How's this for an interesting compromise:

We define the freenet URL as

freenet://<key>

No hosts are in place.

At the same time, we implement a set of well known freenet Web servers that 
simply maintain a changing list of nodes to link into freenet (this is 
something akin to the DNS root servers).

The netscape client plug-in first polls one of these freenet root servers 
for a few freenet entry points.  After which, the system is now self 
sustaining.  Also, if the client turns off his machine for extended periods 
of time, and the cached server lists is no longer valid, then the plug-in 
will just poll the root servers once again.

I agree that the URL/URI should not have a specific node in place.  This 
method should solve the problem.

As well, this method should allow for algorithms to "Akamize" the entry 
points depending upon the client's IP and route.

I believe this gives freenet the proper mix between centralization and 
decentralization.

Comments?


Paul Tan.



________________________________________________________________________
Get Your Private, Free E-mail from MSN Hotmail at http://www.hotmail.com


_______________________________________________
Freenet-dev mailing list
Freenet-dev at lists.sourceforge.net
http://lists.sourceforge.net/mailman/listinfo/freenet-dev

Reply via email to