Hacker News new | ask | show | jobs
by coconut08 1386 days ago
Wait isn't each collection of 32 eth considered a separate validator no matter the source? If that's the case then wouldn't that mean that coinbase and other exchanges make up the massive overwhelming majority of those validators? If that's the case is it really good faith to claim that there are 400+ thousand validators and then arbitrarily put ethereum in first place?
5 comments

Yes, its not genuine to say Eth has 400,000 validators as many of them are the same entity.

It's a similar story for nodes as well. A ton of the reported nodes are actually running on AWS and ultimately its a fairly worthless metric.

For a while the crypto community valued node count as a meaningful number for measuring decentralization, and naturally from that moment forward people have been fully shameless about running hundreds (at some points in history tens of thousands) of nodes just to boost metrics.

> A ton of the reported nodes are actually running on AWS

So? That doesn't mean anything, Amazon doesn't control the nodes just because they are running on its infra.

Because, tomorrow, they could decide to make it against the ToS if they wanted to. Hetzner has already done this.

https://news.ycombinator.com/item?id=32607728

So people move them?

Control of the software is the relevant thing here. If AWS started modifying the software to make it fraudulent that's a problem.

Until that is an issue the hosting provider is as relevant as the bandwidth provider. It's just dumb compute.

Move them where? As we've seen more recently with the OFAC list blocking... once one big player starts to block, the rest follow suit pretty quickly.

Remember also that a lot of the large staking players are staking on AWS and also have other external factors which dictate where they can host (taxes, corporations, shareholders, etc).

Many of the large players are also tied to AWS deployment APIs... moving means rewriting those. Ever worked with terraform before? It isn't just some trivial thing to point at another provider.

> Move them where?

Google Cloud, Azure, deploy a server yourself, etc.

> As we've seen more recently with the OFAC list blocking... once one big player starts to block, the rest follow suit pretty quickly.

There's a big difference between something that is made illegal (eg the OFAC sanctions) and a private action by a company. If your service is illegal then you are going to have other problems than just AWS refusing to host.

> Many of the large players are also tied to AWS deployment APIs... moving means rewriting those.

There's a big difference between using AWS as dumb compute and using AWS features.

The more AWS features you use the more control AWS has. The same applies to any software you use - if a license can be withdrawn there is an element of control. These things are much more important than if the physical machine you are running on is owned by Amazon.

As I said above: Control is the important thing.

Guess what would happen if Amazon would just block certain type of traffic.

US gov or the CIA/fbi just forces AWS to do so.

Since Snowden not unreasonable.

Depends on the custodial setup!
Is this measurable? Is it possible to tell which nodes are running in the cloud, and count them?
At a minimum I would group PoW and PoS coins separately, as the incentives for each category are quite different. This site acknowledges that in a way with separate lists at the top for "nodes" and "validators", but underneath they still have the main list sorted by max(apples, oranges)
More detailed stats on Ethereum validators are here: https://dune.com/hildobby/ETH2-Deposits

Currently there are 79,717 distinct depositor addresses. They list 36 known entities validating, accounting for about 75% of staked ETH. So that leaves about 100K validators split among small unknown entities. Many of those would be home stakers.

Some of the known entities are staking pools, representing lots people who own the stake. But it's only the pool producing the blocks. Bitcoin is similar: lots of home miners but they're mostly in pools, not attempting to create blocks on their own.

For the number of entities creating Bitcoin blocks, I just found this paper, which doesn't give a total number but says 55 to 60 miners control at least half of Bitcoin mining: https://papers.ssrn.com/sol3/papers.cfm?abstract_id=3942181

I wouldn't expect the number of independent (i.e. unpooled) Bitcoin miners to be especially large, since a fairly large investment is required to have a chance of producing a block in a given year.

The point is the shilling as always
1. Yes.

2. Yes.

3. It is not in good faith.