Decentralized network, can it be done and how?

[_eAgLe_]

I am now wandering about Exeem, supposedly they were a decentralized network, but everyone is sasying that it isn't, So whats the deal here, how is the exeem network working?

[Jabryl]

It ( Overnet ) uses Kademlia.
You can download some docs here: http://www.rateless.com/pubs.html

WTF!? I thought that Kad was based on Overnet.

eMule developers just aren't very imaginative with names.

Overnet is based on the Kademlia routing protocol developed by Petar somebody or other and some other peeps at MIT.

I see.
I like the name Overnet better.

[bean999]

I am now wandering about Exeem, supposedly they were a decentralized network, but everyone is saying that it isn't, So whats the deal here, how is the exeem network working?

I think through the node list here:

[HKEY_CURRENT_USER\Software\exeem]

"nodes"="0 217.165.73.8 4000 2005 01 26 17 31 0 etc.
"segments"="1 45875 143 1 42961 141 1 39756 etc.

This is laid down in install. Thereafter it updates (in the registry) while exeem is running.

And this is what the reg fix updates.
---

[_eAgLe_]

Well, a few people on slyck were saying that exeem was not decentralized, but it appears that it is.... Am i wrong?

[bean999]

Well, a few people on slyck were saying that exeem was not decentralized, but it appears that it is.... Am i wrong?

Seems this is based on:

- Idea that the network is blocking exlite - but the conclusion for now in this thread is that this is a version compatibility problem.

- Requirement (at one time) of users to keep client open (as supernodes) - but as pointed out here somewhere this doesn't mean the network isn't decentralized nonetheless.

Unless I've missed something that seems to be about it?

So for now it looks as tho exeem is decentralized.
---

[_eAgLe_]

ok thanks for the info bean999 and everyone.

[d8888]

__________________________________________________________________
the have a unique ID produced hashing infos from their time and their location (this yields a unique hash)
__________________________________________________________________

Don't do such a stupid thing.

Using "Location" and "Time" uses only a very small portion of the available 2^160(or 2^256,depends on what hash you use) hash space.

Do some calculation:86400 secs in a day,30 days in a month,and 365 days in a year

log(86400*30*365) / log(2) = 29.8.

So there are only actually 2^29.8 variations

So the chance of collision is actually HIGHER,and that's not the most important problem.

A hash function with only 2^29.8 outputs can be easily reversed.So Anyone who get your hash can do some little work and calculate the time and location that you use.

If you want to avoid collision,use verylong random number generator.A 32 byte random number is not a burden when transfering,but it provides 2^256 variations.If all 6 billion people log into a p2p network,it's still only 2^32.

It's 2^32 people in a 2^256 space,so rate of collision is 2^-224.More inportantly,the 256bit random number does not contain anything about you,and it can't be reversed.

[_eAgLe_]

Hey guys again. I am under the impression, that when you are downloading something on the fasttrack network, you make a direct connection with the computer you are downloading sdomething from.... is that right... this is what i mean:
'
Computer A (111.111.111.111) is asking for a song from Computer B (222.222.222.222)
So ComputerA makes a direct connection to the IP Address 222.222.222.222. So if PersonB was hosting a fake file, he could get the IP Address of PersonsA....
'
Am i right?

Also if this is so, is there anyway of hiding your IP Address from them?? i think IP Spoofing does something like that, but i actually want to know how it works.... could someone plz help me out here?

TIA.
I hope that all made sense.

[Allied]

...
I was under the impression that Fasttrack, G1, G2, Overnet, Kad and Ares were all decentralized.

So, I don't see anyone challenging Overnet and KAd but what about Fasttrack G1 G2 and Ares? Is there some server out there that could shutdown the entire network making the millions of users on those networks loose their connections?

If so, why hasn't the RIAA gone after that server?

[_eAgLe_]

because the RIAA are stupid

[Jabryl]

When downloading you have a direct connection to the user with the file, searching is another matter. When trying to locate sources a network needs some way to keep track of who has what files, and to connect the person looking for a file with the person who has it.

eDonkey has hundreds of large indexing servers that do this.

Fasttrack and Gnutella use a lot of sort of mini servers, created from any ordinary user that has enough bandwidth and computer power.

Overnet and Kad use distributed hash tables, whereby everyone in the network has some info on some files. Your search basically bounces around the network, with each bounce coming closer and closer to the person who has what you're looking for.

[_eAgLe_]

Thanx Jabryl for the info. But is there any way to stop the user sending you the file from getting your IP Address? I think IP Spoofing does something like this... but could someone let me know how IP Spoofing or any other way of hiding your IP works?
TIA

[_eAgLe_]

Also, i know that when searching it asks from one computer to another, but wouldnt this slow getting reults a lot!!!?
Sorry for posting twice.

[Jabryl]

I know it'e possible but not easy. I don't know much about it myself.

[What IS Real]

Hi, can someone tell me if it is possible to have a file propagated by a filesharing network. ie a file that was forced on every client on the network, one that could be dynamically updated by each and every user. If so, a list of files being shared on the network could be written into this file and then propagated to each client on the network. Each client could update the list with items being shared. Kinda like a Wiki that is hosted everywhere not in one single location. Then, we could have an indexing system that is beyond reach of adversaries. I guess 'sharewiki' would need to be moderated like suprnova/sharereactor were to make sure bogus files from adversaries were not being listed, but its an idea that sprang to mind.
Seems to be the obstacle in a truly decentralised network.
The 'sharewiki' could provide hashed links and tracing information to other elements of the network.

