Topic: Obyte: Totally new consensus algorithm + private untraceable payments - page 12. (Read 1234253 times)

That's all they have given as a pledge? Why they have to be interested in witnessing? What are their independent benefits from that?

They are called Order Providers, right? They're designed to keep an order. Disorder means devastating for the network stability the utilizing which is impossible without. For that reason they must be independent but they aren't by now, right? They might will have got independent in unknown time in the future... Their amount is 13 and 6 of them represent Tony personally, right?

So, I'm not going to rant any more. Sorry, if I bothered you.
You guys have done a great work, but seems like there is a fundamental mistake in those rules. I'm not going neither to strict them nor to teach you. You know better about true decentralization.

Obviously, one shows interest to become a Order Provider when they are interested that the network keeps running (transactions become stable). They have also pledged that if they lose interest then they let the community know, so they could be replaced without downtime. So far, none of the new Order Providers have been replaced, but each new Order Provider has been replacing the old founder's address with the new Order Provider's address.

Order Providers that are being used by users in almost real-time, can be seen from there
https://stats.obyte.org/witnesses.php

Candidate interviews and poll result can be seen from there
https://medium.com/obyte/decentralization/home

Not all candidates had polls, some earlier candidates were only chosen when stats.obyte.org showed support for new candidate, but even with polls, Order Providers replaced their witnesses when users replaced theirs in the app.

What transactions need to be turned back? Order Providers ARE NOT gatekeepers who mark one transaction valid or other invalid - all full nodes do that and they all follow the same rules in the code (if you change the rules, you hard-fork). It does not matter whose transactions they choose parent because it will cause all the other transactions below it to become stable eventually too.

There is no re-orgs or rollbacks, once certain part of the DAG is stable then all those transactions are confirmed, but it's not because somebody marked them confirmed - they are confirmed because order of transactions in that part of the DAG cannot change anymore (majority of Order Providers have posted and cannot have best parent in that part of DAG).

None of these rules have picked randomly, they are there for reason:
* fixed number of Order Providers - whether it's 6, 12, 24 or 100, it would be so much more difficult if each user could have variable number of Order Providers and it would have 0 benefits.
* 12 Order Providers - that's a number big enough to avoid failures, but small enough so users could remember those entitites. there can be unlimited amount of Order Providers on the network because each user can change one in their list if the 11 other stay the same as others.
* publicly known entities for Order Providers - if they are not identified, then attacker could spawn hundreds of Order Providers and announce hundreds of candidates. if majority of even 200 total Order Providers are by single attacker, you have Sybil attack.
* if there is Sybil attack or majority of different Order Providers collude then they can either do a invalid transaction, which will stop the network (unless there is hard-fork to change that rule) or just stop running the nodes (no more stable part of DAG) or not replacing Order Providers based on how users signal them.

In a way, it has similar balance like there is between Bitcoin full nodes and Bitcoin miners. Both have to run the same software, because if they don't then once produces blocks to full nodes, which nobody else but them run, but at the same time full nodes with different software doesn't get any blocks because nobody runs miner for that fork.

On Obyte, there is no middlemen who control what can go to block, but Order Providers and have the same relationship with users, one can't be without the other. Neither of them have the absolute control.

I'm not quite sure if I'm right about it, but I can't then realize what they're running witnesses for if it's that disadvantageous? I can't believe in they're just bunch of altruists, aren't they? How many witnesses have been replaced with new ones that way since the project got started, btw?

Regarding to the witnesses, I'm not going to argue... For now, I'm still have my own opinion about the risks. I agree that community can hard-fork etc, but they, unfortunately, won't turn the transactions back if that collapse had suddenly happened. I also agree that new witnesses have to grow 1 by 1. But I disagree to the point that old witnesses have to approve new ones to add... and that personal identity thing really has to play any significant role in consensus. I will learn more about it.  

Thanks a lot, anyway.

Payload fees are divided by all the witnesses that posted in last 100 MCI.
Header fees go to any address who chose the transaction as parent.
Again, these fees are not a big deal, they barely cover the new transactions that witnesses make. Or if the witness is also oracle then they don't even cover those, but since making transactions on Obyte is so cheap, $1 worth of bytes lasts for 50 000 transactions.
Running a witness node is not done for earning $1 of bytes in maybe 10 years. If any at all.

