In message <[EMAIL PROTECTED]>, [EMAIL PROTECTED] writes
Nick,

Thanks for your thoughts.

I think that the trick is creating a distributed index, so that every node can have access to a more or less complete and accurate index of all of the files available on the LAN. It seems to me that this would be practical. Because we're dealing with a number of files on the order of hundreds of thousands, it seems like the index would take up well under 20mb per node.

Does anyone have tips on how to best propogate that kind of index? I need to keep in mind that not all of the nodes can be trusted. For instance, we can't have one node reporting that all of the other nodes are offline, and wiping the indexes.

Also, after being installed, I would want the daemons to be pretty much autonomous. I won't be able to easily manually stop and restart nodes if something bad happens. To do that I expect to have to make some minor additions to the program (or separate programs to package along with it). The additions would be used to maintain the index.

Having that kind of central index would make searching the files available on the LAN fun and easy. A little Delphi or Cocoa program could be written in a few hundred lines to query the local index and download or stream the file from the Freenet.

Best Regards,
Drew


If there was an easily-accessible index, would not the University lose the plausible deniability which is the whole point of using Freenet in the first place?





----- Original Message ----- From: Nick Tarleton <[EMAIL PROTECTED]> Date: Sunday, August 10, 2003 7:06 pm Subject: Re: [freenet-dev] Distributed LAN Filesystem

On Sunday 10 August 2003 07:45 pm, [EMAIL PROTECTED] wrote:
> Could Freenet be adapted to function only on a LAN, to share
file of
> dubious legality?  It seems that the University could avoid
liability, by
> acquiring common-carrier status for the files that are
distributed.  Also,
> the RIAA would be none-the-wiser to all of the files shared
internally over
> the University Freenet.
I'm not quite sure how this would work, but...

> I don't know much about the inner workings of Freenet, but what
would it
> take to get this started?
(note: this applies to static IP networks only, if you use DHCP
ask someone
else, and all this is just my idea, I've never actually done it)
1. Start 20 or so nodes, then shut them down. Make sure their IPs
are set to
their internal network IPs.
2. Get their public node references (I don't know how to do this,
but it's
*not* the 'node' file), put them in one big text file, and save
this as
seednodes.ref on each node, replacing the default.
3. Start them all back up. Test.
4. To establish a new node, before starting it for the first time,
replace
seednodes.ref with the copy created in step 2.
5. Periodically export new seednodes.ref files and use these instead.
6. Change the updater to get the seednodes.ref file from some
internal
location instead of the Freenet site.

If one local node gets the wrong seednodes.ref, then AFAICS, your
local
network will slowly coalesce into the main Freenet as
announcements pass
through that one node, then a few more, then a few more...
disastrophe. If
this happens, shut down all nodes at the same time, replace all of
their
seednodes.ref, and restart them. Needless to say, in a university,
that won't
be easy.

> It seems to me that the main hurdle (after convincing the University
> administration to go along with it) is to create a useful means of
> searching.
People could create their own indexes on freesites, or use Frost
(jtcfrost.sf.net), which is a message board/filesharing app.

--
http://earth.prohosting.com/tqbay
"I love deadlines. I love the whooshing sound they make as they go
by." - Douglas Adams
Nick Tarleton - [EMAIL PROTECTED] - PGP key available


-- Roger Hayter _______________________________________________ devl mailing list [EMAIL PROTECTED] http://hawk.freenetproject.org:8080/cgi-bin/mailman/listinfo/devl

Reply via email to