Firstly, yes we need simulations, and we need them now. I believe that is Oskar's department.
Secondly, as I have stated already, I think it would be a MAJOR strategic mistake to deploy opennet without some significant and well publicised benefits to getting darknet connections. Admittedly there is the obvious security advantage: Opennet is visible, and vulnerable to Sybil attacks. 0.8's advanced security on premix routing and location swapping will probably not work at all on opennet; the latter may even require a separate pure darknet. But we need additional incentives or releasing an opennet will simply result in nobody using the darknet. Furthermore, we continue to have serious issues with load balancing, and we need to deploy a new storage strategy. Deploying opennet at this point would make a messy situation a lot messier. On Sun, Jul 09, 2006 at 11:02:04AM -0700, Ian Clarke wrote: > I think it is clear to most people that the lack of an opennet option > is proving to be a major inhibitor to adoption. What we have learned > in practice is that if you only give a darknet option to people who > are clearly willing to connect to strangers, then they will create > their own opennet using kludges like ifreed.net and ubernode.org. > Most Freenet users appear happy to connect to strangers, this isn't > surprising, nor is it undesirable. The goal of 0.7 was *never* to > force everyone to form connections manually, the goal was to provide > that option to those that want it. > > I believe that Freenet 0.5-style "destination sampling" should work > fine in conjunction with darknet connection swapping, as both > algorithms have proven themselves robust in practice. We should > simulate this to be sure. > > I really think we need to up the priority of opennet, as I think > currently we are squandering interest in Freenet by forcing people to > manually create connections to users. > > Ian. -- 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/20060710/8a3ad4f7/attachment.pgp>
