The million dollar question here seems to be:
The one who is loyal to the net and reliable (identified by key pair) should get more to do with bigger possibility from the network.
How is loyalty and reliability determined?
Bitcoin's methodology is to fix the value of a person's word to the computing power they can throw at a "hard" computational problem. The idea is simple: any
single "evil" user must outvote the computing power of
all the "good" users in order to hijack the ledger. Hence, the ranking is purely to make it prohibitively difficult for any single "evil" user to get strong enough to do damage.
The idea behind your proposal seems to be, let's not use proof-of-work, it's wasteful and pins voting power to "people who can afford hardware". Let's find some other way to verify that the "votes" will be mostly "good" ones. It seems you want to do this by saying, okay, some nodes are more trustworthy "loyal and reliable" than others, so the "good" nodes have more voting power than even millions of nodes made by an attacker.
But that's easier said than done. What determines loyalty and reliability? If it's anything that is weak to social engineering, you're sunk. If it's anything where an attacker could lie low and act "respectable" for five or ten or fifteen years and then steal millions, you're sunk.