A question I have asked other people before is how to define decentralisation. Radix obviously claims (aspires?) to be decentralised. I thought it would be cool to hear the perspective of those in the Radix community.
Everyone in crypto seems to agree decentralisation is positive/desired, but at the same time it generally remains poorly defined.
Some things that I have seen used to define decentralisation before:
Permissionless
Resistant to state attack
Inability for one actor to take down the network (Whales, government, DDOS attacks etc)
Fair/equitable token distribution
The Nakamoto co-efficient
Most people in crypto would argue that BTC is more decentralised than most/all other crypto, so why exactly is that and can it be quantified in such a way that different crypto projects can be compared as regards decentralisation.
That’s a very good question! I think it’s pretty hard to define exactly what decentralization is. I think the Nakamoto Coefficient is pretty important:
The Nakamoto Coefficient represents the number of validators (nodes) that would have to collude together to successfully slow down or block any respective blockchain from functioning properly
For Radix, that is currently 23 and can be see on @Avaunt’s https://www.radixdashboard.com/.
With Xi’an this number should increase. Currently Bitcoin has the #1 spot with a NC of 7,349.
I think all the factors you listed are part of decentralization, maybe except permissionless. I think a decentralized network can be permissioned, not sure though.
In the sense that not everyone can spin up a node and join the network. You can still have a centralized entity that decides who can become a node and who can’t, but a decentralized (as in, multiple nodes, hosted in various regions with different providers controlled by multiple entities) network.
NC
Don’t quote me on those numbers, but I found this table here:
If the central actor decides who can spin up a node then if it was malicious it could only approve colluding malicious nodes and thereby gain control of the network over time? So in that sense it is a central point of failure? Wouldn’t most consider that to be voiding decentralisaton?
Thanks for the table - I wonder what the NC of ETH is. 23 looks decent in comparison to some of the other L1s
Nice - I’ve been looking for the ETH NC for a while with no success!
That tweet also references this: Herfindahl–Hirschman index - Wikipedia
Although I’m not sure how exactly that’s being applied to crypto projects as it seems designed for corporations.
The other thing I’m unsure of is how bitcoin has such a high score when sometimes people would argue that mining pools tend towards centralisation. Shouldn’t that bring the NC down? Presumably that is because they are applying the NC to BTC nodes, rather than miners, but surely miners perform the consensus so it would be fair to apply the NC analysis to miners?