Martin Stone Davis wrote:

Ian Clarke wrote:

Ian Clarke wrote:

Just browsing through my routing table status servlet and noticed something worrying - a backed-off node whose last estimate was -27143ms. Presumably we shouldn't be getting negative estimates for routing times.



Hmmm, just found another 6 of them in my RT - it looks like this is actually quite common. BTW I am running 6338.


Ian.


I think the reason is:

estimate=(pDNF - pLegitDNF) * (tDNF + requestFailTime)
          + pSuccess * tSuccess
          + pTransferFailed * (tTransferFailed + requestFailTime)

and in the negative cases, pDNF<pLegitDNF.


The perverse consequence of this is that we'll favor nodes with larger tDNF:s! Obviously, the simple fix is:


 estimate=MAX(0,(pDNF - pLegitDNF)) * (tDNF + requestFailTime)
          + ...

I notice that for many of the ones with negative estimates, overall prob of DNF is "0".


-Martin



_______________________________________________ Devl mailing list [EMAIL PROTECTED] http://dodo.freenetproject.org/cgi-bin/mailman/listinfo/devl

Reply via email to