I'd like to see all nodes store *everything* *everytime* that they find data 
in a new short-term datastore.  The existing datastore can hold long term 
specialised information as now.

The short term store does not apply to insertion requests. The short-term 
store can use the current store implementation code (but with a different 
file) as the short-term bit is just a matter of how the client uses the store.

If you are worried that the route could be traced based on retrieval timing, 
then maybe have each node simulate a random delay to pretend that the data 
was not cached locally.

This could drammatically reduce network traffic and node load.

I have always been unhappy at a single tier of LRU store, since data 
popularity is not related to data long term value.
E.g message board daily index.

Also it might be advantageous to have a client specified privacy level on 
each request. Then less private and faster routing algorithms could be used
when privacy was not required.

-- 
Christopher William Turner, http://www.cycom.co.uk/ Java development since 
1996
http://club.cycom.co.uk/tms.htm Terminology Management software
http://club.cycom.co.uk/wt.htm  Wind Turbine blade design software

_______________________________________________
devl mailing list
[EMAIL PROTECTED]
http://hawk.freenetproject.org/cgi-bin/mailman/listinfo/devl

Reply via email to