Blockchain vs. Block Lattice

[Treney]

I'm interested to hear people's opinions on this.

What are the advantages of using a blockchain over a block lattice? Neither has been "hacked", at least at this point, and the block lattice seems to just be way faster (at least with my experience using Nano), and there aren't any confirmations etc so you can use the funds immediately.

Is there any reason to continue using blockchains in the future rather than try to build new coins on a Lattice architecture?

[1714768622]

1714768622

[1714768622]

1714768622

[d5000]

There are many misunderstandings regarding the "block lattice" model and Nano in particular.

Nano has, in reality, two confirmation times: one "in the case that everything goes well", which is nearly instant (1-2 sec.), and another one, which is "in the case of a conflict", which is about ~1 minute. This is due to the fact that when the DPoS algorithm detects a conflict in the DAG, then the nodes proceed to vote which transaction goes through.

But as you can never be sure that a conflict won't appear some seconds later - maybe a "legit transaction" is only late because of latency issues - then it's highly risky to accept a transaction with an "instant confirmation". It may get rolled back.

The hyped instant confirmation at Nano has, thus, equivalent security to the 0-confirmation-transaction in Bitcoin, which is also (almost) instant!

DAGs like the "block lattice" are interesting technologies because the propagation of a "confirmed transaction" in theory could be faster than with blockchains, but all models I know currently (IOTA, Nano, Hashgraph, Byteball) struggle with the double-spending problem and have a centralized (IOTA) or semi-centralized (Byteball, Nano, Hashgraph) security layer consisting of a set of trusted or "delegated" nodes doing the hard work (double-spend checking).

If DAG research, however, should show that there is a viable security algorithm then I would have no problem with Bitcoin changing to a DAG (It will be a hard battle, however, if we take into account the ideologized and "flamy" discussions about Segwit.).

There are also blockchains with a "block interval" of less than a minute, like Ethereum or EOS. But their one-confirmation-transactions are far away from the security Bitcoin's confirmation provides.

@locsta123: Block lattice is a kind of "DAG" used by the Nano cryptocurrency. It is "for real" in the sense that it exists. but if you're thinking that it provides similar security with a higher transaction rate and instant confirmation, you're terribly wrong.

There's until today no cryptocurrency - blockchain-based or DAG-based - that can rival Bitcoin's security.

[pebwindkraft]

What is block lattice tech I have never heard of it? Is this for real?

in English:
https://nano.org/en
in French:
https://courscryptomonnaies.com/nano

A DAG based system, which replaces the TANGLE from IOTA with a system called "block lattice".
This is how far I could get from a quick look, just reading 

As d5000 layed out, the double-spend problem is not really solved (and PoS system based is centralized, aka against the idea of Bitcoin), I would see this more related to IOT and similiar, than for "Bitcoin replacement".
probably s.th. which belongs into the altcoin section.

Edit: there is a PoW component: "Unlike Bitcoin, the PoW in Nano is simply used as an anti-spam tool, similar to Hashcash".
Whitepaper is here: https://nano.org/en/whitepaper

Edit2: and the technical discussion is here: https://bitcointalk.org/index.php?topic=1219264.0

[chrissyg1]

There are many misunderstandings regarding the "block lattice" model and Nano in particular.

Nano has, in reality, two confirmation times: one "in the case that everything goes well", which is nearly instant (1-2 sec.), and another one, which is "in the case of a conflict", which is about ~1 minute. This is due to the fact that when the DPoS algorithm detects a conflict in the DAG, then the nodes proceed to vote which transaction goes through.

But as you can never be sure that a conflict won't appear some seconds later - maybe a "legit transaction" is only late because of latency issues - then it's highly risky to accept a transaction with an "instant confirmation". It may get rolled back.

The hyped instant confirmation at Nano has, thus, equivalent security to the 0-confirmation-transaction in Bitcoin, which is also (almost) instant!

DAGs like the "block lattice" are interesting technologies because the propagation of a "confirmed transaction" in theory could be faster than with blockchains, but all models I know currently (IOTA, Nano, Hashgraph, Byteball) struggle with the double-spending problem and have a centralized (IOTA) or semi-centralized (Byteball, Nano, Hashgraph) security layer consisting of a set of trusted or "delegated" nodes doing the hard work (double-spend checking).

