DagCoin: a cryptocurrency without blocks

[hieuho381]

I believe that time will eventually show that pre-sales simply don't work to do anything but temporarily enrich the launchers.

[33bitcoin]

DagCoin: a cryptocurrency without blocks

Back in 2012 I thought a lot on a new cryptocurrency that could merge the concepts of transaction and block. Each transaction would carry a proof-of-work and reference one or more previous transactions. The resulting authenticated data structure would be a Direct Acyclic Graph (DAG) of transactions where each transaction “confirms” one or more previous transactions. The confirmation security of a transaction would be measured in accumulated amount of proof-of-work referencing (or confirming) the transaction. This structure is well suited for a cryptocurrency without subsidy (such as a side-chain). On the past years I’ve read a couple of similar proposals on bitcointalk (although I cannot find the references now). When the GHOST paper was published, I perceived it as a reinforcement of my idea that a tree could give more security than a chain in case of high rate of transactions.

My open problems…

The problem that I could not solve in 2012 is how to limit the maximum cut of the generated DAG or, in other words, how to prevent all new transactions from referencing the same set of parent transactions. How to create the incentive to “move forward”? The DAG must not increase in “width”, and it should look more like a DAG-chain. Also one must prevent users from choosing old transactions to extend the DAG. I tried several monetary incentive structures to force users to choose newer transactions, but with no result. To know the last “ledger state” there must be a way to consolidate branches. Merging branches should be good, but not too good such that everyone starts merging the same branches over and over. The problem of spam was also less important, as no transaction would be able to get a “free ride” in a block, as each transaction carries PoW. Ultimately the owners of a computer that is being part of a spamming botnet would realize their computers have been hijacked based on the amount of CPU consumed. For instance, if a transaction requires a proof-of-work that takes 1 second in a standard PC, and each transaction is 400 bytes in size, then a botnet consisting in 10K computers may create transaction reaching 3 Mbytes/second. This high network bandwidth usage itself is not a problem, since it can disrupt the network only as long as the attack is active. However, there must be a way to prevent the DAG-chain from growing at that pace. It turns out that the election of an optimal data structure allows the DAG-chain to be compressed, but it requires us to change how we think about double-spends, and how we conceive the “ledger state”.

A Radical Change

The leap of faith required to find an out-of-the-box solution is to think about double-spends not as a boolean attribute, but as a probabilistic attribute, based on comparing the confirmation work on competing transactions. An the security of a transaction, as the confirmation work compared to the the work expected that an adversary may use. Also it requires to forget about the concept of a “global ledger state”. In Bitcoin there is a global ledger state. Chain reorganizations can always rollback the state, but the state is globally consistent. There is a certain probability of the last block rolling back, but the probability is the same for every transaction in that block. In this proposal, the ledger state is just the overlap of all possible transactions, each with its own confirmation probability, and there is no consistent global state.

Design Premise: “The cryptocurrency network benefits from creating a DAG growing as “thin” as possible.”

In other words, having the average maximal cut as low as possible. It seems that referencing many previous transactions (high out degree) can make the DAG thinner only if the following transactions reference the transaction with high out degree, but are themselves of low out degree. So we want high out degree some times, but low out degree another times.

I designed a DAG that tries to fulfill that premise, and an associated incentive structure such that:

 There is a benefit for users to reference as many previous transactions as possible
 Referencing many previous transactions is incentivized only when there are many previous transactions unreferenced.
 There is no competition between users to reference a previous transaction.

Here is the paper draft  –> DagCoin-v4 https://bitslog.files.wordpress.com/2015/09/dagcoin-v41.pdf

This same article can be found in my blog: https://bitslog.wordpress.com/2015/09/11/dagcoin/

to be honest, i really like this but won't that make de centralized concept little more off the edge ?

[Panda Trump]

It sounds nice, but I don't think this would get very far... It's just not a very smooth concept and it would bring nasty disadvantages with it too... ;/

[Panda Trump]

It sounds nice, but I don't think this would get very far... It's just not a very smooth concept and it would bring nasty disadvantages with it too... ;/

could you elaborate? why not very smooth? what disadvantages? it seems like two coins already up and running on this concept.

Could you specify? What coins? Are they exactly the same as this one?

[trout]

are there any other DAG coins now besides IOTA and Byteball?

[nativehasher]

It sounds nice, but I don't think this would get very far... It's just not a very smooth concept and it would bring nasty disadvantages with it too... ;/

could you elaborate? why not very smooth? what disadvantages? it seems like two coins already up and running on this concept.

One of the issues is that these DAG-based approaches give away the benefit of making "blocks" and merkle trees. This means that the ledger side rises linearly with tx count..there is no way to "prune" it..whereas in blockchain you could just store the headers and have merkle proofs to verify block membership.

Other issue is that the resulting structure is unwieldy and exposes lots of additional attack surfaces. Blockchain on the other hand is ideally a single linear chain or in the worst case some orphan blocks and a few forks..which makes it easier to analyze.

[fan_of_things_and_stuff]

This sounds really interesting, but my biggest concern is that the individual user actually loses out by being forced into mining. Right now, users of a blockchain pay a fee to miners for their services. As a result, big business has developed around mining as cheaply as possible to maximize profits. And this works out well for the end user because they don't have to worry about owning a mining rig or paying the cost in electricity that would be required to run all the hash attempts.

What happens when every person has to mine their own transaction? First of all, the user will be on the hook for that cost in electricity/mining hardware cost. Second, their transaction is dependent on them being able to solve someone else's unreferenced transaction. Doesn't that end up being cost-prohibitive?

