kjn311
Member
Offline
Activity: 319
Merit: 10
|
|
November 20, 2014, 06:56:42 PM |
|
From what I understand, and the reason I mentioned it here, is that tendermint confirmations times are limited only by the bandwidth of the network, for example, with about 10,000 validators = about 1min confirmations times and also unlike dpos doesn't rely on having trust in the 101 delegates. "On the other hand existing proof-of-stake protocols do not have a well defined intrinsic penalty for instigators of a double-spend attack. This is commonly called, ironically, the “nothing at stake” problem. Newer protocols like the BitShares delegated-proof-of-stake protocol attempt to address this problem by placing the role of ranked-delegate at stake [3], but security is dependant on the extrinsic ability of stakeholders to accurately predict the future performance of delegates." Reference on confirmation time about 19:00 mark http://letstalkbitcoin.com/blog/post/beyond-bitcoin-21-consensus-without-miningDirectly from the developer: Hi Keith, the conf time depends on the available bandwidth of each node and the number of validators. With say a 10Mbit connection for everyone and 5000 validators it should get sub minute confirmations.
|
|
|
|
jaekwon
Member
Offline
Activity: 70
Merit: 10
|
|
November 20, 2014, 07:40:52 PM |
|
what I don't like about tendermint:
"If the validator causes the blockchain to fork while its coins are locked in bond, all of its coins are destroyed."
so if the developers make some mistakes, or I have some network problems or whatever and it forks, I loose the coins. Wow!
That's a valid concern, though one that isn't impossible to fix with proper technology. A modification to the protocol could be made to only destroy a fraction of the bonded coins (and the guaranteed security would likewise be reduced), but I don't mention that in the whitepaper because it doesn't help clarify the algorithm.
|
|
|
|
ohad
|
|
November 20, 2014, 08:39:18 PM |
|
To be honest, I also reconsider plain old POW, counting on the ASICs not to bite from zennet's computing resources. I talked with HMC about his "kill switch" and he explained that in order to restart the chain, one would have to invest the resources needed to create the original chain. Of course, it's much more difficult on BTC-like POW.
|
Tau-Chain & Agoras
|
|
|
kjn311
Member
Offline
Activity: 319
Merit: 10
|
|
November 20, 2014, 09:06:40 PM Last edit: November 20, 2014, 09:22:35 PM by kjn311 |
|
To be honest, I also reconsider plain old POW, counting on the ASICs not to bite from zennet's computing resources. I talked with HMC about his "kill switch" and he explained that in order to restart the chain, one would have to invest the resources needed to create the original chain. Of course, it's much more difficult on BTC-like POW.
I don't mind POW but I thought the concern was confirmation times?. Personally, I would like to see Zennet as the most energy efficient super computer in the world but that decision is ultimately yours. As long as the cost of selling my computer resources can also cover the cost of POW I guess it would be fine? I do hope that POW will usher in new renewable energy ideas before the next BTC halving in 2 years but that's another topic.
|
|
|
|
ohad
|
|
November 20, 2014, 09:37:00 PM |
|
I don't mind POW but I thought the concern was confirmation times?.
POW can be implemented with some new advancements (cf Ethereum, GHOST) and even disregarding those advancements, block time can be decreased. It does of course increase the probability of double-spend, but the math has to be done again when considering the micropayment nature, and looking at the expectation of the loss, rather then the probability. And, as above, considering new POW-based schemes. Personally, I would like to see Zennet as the most energy efficient super computer in the world but that decision is ultimately yours.
* we're altogether. that's the spirit. i strongly encourage deep discussion here in order to study and do the right thing, and there are many people here knowing better than i do, especially on the cryptocurrency algos field. As long as the cost of selling my computer resources can also cover the cost of POW I guess it would be fine?
i currently count on the tx fees. no coin generation (as i explained why here). again, math has to be done. note that ASICs will do the work and not take from zennet's computational power.
|
Tau-Chain & Agoras
|
|
|
kjn311
Member
Offline
Activity: 319
Merit: 10
|
|
November 20, 2014, 10:03:46 PM |
|
Well my vote is POS. POS is the answer and will replace POW long term. POS is the more distributed model that I feel is perfect for this project. Just not dpos.
|
|
|
|
ohad
|
|
November 20, 2014, 10:15:12 PM |
|
Well my vote is POS. POS is the answer and will replace POW long term. POS is the more distributed model that I feel is perfect for this project. Just not dpos.
which specific kind of POS? there's a whole zoo
|
Tau-Chain & Agoras
|
|
|
kjn311
Member
Offline
Activity: 319
Merit: 10
|
|
November 20, 2014, 10:37:03 PM |
|
Well my vote is POS. POS is the answer and will replace POW long term. POS is the more distributed model that I feel is perfect for this project. Just not dpos.
which specific kind of POS? there's a whole zoo That's the million dollar question. Honestly, I would message Jaekwon. He is more of the expert on the matter.
|
|
|
|
ohad
|
|
November 20, 2014, 10:40:59 PM |
|
Well my vote is POS. POS is the answer and will replace POW long term. POS is the more distributed model that I feel is perfect for this project. Just not dpos.
which specific kind of POS? there's a whole zoo That's the million dollar question. Honestly, I would message Jaekwon. He is more of the expert on the matter. if I had to choose between POS and POW or hybrid, I'd choose pure POW as for what I know and heard so far. extreme cases have to be mapped. risk expectation has to be calculated. any such researches on POS? not afaik
|
Tau-Chain & Agoras
|
|
|
Hueristic
Legendary
Offline
Activity: 4032
Merit: 5589
Doomed to see the future and unable to prevent it
|
|
November 20, 2014, 10:46:23 PM |
|
Well my vote is POS. POS is the answer and will replace POW long term. POS is the more distributed model that I feel is perfect for this project. Just not dpos.
