[EGreg]

There are lots of other approaches - IOTA DAG, HashGraph, Ripple Consensus Process etc.

I am not a fan of blockchains, though. They are overkill for most uses. But here is an example of a non-blockchain system that doesn’t even require global consensus:

http://intercoin.org/technology.pdf

Also check out the Autonomi network

[cyberax]

It's not really going to work. Without a centralizing consensus, any such scheme is vulnerable to be drowned in forks.

A malicious notary network can simply flood the ledger with conflicting views. So clients will have to somehow find a set of notaries that is the "best".

Proof-of-stake means that there's effectively a vote on the set of "reliable" agents, and proof-of-work works because the malicious notaries can't outrace everyone else.

[EGreg]

Sorry, but you're not exactly an expert on this. There is a huge body of literature that says otherwise, and reference implementations.

You don't NEED "a centralizing consensus". IOTA did have one, called a "governor". And now they also did away with it.

I had a discussion with the CTO of Ripple (back then their chief cryptographer) David Schwartz about this exact issue in 2018, when I was also connecting with Leslie Lamport and others in the industry to discuss why and how global consensus was even needed

https://community.intercoin.app/t/intercoin-technology-conse...

You can also read this post here: https://community.intercoin.app/t/intercoin-technology-recov...

At the end of the post, it links to the mathematical results on arxiv if you're interested: https://arxiv.org/pdf/0802.0832v1

[cyberax]

> Sorry, but you're not exactly an expert on this. There is a huge body of literature that says otherwise, and reference implementations.

Yep, and none of them managed to solve the issue of resiliency without some kind of a stake.

From your own link:

> Intercoin’s ledger technology requires the sender to endorse a transaction after a supermajority of validators have approved it.

> Validators periodically check one another with “proof of resource” techniques.

Basically, it just moves the problem of validating individual transactions to validating the set of trusted notaries via proof-of-stake.

Just another rehash of crapcoin bullshit.

[EGreg]

There is no “the set” of trusted notaries

Each coin has a few notaries — not the majority of the network. Did you even glance at all at the arxiv paper?

[cyberax]

Yes. How do you find a notary that is not malicious?

A malicious subnetwork of notaries can flood it with bogus transactions. To prevent that, you have to make sure that transactions can't happen without a significant expenditure of real-world resources.

[EGreg]

That's not true. Notaries shouldn't be able to flood a network at all. Each participant in the network is supposed to stop accepting messages from a malicious participant. It's one strike and you're out.

That's what Proof of Corruption is about, in our technology, for instance.

Every participant has to sign their claims. If a participant signs two contradictory claims, this Proof of Corruption can be gossipped and the participant is excluded.

In distributed systems, 99% of the time you have finality, but 1% of the time you may have a conflict due to race conditions or corrupt nodes, etc. Blockchains take the unfortunately brute force approach of gathering all the conflicts / ambiguities into one or another consistent chain of transactions, and then "duking it out" with a lot of expensive "stake" or "work" or whatever. But it doesn't have to be that way. The end-users are ultimately the ones to either endorse a transaction or not, there is no reason to have the network be the source of truth for the remaining 1%, there are hugely diminishing returns from all that waste of electricity. So even the double-spend problem can be solved without blockchains.

But even without this, in other decentralized architectures such as the PTN (Permissionless Timestamping Network) I linked you to above (https://intercoin.app/technology.pdf) there is no blockchain, no consensus, just nodes talking to each other and data structures accruing in eventually-consistent ways. And the nodes can just as easily stop listening to you and forwarding your messages.

Similarly in the SAFE network. Even the routing is done in a way that the routing info is deleted after one hop, so you can't DDOS the network the way you can a regular IP / BGP network or even a regular DHT (such as Bittorrent's Mainline DHT). Because the nodes will just refuse to pass on your message. Every node expends only the resources it is prepared to, and nothing more. This idea of "flooding" or "DDOS"ing is more of a legacy idea due to the federated systems we have today, like email and DNS (where the whole world can spam a person's email, and you play cat-and-mouse).

Again, blockchain is a tiny part of this space of decentralized networks. You can have CRDTs syncing, or you can have append-only logs such as Hypercore (now called Holepunch / Pears) or you can have Freenet (the new one, I interviewed the founder a couple years ago when it was still called Locutus: https://www.youtube.com/watch?v=yBtyNIqZios) you can have Secure Scuttlebutt, or Nostr etc. etc. etc.