When trying to make rough estimations about the POI
https://forum.ournem.com/index.php?topic=2618.msg9009#msg9009
it looks like there are many factors, which affect to the importance and harvesting.
Eg. if someone keeps a shop in net, he is causing a lot of transactions. For him those are 'incoming' and possibly he is sending only quite few or even zero NEMs.
But still he has an important effect to the NEM network - and has earned the importance, right?
That leads to the robots, which are now used for stress testing.
The robots are good for testing, how the network tolerates stress and data load, but maybe they will give wrong picture about the POI and harvesting:
- robots are sending NEMs to many accounts
- those accounts are treated like shop keepers and
they are getting too high importance
- too high importance may give a too good picture about the harvesting
Should the stress testing (with robots) be done only for some time per each day ?
Or only 1 - 2 days and then 1 - 2 days' pause?
Beside that once NEM is released the question is how much importance will exchanges gain
because people will want and start to trade NEM and they will deposit
big amounts of nem on exchange account. The exchange will make a lot of transactions
and will have a huge amount of NEM in the account therefore it will for sure gain a lof of importance.
What I' m afraid of is that the exchanges will become to NEM the same as the
mining pools are to POW coins and this could lead to centralization .
Beside that once NEM is released the question is how much importance will exchanges gain
because people will want and start to trade NEM and they will deposit
big amounts of nem on exchange account. The exchange will make a lot of transactions
and will have a huge amount of NEM in the account therefore it will for sure gain a lof of importance.
What I' m afraid of is that the exchanges will become to NEM the same as the
mining pools are to POW coins and this could lead to centralization .
That problem would exist for POS as well but it hasn't popped up so far. Good exchanges store the majority of funds in a cold wallet so they won't forge/harvest with those.
The majority of transactions that happen in exchanges are off-blockchain so they might not gain as much importance as you'd think.
When trying to make rough estimations about the POI
https://forum.ournem.com/index.php?topic=2618.msg9009#msg9009
it looks like there are many factors, which affect to the importance and harvesting.
Eg. if someone keeps a shop in net, he is causing a lot of transactions. For him those are 'incoming' and possibly he is sending only quite few or even zero NEMs.
But still he has an important effect to the NEM network - and has earned the importance, right?
That leads to the robots, which are now used for stress testing.
The robots are good for testing, how the network tolerates stress and data load, but maybe they will give wrong picture about the POI and harvesting:
- robots are sending NEMs to many accounts
- those accounts are treated like shop keepers and
they are getting too high importance
- too high importance may give a too good picture about the harvesting
Should the stress testing (with robots) be done only for some time per each day ?
Or only 1 - 2 days and then 1 - 2 days' pause?
Do we want to make useful POI testing now with the Beta?
If yes, then too active robots (too many TX's from A to B) will bias the results of POI testing.
If robots are used, then there should be more circulation:
TX's from A to B, C, D, E ...
TX's from B to A, C, D, E ...
...maybe there is not enough accounts ... or robots.
I would participate, but 0.4.20-BETA is still having performance problems (too many open file handles), so the software crashes on my cubieboard every few hours.
I would participate, but 0.4.20-BETA is still having performance problems (too many open file handles), so the software crashes on my cubieboard every few hours.
... corrected in 0.4.28 ?
The stress testing with robots should be planned ... by centralized manner :)
And then let some space for the "normal" behaviour; we could simulate shops etc. by sending NEMs to some certain accounts.
Otherwise the estimations about POI are quite near about good guesses.
I just updated to 0.4.28 and am testing now. I will inform you about how its working now.
What I noticed: The node was booting in like 3 seconds. I didn't understand that^^
Works really good. My robot is doing transactions every 20 seconds.
My node: mixmasters node
My address: TCXXVB-FMSJ2W-UA6FH6-AV5SUG-YGU6WN-GXWR27-X523
Works really good. My robot is doing transactions every 20 seconds.
My node: mixmasters node
My address: TCXXVB-FMSJ2W-UA6FH6-AV5SUG-YGU6WN-GXWR27-X523
C'mon...pump it up :)
Try to do more and more tx per second until something breakes. The only way to find limits is to go beyond.
I already increased the tx/s from 1/60 to 1/20. I will increase more, but that is only showing the limit of my cubieboard, not the network.
Whe I get home I'll make some ajustments.
My desktop can handle really fine ( APU 7850K AMD + R9 290), but my laptop doesn't like so much transactions below 10 seconds interval.
I'll see how it accepts!
Have a nice day!