[DaBlade]

The only way I see possible is that all peers are equal. No primaries, no secondaries. Nodezilla have done a good job this way from what I've heard.

[xxxsidxxx]

it can be done simply...

http://msdn.microsoft.com/msdnmag/issues/01/02/netpeers/default.aspx

[What IS Real]

So all we need is some SOAP ;p

Surely this is the way forward, then complete decentralisation (Pure p2p) is possible, combine that with the work being done on ANts and clients performing proxychaining, and a hypernet is born, impregnable.
Sure to begin with we'd have to face a slowing bandwidth wise, but as my ISP now gives me 400up and 4000down, it won't be long before we've got the bandwith to not notice.
UploadWidth needs to equal DownloadWidth across the network.
The future's looking safe and decidedly different.
Welcome to Subscription-based life!!
Movies already are, so are telecommunications, music soon will be. Games next?
How long before an all you can eat pay monthly supermarket, OR an all you can drink off-license ;P

[jonne]

most p2p networks are decentralised.

Gnutella has been decentralised from the start.
Webcaches are just used to get connected to the network, but you might aswell add the IP of your hub manually (ask a friend on the network to give you the IP of a hub).

same thing with Fasttrack (they have one server somewhere that gets queried after you just installed it, after that you just use a list of supernodes stored in the app), G2 (same thing as gnutella), Ares (I think, it's probably the same mechanism as fasttrack), etc.

Servers are just used to get connected to the network, so you don't have to go on IRC to ask an IPaddress of a hub. If the servers are taken down, the network doesn't necessarily fail, you can keep connecting, if you want to spend some energy.

I really don't see the point of this thread.

[Ooble]

By true decentralization, I imagine a network where finding that first IP address to go to from a central location is unnecessary.

[Asuran]

By true decentralization, I imagine a network where finding that first IP address to go to from a central location is unnecessary.

It's not required to obtain it from a central location. You can connect to any peer in the network. Webcaches are just practical to make it easier. They are not required for the system to work.

[AussieMatt]

Well, a few people on slyck were saying that exeem was not decentralized, but it appears that it is.... Am i wrong?

Seems this is based on:

- Idea that the network is blocking exlite - but the conclusion for now in this thread is that this is a version compatibility problem.

- Requirement (at one time) of users to keep client open (as supernodes) - but as pointed out here somewhere this doesn't mean the network isn't decentralized nonetheless.

Unless I've missed something that seems to be about it?

So for now it looks as tho exeem is decentralized.
---

Heres some info from Infoanarchy that points out why Exeem isnt centralised many p2p apliciactions have numerous methods to obtain a network 'bootstrap ip' exceem has one for the average user.There are many 'nodelist' servers or webcaches on many p2p networks Exceem has 4 DNS servers that resolve to one IP therefore its not decentralised.The File Transfer is decentralised.

Exeem/Protocol
From infoAnarchy Wiki
< Exeem

The Exeem protocol is the network protocol that Exeem uses.

Exeem only begins a connection to the Exeem network when it begins an operation. On Exeem's initial connection to the Exeem network, it bootstraps. It does this by doing a DNS lookup on four FQDNs, all of which as of January 22, 2005 point to the same IP address. Exeem then connects to this IP address on port 1026, sends an Exeem packet with function "6". The server responds with a list of 20 or so Exeem Superpeers to connnect to. Subsequently, when one does a search or new files research, the Exeem client connects to these 20 hosts and does a search, or a request for new files.

http://www.infoanarchy.org/wiki/index.p ... m/Protocol

[_eAgLe_]

Ok thanks AussieMatt.

[larytet]

Hey guys again. I am under the impression, that when you are downloading something on the fasttrack network, you make a direct connection with the computer you are downloading sdomething from.... is that right... this is what i mean:
'
Computer A (111.111.111.111) is asking for a song from Computer B (222.222.222.222)
So ComputerA makes a direct connection to the IP Address 222.222.222.222. So if PersonB was hosting a fake file, he could get the IP Address of PersonsA....
'
Am i right?

Also if this is so, is there anyway of hiding your IP Address from them?? i think IP Spoofing does something like that, but i actually want to know how it works.... could someone plz help me out here?

TIA.
I hope that all made sense. :D

IP spoofing is impossible if you use TCP protocol and this is generally the case in P2P. every packet your TCP/IP stack sends out is stamped with your IP source. uploader will use this IP address to send requested data back to you. Imagine envelope with prepaid stamp and your address on it. you send the envelope to the peer, peer fills the envelope and sends back.

In some TCP networks (MUTE, Ants) was made an effort to PROXY all your packets both outgoing and incoming, so destination peer sees only IP of PROXY, not yours IP.

see also http://en.wikipedia.org/wiki/IP_spoofing for more details, links

try also google IP spoofing P2P
http://www.google.com/search?hl=en&q=IP ... gle+Search

[countrypirate]

Anything is possible be cause everthing including technology evolvs and intergrates. What was'nt possible ten years ago is now commonplace.