus ultrapeers LimeWire cannot make sure how many hosts you can reach.

But Morpheus 1.9 sucks anyway.

Taliban, what's up with your sig? Why does Shareaza suck? I've used it once or twice and I thought it was a perfectly good client. It incorporates Limwire's XML metadata scheme, ultrapeers and hashing. What do you have against Shareaza?

-Chuck

I think the problem with Morpheus Ultrapeers is not that they can not offer as many files, but that, because many of them should not be given ultrapeer status (the default setting), the downloads are not reliable.

The stats have been broken/inaccurate for a long time now, but the new morpheus/gnucleus has pretty much mutilated it. Getting accurate stats with low bandwidth reqs on a decentralized P2P is not easy. But really, why does it matter if 30,000GB are available? As long as what you're looking for is there, isn't that good enough?

Cheers! :)

that's the problem, my search results are noticeably worse ever since i've been swamped with morpheus and gnucleus ultrapeers... guess either party should try to fix it, i don't want to block them out, but i will.

I added a few lines of code to my version of LimeWire so it will only connect to LimeWire/AquaLime ultrapeers. In case you are willing to compile LimeWire yourself this might help you:

Just edit the .../core/com/limegroup/gnutella/ManagedConnection.java and replace the initialize() method with the following one.

(lines 345 following)
public void initialize()
throws IOException, NoGnutellaOkException, BadHandshakeException {
//Establish the socket (if needed), handshake.
if (_isRouter)
super.initialize(); //no timeout for bootstrap server
else {
super.initialize(CONNECT_TIMEOUT);
if (getUserAgent() != null) {
if (getUserAgent().indexOf("Lime") == -1) // close connection if non-LimeWire
this.close();
}
}
//Start the thread to empty the output queue
new OutputRunner();
}


I know that relying on the user agent string in the handshake is not very safe, but it works.