On Tue, Jan 03, 2006 at 05:59:33PM +0000, Ian Clarke wrote: > Its a good idea, but I think we need to put these ideas on the TODO > list for after 0.7.0.
Don't you think the transition would be rather messy? It's not like it's going to bog us down for months on end; why should we do this later rather than now? (We will have to migrate to the database backed datastore before 0.7.0 ANYWAY, because of memory usage/scale issues). > > Ian. > > On 1/3/06, Matthew Toseland <toad at amphibian.dyndns.org> wrote: > > Freenet is designed to be a meritocracy for content. If somebody inserts > > several gigabytes of pictures of mating chickens, for instance, this > > will only persist if it is popular, or so the theory goes. This > > meritocracy is one of the main principles underlying the moral, > > philosophical and practical aspects of Freenet. > > > > The only problem is this: If an attacker can continue to insert a file, > > he can keep it in the network forever. Inserts don't stop when they find > > the data (for good reasons IMHO), so this is more effective than > > requesting the data. And he doesn't have to push the same data, he can > > just keep inserting random data chunks. He doesn't need to find lots of > > nodes to connect to. These will then be promoted to the top of the LRU > > list, and therefore will push out moderately popular content - or even > > popular content, if he has many links to many nodes and lots of upstream > > bandwidth. > > > > Solution? > > All we have to do is not promote inserted content to the TOP of the LRU > > list. Set it say 3 days ahead of the MRU. This can be done efficiently > > with the new database-backed datastore. Then when it is requested, > > promote it to the top. If it isn't requested, for example if it is > > random junk inserted by Mallory, it will expire in 3 days. We could make > > it 7 days, or some fraction of the overall most-recent to least-recent > > period (check your Environment page, Data Store box; most nodes have at > > least a month between most recent and least recent access time). > > > > This should make flooding attacks considerably harder, and move freenet > > somewhat closer to the meritocratic ideal we claim. > > -- > > Matthew J Toseland - toad at amphibian.dyndns.org > > Freenet Project Official Codemonkey - http://freenetproject.org/ > > ICTHUS - Nothing is impossible. Our Boss says so. > > > > > > -----BEGIN PGP SIGNATURE----- > > Version: GnuPG v1.4.1 (GNU/Linux) > > > > iD8DBQFDurnNHzsuOmVUoi0RAiLIAJ44TGLQozq6RDi4zswMV9OnfdypqQCfYURA > > vteo3spHLTfiFFRh82l03S4= > > =00Cd > > -----END PGP SIGNATURE----- > > > > > > _______________________________________________ > > Tech mailing list > > Tech at freenetproject.org > > http://emu.freenetproject.org/cgi-bin/mailman/listinfo/tech > > > > > _______________________________________________ > Tech mailing list > Tech at freenetproject.org > http://emu.freenetproject.org/cgi-bin/mailman/listinfo/tech -- Matthew J Toseland - toad at amphibian.dyndns.org Freenet Project Official Codemonkey - http://freenetproject.org/ ICTHUS - Nothing is impossible. Our Boss says so. -------------- next part -------------- A non-text attachment was scrubbed... Name: signature.asc Type: application/pgp-signature Size: 189 bytes Desc: Digital signature URL: <https://emu.freenetproject.org/pipermail/tech/attachments/20060103/b557360f/attachment.pgp>
