On Thu, Dec 20, 2001 at 05:03:51PM -0600, Mark J Roberts wrote: > You're right, except there's no multiplexing. We don't need > multiplexing to implement the NAT workaround since we can always > create a new connection the same way we created the first one.
But, how do we signal that we need another connection? Perhaps by implementing a signal system as a special message on a busy conection, sort of like out-of-bound data on TCP. 1. ask ocm for connection 2. if busy do signal on busy connection 3. sleep a little 4. retry This could be as complex as multiplexing, and if UDP transport is to be implemented multiplexing is needed anyway. But it is late now, so perhaps I'm missing the obvious and simple solution? > There's still a very real danger of > centralization, though, and it's something we have to seriously > consider. Yes, this relay thingy should only be used when absolutely needed. The alternative in some cases is not running freenet at all, which is hardly a better alternative. /g?ran -------------- next part -------------- A non-text attachment was scrubbed... Name: not available Type: application/pgp-signature Size: 239 bytes Desc: not available URL: <https://emu.freenetproject.org/pipermail/devl/attachments/20011221/5c287664/attachment.pgp>