Your assumption that witnesses would not want to give up their role is not correct and not even remotely valid. Witnesses are replaced 1 by 1 gradually, it takes time and only majority is needed.
And on top of that, if that doesn't align with what users want, community can hard-fork the shitty behaving witnesses out.

The point might be the commission divides equally among witnesses?
Thanks for your answers.

No, if you have 120 anonymous witnesses and majority of them are actually the same person, you have Sybil attack. If all the 12 witnesses have been identified then there is no Sybil attack because you know that each one is unique.



Not sure if you are now trolling, this was already answered. I am sorry if you still don't get it.


But you were not right, you made it up.
And witnesses don't decide stability point, it's just that once they have posted enough transactions on top of other transactions that they know about, the stability point appears on unit, which order is impossible to change without breaking the rules. Everybody can do that, witnesses just add their transactions also there. Witnesses can't exclude any transactions because others will choose these transactions as parents. If they would want to censor one transaction, they would need to stop the posting altogether and censor every transaction, which results a hard-fork by community again. I think you skipped reading this https://medium.com/obyte/dag-vs-blockchain-6d2d99f10bd9

There is no such thing as shitty stability point, the witness who doesn't run their node can be considered shitty because they are not doing what they were suppose to do - run the node.



Not sure how that proves your point that witnesses have no reason to choose a new witness? The data on DAG proves that they have done exactly what you say they would not do.
Also, you are forgetting that it's users who signal the witnesses by changing their witness list in their wallet app. If majority of the witnesses don't change the witness that users want then the community can always hard-fork the project.

Feels to me that you are trying to apply your blockchain understanding of how consensus works onto DAG consensus and you are assuming things.

Only in case the amount of witnesses such small.


Why the amount is still just 12?  Apparently, users haven't made any significant changes in years... They couldn't have made much because of the rules.


Who decides whether it would be a shitty witness or normal? The answer is: majority of old witnesses which too small bunch of people to be a real majority. Let's just imagine that I was right about dev related to the witnesses. That would've meant he might decide every stability point to be shitty or not... That's too high risk to me.


From WP (page 21):
"1. best parent is selected only among parents whose witness list has no more
than 1 mutation;
2. there should be no more than 1 mutation relative to the witness list of the
last ball unit;
3. there should be no more than 1 mutation relative to the witness lists of all
the unstable MC units up to the last ball unit;
4. the stability point advances only when the current witnesses (as defined in
the current stability point) post enough units after the current stability
point.

These rules are designed to protect against malicious and accidental forks. At the same time, they imply that any changes of the predominant witness list have to be gradual, and each step has to be approved by the majority of the current witnesses."

Each transactions having 12 witnesses attached to it is consistent. What is not consistent would be having variable amount of witnesses on each transaction.

There can be more than 12 at the same time, but each transaction need to have 12 witnesses. That's consistent.

If you don't have identified witnesses then you could Sybil attack it.

1. No, the network launched with all witnesses belonging to founder, in past 2-3 years, there has been interviews, polls and users have replaced majority of witnesses with new ones. Witnesses have followed the changes that users have made.

2. The only power that witnesses have is stopping the network from finding the stability point when majority of witnesses collude and turn off the node, but even then, the community can hard-fork. There is no real power for witnesses (they can't censor) and the commissions barely cover fees for new transactions. Yes, it is interest of all witnesses that majority of witnesses keep running their nodes and if some won't, majority of witnesses or even users can signal to replace a shitty witness.

3. Not sure what you are talking about, each time there was a witness change, all the witnesses and users changed their witness lists, but actually only majority is needed. There is years of public DAG data that Obyte consensus mechanism works.

Thanks, I've read it, as well. Unfortunately, I'm not satisfied by that, though.

This one is the vital and understandable requirement and there is no questions with it.

This is the really weird one... There is no consistency at all. imho
If users are that bounded to check they're not going to be able to get the identity of 13 or 14 or 50 witnesses? That's not true, from my point of view, until I haven't got any evidence of vice versa stuff. And, I guess, so much people really don't care if those witnesses are identified or not.

