>In effect, a node itself could act in a similar manner to these=20
>"gwebcache" scripts, which could be harvested centrally and mixed=20
>together.

It actually does, through servlet/nodestatus/noderefs.txt.  Unfortunately that
servlet is not accesible from the outside in the default installation.  If it
were, it would take _literally_ 30 minutes to adapt the gwebcaches.  

>Having said all that, is there any evidence that the current mechanism=20
>isn't working?

Currently it could be argued that the very uneven distribution of the load is in
part due to all new nodes getting similar entry points to the network.  The
problem will manifest when the number of new nodes starts to increase rapidly
(i.e. when a new release is announced).

I know this is far from perfect, but it is very fast to deploy.  As I said above
if public access is given by default to the above servlet, it will take 30
minutes to have the caches running.  If that doesn't happen (there are probably
security/privacy issues), but if someone experienced in php is willing to help
it could be done in few hours.  Even if I do it all by myself it would not take
more than 2-3 days. 

_______________________________________________
devl mailing list
devl at freenetproject.org
http://hawk.freenetproject.org:8080/cgi-bin/mailman/listinfo/devl

Reply via email to