Re: Faulty node lookup in LimeWire

"pwang708" <[email protected]> Thu, 28 Aug 2008 18:27:23 -0000
Newsgroups gmane.network.gnutella.devel
Message-ID <[email protected]>
--- In [email protected], Raphael_Manfredi@... wrote:
>
> Here are more statistics about "foreign storage":
> 
> I ran a node for 20 hours during which I got 4475 keys stored.
> Out of these, only 1683 fell within my k-ball.
> 
> That means 62% of all STORE requests come to my node when it is not
> the ideal target for the key, and actually not within the k-ball for
> that key.
> 

From my data, I computed the common prefix length between my ID and
keys in the STOREs. Look like most STOREs sent to me were OK. But I
agree that we need to speed some time to find out what happened with
those STOREs with less common prefix length, although a small percentage. 

Common prefix length >= 16, count =  57181
Common prefix length  = 15, count =  92628
Common prefix length  = 14, count =  92912
Common prefix length  = 13, count =  17476
Common prefix length  = 12, count =  656
Common prefix length  = 11, count =  43
Common prefix length  = 10, count =  28
Common prefix length  = 9, count =  50
Common prefix length  = 8, count =  57
Common prefix length  = 7, count =  24
Common prefix length  = 6, count =  23
Common prefix length  = 5, count =  33
Common prefix length  = 4, count =  26
Common prefix length  = 3, count =  27
Common prefix length  = 2, count =  9
Common prefix length  = 1, count =  16
Common prefix length  = 0, count =  26


Best,
Peng