Could DPos and POA be briefly explained so I know what I'm voting for?
DPoS would basically be PoS with preselected validators which have to be online.
If a validator is not online, the block will be skipped.
DPoS is chosen instead of PoS to have a minimum weight for every participant, which ensures that the block creator will be online.
In the beginning, I would like to set the minimum to 0.1%, which of course would be lowered later on.
PoA is PoW combined with PoS, to ensure that there are no 51% attacks and no orphans. First one will randomly find a proof of work block and submit it. Then 100 previously selected (dependent on the previous block) validators will sign the block, and if one of them did not sign the block, the block will be invalid and you will be forced to wait for the next one.
PoA would be much more difficult to implement, add more garbage that has to be sent from one node to another. It would drastically increase the time needed for each block to be confirmed, implying that the block time has to be increased from 15s to 2 minutes or more, increasing the average confirmation time to approximately six minutes, assuming just 1% of the nodes are offline; with 10% of the nodes being offline, it would be one block every 26 days.
On the other hand PoA would be the only way of having not just a decentralised, but also a distributed consensus.
As you might see, I, as the developer and visionary, definitely vote for DPoS, because this is much better for the future of the network.
Unless someone has a better solution, I will vote for DPoS, but I can also see why other people would vote for the only way of having a distributed consensus that's resistant against orphans.
Thanks
DPos sounds a little too centralized, but I do understand the selected validators are needed to ensure the network is secure when using super fast block times.
Would it be possible to for the network to randomly select which nodes are validators periodically, for a determined number of blocks, then repeat?
Such nodes would need to qualify in some way to validate, ( time online, balance, connectivity etc).
Is the proof of work in your version of POA really needed?
In pure POS currencies you would need 51% of the coins to attack the network, why would you want to attack something you have a 51% stake in?
This version of POA sounds a little too complicated.
Can bullions POS algorithm not be used and tweaked to how you need it? I like what was done with it, and it works really well.