On Tue, Dec 02, 2003 at 11:41:23AM -0800, Martin Stone Davis wrote: > Even with iTTL and unobtanium, we would still get sucked into such black > holes if estimate() were changed to reflect only the chance of success.
Indeed. But we wouldn't need to, if we had iTTL. > > > > >>So, before we try some alchemical modification of tRetry, shouldn't we > >>consider using the new computation of estimate() I derived (see > >>http://article.gmane.org/gmane.network.freenet.devel/8143)? > > > > > >If we have to have a tRetry based on failing the whole request, it has > >to be alchemically made enormous. Which suffers from the outlined > >possible attacks. > > Eh? Are you trying to say that my new estimate() is alchemical? It > ain't. According to my derivation tRetry (I call it tGlobalRetry) might > be calculated as enormous (if pGlobalSuccess is sufficiently low) or it > might not be so enormous. It all depends. But the derivation was all > math, so what alchemical stuff are you talking about? No, I am saying that we need to include some alchemy because the consequences of a request going back to the source node with a DNF and failing at the client level are unpredictable and severe, even if we do have some idea w.r.t. retries under certain reasonable assumptions. > > -Martin -- Matthew J Toseland - [EMAIL PROTECTED] Freenet Project Official Codemonkey - http://freenetproject.org/ ICTHUS - Nothing is impossible. Our Boss says so.
signature.asc
Description: Digital signature
_______________________________________________ Devl mailing list [EMAIL PROTECTED] http://dodo.freenetproject.org/cgi-bin/mailman/listinfo/devl
