[P2PSocialNetwork]

Hi all, I have some thinking on the P2P network things in the following pdf link, it is immature opinion and open for discussion.
If this can give you any benefits or make any difference to any P2P program, I will be very pleased.
download the pdf

This post has been edited by P2PSocialNetwork: 04 April 2012 - 11:16 AM

[P2PSocialNetwork]

Am I talking about a naive idea here? or the content is poorly presented?
Any comment will be appreciated.

[coluche]

P2PSocialNetwork, on 05 April 2012 - 03:56 PM, said:

Am I talking about a naive idea here?

no idea. because :

Quote

or the content is poorly presented?

definitely.
I am not going to d/l some pdf to read about sth.

imo. pdf is for printing, and should never have started to be seen as good acceptable method to present stuff on-screen.

[someQuestion]

This website make some mistakes yesterday,so I just think you just kid me!
while it`s Ok now!

[P2PSocialNetwork]

coluche, on 05 April 2012 - 02:18 PM, said:

I am not going to d/l some pdf to read about sth.
imo. pdf is for printing, and should never have started to be seen as good acceptable method to present stuff on-screen.

OK I will rearrange the content to fit the web form.
And 51 pdf views in 180 topic views proved you are right.

This post has been edited by P2PSocialNetwork: 07 April 2012 - 01:08 AM

[Stulle]

Don't bother. The bits I understood seemed fairly stupid. It seems you have little idea about the technicalities and come entirely from a users point of view which is a bit of pointless view if you want to create a network of some sort.

[P2PSocialNetwork]

Stulle, on 07 April 2012 - 06:53 AM, said:

Don't bother. The bits I understood seemed fairly stupid. It seems you have little idea about the technicalities and come entirely from a users point of view which is a bit of pointless view if you want to create a network of some sort.

You are quite right!
Firstly, I skipped the technical part because I think maybe it is too early to discuss the technical implements, can you or any other point out things seems impractical? then I may present a technical solution or find a work around.
Secondly, this presentation should be stupid, because it is beta version and I have no experience in bottom layer development, Could you please kindly point out the things that you find doesn't make sense? no need for all, just take the one that you think is most stupid. For example, XXX is not necessary we already have YYY. or XXX cost too much could not possibly work.
The purpose I discuss it here is want to figure out whether we can possibly make the P2P software like emule more robust in peer trust peer way.

[Stulle]

-.- It should not be technical and yet again you speak about sending emails using the POP3 or IMAP protocol? Well, you had me at that point because both protocols are made to retrieve emails...

After skimming through the first few pages of that ppt made pdf I had exactly no clue what you are suggesting because you continued elaborating pseudo-technicalities. To the date I do not even see why you would want to append the eD2k or KAD network and why I need to transfer some list or other to do just that.

[P2PSocialNetwork]

Stulle, on 08 April 2012 - 07:26 AM, said:

-.- It should not be technical and yet again you speak about sending emails using the POP3 or IMAP protocol? Well, you had me at that point because both protocols are made to retrieve emails...

Sorry, I was thinking SMTP as well, forget to write it on. actually it is a method to update peer info(like IP address) to offline peers, so that they know how to reconnect each other even the IP address has been changed during the offline period. It is possible to implement I think.

Quote

After skimming through the first few pages of that ppt made pdf I had exactly no clue what you are suggesting because you continued elaborating pseudo-technicalities. To the date I do not even see why you would want to append the eD2k or KAD network and why I need to transfer some list or other to do just that.

Yes, my fault, I did not mention the reason why we need the pseudo-technicalities.
The problem is using a server to keep peers information (like ed2k) is vulnerable, if servers down or be banned the whole network will not work.
Just for this reason we created KAD. The disadvantages of KAD are:
1. take a long time to train node list when reconnecting to the KAD network;
2. If start without ed2k server, only using www.nodes-dat.com, I personally found it is painful to wait for the useful nodes and there are too many fake or invalid nodes.
3. KAD network can be attacked as this thread discussed.

So I want to know if we can improve this by remembering and trusting the nodes(peers). To do this we have to solve these problem:
1. Distinguish each node;
2. Evaluate the nodes if they can be trust;
3. Find and connect trust node.

So I made the PPT try to presenting a solution for the three problems above.
Now I realized that this is also too early to be discussed. The proper question is whether KAD can be improved as the three things above.

This post has been edited by P2PSocialNetwork: 08 April 2012 - 01:56 PM

[Stulle]

Do you have any experience in coding and/ or network design? Because your conclusions and solutions seem very simplistic and more born from good intention than actual knowledge.

Besides, while KAD can be attacked I don't see how your proposed changes can make a difference. KAD is not even slow or anything. There is no need to have a huge nodes list as long as we are connected to network. In fact, the idea of uniformity within the KAD network makes is safer to attacks than some kind of trust system could. Such a trust system is just another point to attack. Anyhow, KAD is rather robust right now against most any attack so there is really not a huge need for change.

[P2PSocialNetwork]

Stulle, on 08 April 2012 - 03:08 PM, said:

Do you have any experience in coding and/ or network design? Because your conclusions and solutions seem very simplistic and more born from good intention than actual knowledge.

Thank you for joining the conversation.
I have good experience in DB design and OO coding, but limited network knowledge.
You are right, I should do more study on KAD behavior before jump to the conclusion KAD need change.
This is only based on my personal experience of KAD.
Then, I will try to look into it with statistical eyes

This post has been edited by P2PSocialNetwork: 09 April 2012 - 02:13 AM

[Stulle]

Yeah, thought as much. Before trying to change things make sure you really understand them. Otherwise you will either not be taken seriously or break things or even both.