fix verification network behaviour:

[UdjinM6]

• use up to 10 connections only
• save state (do not ask same nodes after restart)
• initiate on time based intervals (every 5 minutes) instead of new blocks (should keep connections alive longer)

This address complains that some vps providers consider such activity (a lot of very short-life connections) a ddos-attack.

[thelazier]

I found that we will send verify more than one to some nodes.

Ex: Rank#1 starts sending from Rank#11 and Rank#2 starts sending from Rank#21. 
If there Rank#14 is verified and get skipped by verified/banned,
then Rank#1 will also send verified to Rank#21 which is already verified by Rank#2.

My solution is ...

Rank#1 starts sending from Rank#11,
then next to Rank#n+MAX_POSE_CONNECTION (#21,#31,#41,#51,...),
until nCount = MAX_POSE_CONNECTION. 

This way Rank#1-#10 will not send verify duplicate with other Rank#, IMO.

[UdjinM6]

Actually no,

if(netfulfilledman.HasFulfilledRequest(addr, strprintf("%s", NetMsgType::MNVERIFY)+"-request")) {

should take care of recent connections.

And mWeAskedForVerification is serving different purpose, reverted serialization.

EDIT: but thanks for bringing my attention to this again :)

[thelazier]

Actually, I don't mean that a node will duplicate send to same MN for verify again, I mean node A on Rank#1 and node B on Rank#2 could send verify-request to node C on Rank#21 in same period of time. So, node C need to calculate verify response twice while it should calculate only once for verify-request from node B :)

[UdjinM6]

Ah, I see now :)

[tgflynn]

utACK