On Wed, 20 Feb 2002, Timm Murray wrote:

> <>
> > The main reason I want to separate queries 
> > for the location of the data and queries to retrieve the actual data is to 
> > make it possible to accesses any data on that is still on the network no 
> > matter how unpopular it is . . . 
> <>
> 
> Doing a direct connection won't help its popularity.  So very unpopular 
> content that is still somewhere on the network will stay that way, even 
> when if later on it gets a sudden surge in popularity.  

By my method if very unpopular and data gets a sudden surge of popularity 
it will be on the nodes which requested the data.  The more nodes that 
request the data the more nodes in which the data will spread.  So 
suddenly popular data WILL spread throughout the network, just not as 
fast as at will on freenet.

> To solve the 
> problem, you need to have the data spread to other nodes whenever a 
> request for that data occurs.  Freenet's meathod is to "pull" that data 
> through the request chain.  It could also be accomplished by doing a 
> direct connection, but then having one of the nodes "push" that data via 
> a standard insert.

Yes a push would work.  So would telling a node B to get data from C but 
then also getting the data directly from C.

--- 
http://kevin.atkinson.dhs.org



_______________________________________________
freenet-tech mailing list
[EMAIL PROTECTED]
http://lists.freenetproject.org/mailman/listinfo/tech

Reply via email to