zzz
orignal, a quick clarification on how Java treats fU routers:
zzz
I don't think I explained it very well the other day
zzz
you asked if Java treats fU as floodfill, I said yes, that's correct
zzz
however:
zzz
due to profiling, ff selection policies, netdb ejection policies, and peer selection policies, we very rarely do lookups/stores to fU routers
zzz
so in practice, we mostly exclude them
orignal
so my question is
orignal
do you count them as floodfills or not?
orignal
e.g. I'm still trying to understand the difference in number of floodfills between Java and i2pd
orignal
say I have 700 you have 1200
orignal
and second question why you don't like Uf floodfills
zzz
1) yes we count them in what's reported in the router console
zzz
2) it's not a 'don't like', but U routers in general get evicted from the netdb faster, and don't build up a high-rated profile
zzz
3) we aggressively expire known routers if we have more than 4000 total, and i2pd regularly reports 10-20K knwonRouters, so I don't think you can directly compare java and i2pd floodfill counts
orignal
2. how about flood?
orignal
1. do you know how many "real"(according to profiles) you have in netdb?
zzz
1) I don't track real vs. not real but here's a current count of ffs by caps:
zzz
634 XfR
zzz
101 NfR
zzz
66 XfU
zzz
63 PfR
zzz
44 XfRE
zzz
25 XfRD
zzz
13 PfRE
zzz
11 OfRE
zzz
9 PfRD
zzz
9 OfR
zzz
2 XOfR
zzz
2 XfRG
zzz
2 PfU
zzz
2 Pf
zzz
2 OfRD
zzz
2 LfR
zzz
1 PfUE
zzz
1 OfUG
zzz
1 OfRG
zzz
1 NfRE
zzz
1 NfRD
zzz
2) flood selection is the same algorithm as lookup/store selection
orignal
66 XfU why are they real?
orignal
never mind