why not add a property to every key that says: pls do not remove me
from your cache unless you 1st try to republish me to freenet?
again no garantees with this approach but so it goes...
On Sat, May 19, 2001 at 07:23:57AM +1200, David McNab wrote:
> OK - I *do* know that the whole Freenet architecture is dependent on
> ephemeral storage, and the ability of nodes to purge older less popular
> files.
>
> But an idea came to me for how *some* files could be kept available
> permanently.
>
> The idea would require that some people anonymously volunteer to work as
> 'Freenet archivists'.
>
> All this would require is:
>
> 1) For archivists - using an 'archive' program which copies retrieved keys
> into a 'permanent cache' directory on their hard disk, and also maintains a
> database of such keys.
> This is simple - archivist is browsing, and thinks 'i feel this should be
> permanently available'; (s)he then runs:
> freenet_archive key_uri [-scomment]
>
> 2) For clients - anyone who requests a key and fails, can run an 'archive
> request' client which simply writes an archive request to a keyindex:
> freenet_archive_request key_uri
>
> 3) For archivists - run an 'archive server' program which regularly harvests
> the 'archive request' keyindex and re-inserts any requested files which are
> found on the archivist's permanent store, also logging such re-insertions to
> another 'archive response' key index.
>
> 4) For clients - 'archive request' client can check the 'archive response'
> key index (say over the last 2 days) and see if their request has attracted
> a response:
> freenet_archive_status
>
> This won't guarantee permanence of all material, but it does give
> individuals the power to guarantee that materials of their choice will
> always be available within a (say) 48 hour window.
>
> Summary - software required:
> 1) client - send archive request to key index, log this in local database
> 2) server - add files to archive, browse archive, prune archive
> 3) server - daemon - read archive requests, reinsert keys found in archive,
> log successful reinsertions to an 'archive response' key index.
> 4) client - daemon which periodically peruses 'archive response' and seek
> items previously logged to local database.
>
> If no-one has any objection to this, I'd like to make it my next project
> once FreeWeb goes beta.
> If someone beats me to it, I'll support totally.
>
> Suggestions - very welcome and appreciated.
>
> Cheers
> David
>
>
>
>
> _______________________________________________
> Chat mailing list
> [EMAIL PROTECTED]
> http://lists.freenetproject.org/mailman/listinfo/chat
_______________________________________________
Chat mailing list
[EMAIL PROTECTED]
http://lists.freenetproject.org/mailman/listinfo/chat