[Vigens]

This sounds really interesting, but my biggest concern is that the individual user actually loses out by being forced into mining. Right now, users of a blockchain pay a fee to miners for their services. As a result, big business has developed around mining as cheaply as possible to maximize profits. And this works out well for the end user because they don't have to worry about owning a mining rig or paying the cost in electricity that would be required to run all the hash attempts.

What happens when every person has to mine their own transaction? First of all, the user will be on the hook for that cost in electricity/mining hardware cost. Second, their transaction is dependent on them being able to solve someone else's unreferenced transaction. Doesn't that end up being cost-prohibitive?

I have this same question. Thats the only reason i havent bought dag yet.

[Pedagang]

Here is the announcement thread of a token based on DAG tech.

here's the visualization of it https://tangle.blox.pm/

[Pedagang]

Have you looked at https://byteball.org what is your opinion on that?

Yes, would be interesting to get an opinion because...

...I'm still not convinced that Byteball is a pure DAG coin, to prove my position I would need to generate a lot of transactions on Byteball network to show that in certain conditions (related to DAG topology) TPS growth is negatively impacted by necessity to pick the main chain. If you compared Ethereum (which calls itself blockchain) and Byteball you would see that they don't differ much:

You keep spamming that image.

The best image of Byteball is this, an actual DAG.
https://ip.bitcointalk.org/?u=http%3A%2F%2Fi.imgur.com%2Fwf22yutl.png&t=576&c=RbxRGi7jK-Tm9w

Byteball is the first DAG-coin, the first! The first IoT coin, and the first on exchanges and actual use in livenet.

what's with the first thing? LoL

what a shill

[tendermartin]

Is your system a Quorum Based Qubic system???

Very interesting. I'm coding a very similar system right now. 80% is already done.

[tendermartin]

Exactly...

this is the reason i have reservations with dag.

This sounds really interesting, but my biggest concern is that the individual user actually loses out by being forced into mining. Right now, users of a blockchain pay a fee to miners for their services. As a result, big business has developed around mining as cheaply as possible to maximize profits. And this works out well for the end user because they don't have to worry about owning a mining rig or paying the cost in electricity that would be required to run all the hash attempts.

What happens when every person has to mine their own transaction? First of all, the user will be on the hook for that cost in electricity/mining hardware cost. Second, their transaction is dependent on them being able to solve someone else's unreferenced transaction. Doesn't that end up being cost-prohibitive?

[rpietila]

What is reason that one needs to buy these?

(The optimal way to get anything bootstrapped is a free spinoff, eg. a cryptocurrency receives the highest marketcap if it is given for free according to the amount one paid personal (fiat) taxes last year... The previous would give everyone according to their ability to spend, and propel any reasonably transactable token to billion$ of marketcap in an instant, without moving any other currency in the bootstrapping process.)

[d5000]

Having observed Iota a bit and being a bit disappointed by Byteball, I must say that I haven't understood the exact advantages of a DAG with respect to a blockchain, at least of a DAG based on the "transactions as blocks" paradigm like Iota and Byteball (instead, a limited DAG-based approach like in Ethereum seems to have some advantages).

Iota cites scalability (mainly, as far as I understand it, because of fast transaction propagation/validation), but small-block blockchains should have the same advantages. Iota's zero-fee paradigm seems specific to its design and not to a DAG (Steem, for example, being blockchain based also has zero fees). And then there is the issue to get transactions confirmed that looks pretty difficult in Iota. In both Iota and Byteball, as far as I know, also still centralized servers work as the "clock" of the system.

So I would be really grateful if someone could explain the advantages of a DAGcoin and what is its real potential. (This thread also was not helpful, but I have some hope that in this subforum there are more technically skilled people than in the altcoin forum.)

Also, I have heard also that research is going on to integrate the concept into Bitcoin in the far future. If someone knows more, I'd be grateful for links about that.

[Cconnected]

https://github.com/DagcoinOY/dagcoin

[el3f]

Do you have exchanger for Dagcoin?

[d5000]

Do you have exchanger for Dagcoin?

DagCoin is not an altcoin. It is a concept that originally was meant to be used as a Bitcoin sidechain.

As far as I know, the original idea was never transformed into a working cryptocurrency. But with Byteball and Iota there are two altcoins attempting something similar.

I am strongly in favor to limit the scope of this thread to the "DAG" concept in general and the original DAGcoin implementation, and not discuss Byteball and Iota here, as they are discussed in several threads in the Altcoin subforum.

https://github.com/DagcoinOY/dagcoin

First I thought, "Interesting." But then I read this:

Dagcoin uses byteball network as an underlying platform.

So it's a second-layer currency built on an altcoin - not really what we're discussing here.

[snakes-r-us]

What is the present list of DAG based ledgers? I could find tangle/iota, byteball and this. I'd like to research more into this, so any additional references would be helpful.

For completeness' sake there's also XRB - RaiBlocks. Very cool distribution model btw - proof of reCAPTCHA )

Resources worth checking out:

ANN: https://bitcointalk.org/index.php?topic=1381323.0
XRB vs BTC vs XRP: https://bitcointalk.org/index.php?topic=1381323.msg19698220#msg19698220
Github: https://github.com/clemahieu/raiblocks

[silverkamote]

Cool! I was working on a similar problem last year. My approach was to have txs delivered in rounds (also with PoW, in a DAG), such that low value txs happen early in the round and higher value txs happen later in the round (and must point to previous, lower value txs in the round). A “highest value tx in this round” acts as a shelling point to “end the round”, ie. new low value txs start pouring in on top of it for the next round. You’d want to get at least a few rounds of confirmation before accepting a transaction, and an interesting auction economics emerges for rolling into the next round. A more complex variation would target the distribution of tx amounts directly, such that the modes happen early in the round, and the tails end the round.

[Soots]

sir do you have another thread about the information of this project you we can read or signature campaign?