If DAG research, however, should show that there is a viable security algorithm then I would have no problem with Bitcoin changing to a DAG (It will be a hard battle, however, if we take into account the ideologized and "flamy" discussions about Segwit.).

There are also blockchains with a "block interval" of less than a minute, like Ethereum or EOS. But their one-confirmation-transactions are far away from the security Bitcoin's confirmation provides.

@locsta123: Block lattice is a kind of "DAG" used by the Nano cryptocurrency. It is "for real" in the sense that it exists. but if you're thinking that it provides similar security with a higher transaction rate and instant confirmation, you're terribly wrong.

There's until today no cryptocurrency - blockchain-based or DAG-based - that can rival Bitcoin's security.

This is an interesting analysis but actually incorrect - your post has been linked in the Nano subreddit where the developer has responded indirectly to your claims. You can check it out here - https://reddit.com/r/nanocurrency/comments/9qdpkw/relevance_of_comment_on_nanos_instant/

[bob123]

There are many misunderstandings regarding the "block lattice" model and Nano in particular.

Nano has, in reality, two confirmation times: one "in the case that everything goes well", which is nearly instant (1-2 sec.), and another one, which is "in the case of a conflict", which is about ~1 minute. This is due to the fact that when the DPoS algorithm detects a conflict in the DAG, then the nodes proceed to vote which transaction goes through.

But as you can never be sure that a conflict won't appear some seconds later - maybe a "legit transaction" is only late because of latency issues - then it's highly risky to accept a transaction with an "instant confirmation". It may get rolled back.

The hyped instant confirmation at Nano has, thus, equivalent security to the 0-confirmation-transaction in Bitcoin, which is also (almost) instant!

[...]

This is an interesting analysis but actually incorrect - your post has been linked in the Nano subreddit where the developer has responded indirectly to your claims. You can check it out here - https://reddit.com/r/nanocurrency/comments/9qdpkw/relevance_of_comment_on_nanos_instant/

Regarding the statement about two confirmation times: From a technical standpoint, it is wrong.
Propagation of votes in this specific block-lattice-based coin does happen within a few seconds. So there can't be a conflict happening within the next minute.

The only case where this could be a feasible attack would be when the majority of the network does not vote (or is seperated from the rest of the network) and then after a given time (e.g. 30 seconds) does publish a countering vote.
But this could easily be spotted if you look at the vote weight and if you have a full node running.


So.. in the first few seconds there is a decision whether it is confirmed (almost 'instantly'; up to 3 seconds) or will take up to 1 minute until there might be a confirmation.

But a scenario where a transaction is confirmed after 5 seconds, with opposing votes 20 seconds later can not work.

You might as well read the link posted by chrissyg1, which probably leads to the same conclusion.

[d5000]

Regarding the statement about two confirmation times: From a technical standpoint, it is wrong.
Propagation of votes in this specific block-lattice-based coin does happen within a few seconds. So there can't be a conflict happening within the next minute.

Yes, you're right - I never meant it that way.

The point is that if a conflict appears in the first seconds - e.g. an attacker sends one transaction to one half of his connected nodes, and a conflicting transaction just 2-3 seconds after to the rest of the network - then the voting process can be longer - at least it's no longer "instant".

It may be confirmed "enough" for a merchant to accept it in significantly less than a minute (as it seems to be the case, according to the Reddit discussion), but this is mainly an "achievement" of the semi-centralized DPoS system, and not of the block lattice/DAG.

So.. in the first few seconds there is a decision whether it is confirmed (almost 'instantly'; up to 3 seconds) or will take up to 1 minute until there might be a confirmation.

But a scenario where a transaction is confirmed after 5 seconds, with opposing votes 20 seconds later can not work.

Yep, I also understood it that way. The "two confirmation times" part of my post was simply trying to describe what you say in your first sentence.

I want to clarify that I'm not a Bitcoin maximalist (although I do support mainly coins which follow the Bitcoin model - i.e. decentralized coins not distributed via ICOs) nor do I have something against Nano. It only annoys me that some Nano users try to market the coin as an alternative to Bitcoin with LN, which it is not, because confirmations are not always instant and the security model is drastically different and much more centralized. Also, the myth of "infinite capacity" is basically the old big blocker argument. In a scenario when Nano gets used more, then conflicts will be also more likely, so the instant confirmation time will be matter of luck.