Hi all.

This is in continuation to the earlier discussions held about an adhoc
metalink swarm system. (http://tinyurl.com/an9rg7)

The main thing that came out of the above mentioned discussion, is
that most of the things required for such a system are very well in
place. BitTorrent provides a good enough method for file exchange. I
haven't read the complete and latest specs of BitTorrent, but the
thread says that BT fails when it comes to different sized pieces.
So the question concerning such a system is the protocol that needs to
be followed for exchanging files. BitTorrent cannot be used if pieces
are supposed to be of different size, and HTTP/FTP seem to be too
rigid for this purpose. Then how should the data transfer between
nodes occur? And how are different sized pieces advantageous as
compared to fixed sized ones?

I have another query here -- If we are building a system that is
supposed to be better than other P2P systems, we should avoid a
central point of failure, ie, no trackers or any central database. In
such a situation, how is the discovery of nodes supposed to take
place? The Gnutella system is nice; a predefined node (which *has* to
be on the network) is pre-programmed into the client, so that the
client can send all the query messages to this node, which can then
forward this query to other nodes it is connected to. But since it
effectively is pinging the entire network, should we avoid such a
system? In other words, how is the discovery of the nodes and the
files present on these nodes, supposed to work?

Also, a P2P and client/server hybrid system was suggested, where a
client could query a tracker to get a metalink file and once it has
finished downloading, it can register itself at the tracker, so that
other nodes joining in future could know about this client. Would
skipping a central tracker be feasible here? If yes, then what could
be an alternative to this?

Looking forward to more inputs.

And I hope I made sense. :-)

--~--~---------~--~----~------------~-------~--~----~
You received this message because you are subscribed to the Google Groups 
"Metalink Discussion" group.
To post to this group, send email to [email protected]
To unsubscribe from this group, send email to 
[email protected]
For more options, visit this group at 
http://groups.google.com/group/metalink-discussion?hl=en
-~----------~----~----~----~------~----~------~--~---

Reply via email to