For some reasons, actually.
1. All of those 12 were declared by dev at the time of the system was first published (WP) years ago, so it's not a tough issue to realize them to be somehow related to one another.
2. It's described the case of attack in the WP as the only majority of witnesses are able to distinguish the main chain using their data. Certainly, we're talking about just 12 people or organizations somehow interested in keeping the network healthy. And they're have the real power. And they're interested in getting commissions, as well.
3. There is the rule of the protocol saying the new witnesses might be added just after old 12 witnesses have approved the new one. There is no way they would have done that. They haven't any reason for that. But they've got all the reasons not to do it.

"Rule 4. the stability point advances only when the current witnesses (as defined in
the current stability point) post enough units after the current stability
point"
These rules are designed to protect against malicious and accidental forks. At the same time, they imply that any changes of the predominant witness list have to be gradual, and each step has to be approved by the majority of the current witnesses. (From the WP, page 21)

That's my personal opinion. I can see a lot of risks... Because protectability of the network depends on 12 people.

---
So, please, answer my question about the other real reason in case I was wrong.
Thanks

What makes it look completely centralized? Order Providers are not block producers or any other way between you and the network.

There is an explanation on Medium blog article, I recommend to read it:

Oh, thank you very much.. I've learned a lot about how it works from WP and from source code trying to find out. And I sourced no one real cause why they're just up to 12, despite all the explanations over there. Technically, there is definitely no one fundamental restrictions about it, especially, in terms of consensus. So, my question was what that magical restriction rule of 12 for? Why they're not 120 or not 1200?

According to the White Paper, the majority of witnesses has the power over the network. And if there are just 12, it looks like completely centralized, doesn't it? Please, correct me in case I was wrong explaining at least one real cause of that design.

I'm not against it, I'm just trying to figure it out from risk perspectives of using. Personally, it doesn't really matter to me if those 12 public decent persons or not.  

GB2 is a 2x leverage of GBYTE/USD price. If the price is going up in bull market, more people are likely to buy more GB2 to profit from the increase of price.
The more GB2 is bought/issued the more valuable GRGB becomes (it's price is calculated by the bonding curve) and more fees go to both capacitors (because buying GB2 right now would make the peg even worse).

The fast capacitor is emptied only when somebody fully fixes the peg. Partially making the peg more closer only takes some of the reward from fast capacitor. There is also reward in slow capacitor, from which 10% can be moved to fast capacitor.

What can gives the initiative to buy governance tokens in such cases?
Capacitor may be almost empty when all previous trades was near optimal (bought/sold both t1&t2 with target proportion).

That's how the consensus is designed to work. Those Witnesses (Order Providers) doesn't quite work as other witnesses on other blockchains.
Order Providers are just like any other full node on Obyte DAG, but they have pledged to run a wallet software that makes transactions in certain interval (depending on activity of others).
All they basically do is that they show, which direction the Directed Acyclic Graph should be growing.

You can read more about it there:
https://medium.com/obyte/dag/home
https://medium.com/obyte/from-blockchain-to-dag-getting-rid-of-middlemen-28afa7563545
https://obyte.org/technology
https://obyte.org/technology/order-providers

Hey there.
Could anybody explain why there are just 12 witnesses or about within the network?
Thx

Nope, there was just pump in price, so there is now opportunity to fix the peg by buying more governance tokens or redeeming GB2 tokens. Fixing the peg would give reward from fast capacitor.

Hmm...  Last three days GB2 price is 15-30% below target. Is it broken?

https://medium.com/obyte/ask-not-what-obyte-can-do-for-you-ask-what-you-can-do-for-obyte-7c07b30d894c

I'd love to help. however I find this whole Stablecoin thing way to complicated. If I can in anyway help please DM me.

Seems just that some speculators noticed that Bitcoin is pumping, so they just dumped their GBYTE to get more GBYTE.

The amount of GBYTE on Bittrex hot wallet has not changed much recently, so these probably were never Obyte users anyways.

Bonded Stablecoins and Oswap.io liquidity distribution on the other hand has more and more GBYTE value deposited.

Just more people needs to know about it, it's difficult to reach new people when nobody cares to spread the news.