which specific kind of POS? there's a whole zoo That's the million dollar question. Honestly, I would message Jaekwon. He is more of the expert on the matter. if I had to choose between POS and POW or hybrid, I'd choose pure POW as for what I know and heard so far. extreme cases have to be mapped. risk expectation has to be calculated. any such researches on POS? not afaik I thought the clients supplying the Cycles would be securing the network? And you expect them to also do POW? I guess I'm not understanding this Idea at all.
|
“Bad men need nothing more to compass their ends, than that good men should look on and do nothing.”
|
|
|
ohad
|
|
November 20, 2014, 10:56:06 PM |
|
Well my vote is POS. POS is the answer and will replace POW long term. POS is the more distributed model that I feel is perfect for this project. Just not dpos.
which specific kind of POS? there's a whole zoo That's the million dollar question. Honestly, I would message Jaekwon. He is more of the expert on the matter. if I had to choose between POS and POW or hybrid, I'd choose pure POW as for what I know and heard so far. extreme cases have to be mapped. risk expectation has to be calculated. any such researches on POS? not afaik I thought the clients supplying the Cycles would be securing the network? And you expect them to also do POW? I guess I'm not understanding this Idea at all. Here you go sir go read about it
|
Tau-Chain & Agoras
|
|
|
Hueristic
Legendary
Offline
Activity: 4032
Merit: 5589
Doomed to see the future and unable to prevent it
|
|
November 20, 2014, 11:14:45 PM |
|
Well my vote is POS. POS is the answer and will replace POW long term. POS is the more distributed model that I feel is perfect for this project. Just not dpos.
which specific kind of POS? there's a whole zoo That's the million dollar question. Honestly, I would message Jaekwon. He is more of the expert on the matter. if I had to choose between POS and POW or hybrid, I'd choose pure POW as for what I know and heard so far. extreme cases have to be mapped. risk expectation has to be calculated. any such researches on POS? not afaik I thought the clients supplying the Cycles would be securing the network? And you expect them to also do POW? I guess I'm not understanding this Idea at all. Here you go sir go read about it Sorry, that is what I took from this. A Provider in search of employment looks for these announcements on the blockchain. XenCoin is not a currency. It is not meant to be a method of payment used in everyday business transactions. Its purpose and design goal are to be a token for activating computational machines.
|
“Bad men need nothing more to compass their ends, than that good men should look on and do nothing.”
|
|
|
kjn311
Member
Offline
Activity: 319
Merit: 10
|
|
November 21, 2014, 12:23:19 AM |
|
Well my vote is POS. POS is the answer and will replace POW long term. POS is the more distributed model that I feel is perfect for this project. Just not dpos.
which specific kind of POS? there's a whole zoo Thank me later. http://bravenewcoin.com/assets/Uploads/TransactionsAsProofOfStake10.pdfI think the tendermint algorithm should be modified in away that utilizes the above white paper.
|
|
|
|
Djinou94
Legendary
Offline
Activity: 1036
Merit: 1000
|
|
November 21, 2014, 01:36:36 AM |
|
When the presale open per favor?
|
|
|
|
ohad
|
|
November 21, 2014, 01:22:35 PM |
|
Well my vote is POS. POS is the answer and will replace POW long term. POS is the more distributed model that I feel is perfect for this project. Just not dpos.
which specific kind of POS? there's a whole zoo Thank me later. http://bravenewcoin.com/assets/Uploads/TransactionsAsProofOfStake10.pdfI think the tendermint algorithm should be modified in away that utilizes the above white paper. See DPOS docs for TaPOS vulnerabilities..
|
Tau-Chain & Agoras
|
|
|
|
mishax1
Legendary
Offline
Activity: 2898
Merit: 1017
|
|
November 25, 2014, 12:02:42 PM |
|
Shit, this makes me wanna invest my bitcoin in it
|
|
|
|
DrGrid
Member
Offline
Activity: 101
Merit: 10
|
|
November 27, 2014, 03:21:40 AM |
|
So I finally read through the entire thread, Hunter Miner Crafter sure posts the longest explanations I have seen for quite some time. The discussion was very good though and will be following it on the irc more closely, now that I have become conscious of it. Basically, I have three questions. I tend to post in a brief manner, I can expand though, if needed. If another protocol besides DPOS will prove its utility until Q3 2015 (if I remember correctly this was the launch date for the core client), will you be willing to adapt to it, even if this would cause a delay (I am particularly eyeing at slasher here), or is your protocol based on a certain aspect that only DPOS can provide? Secondly, I would like to ask if Little Duke's IDCoin proposal would be feasible for the trust system. Lastly I see that you use checksums in the current system to prove that the data still exists. How are checksums more effective and secure than Filecoin's "challenge" system?
|
Bitrated user: DrGrid.
|
|
|
ohad
|
|
November 27, 2014, 01:32:55 PM |
|
So I finally read through the entire thread, Hunter Miner Crafter sure posts the longest explanations I have seen for quite some time. The discussion was very good though and will be following it on the irc more closely, now that I have become conscious of it. Basically, I have three questions. I tend to post in a brief manner, I can expand though, if needed. If another protocol besides DPOS will prove its utility until Q3 2015 (if I remember correctly this was the launch date for the core client), will you be willing to adapt to it, even if this would cause a delay (I am particularly eyeing at slasher here), or is your protocol based on a certain aspect that only DPOS can provide?
the plan is to adopt new algos if they come. see last comments where I even think of POW rather DPOS. Secondly, I would like to ask if Little Duke's IDCoin proposal would be feasible for the trust system.
i dont know what you're talking about. please provide a link. Lastly I see that you use checksums in the current system to prove that the data still exists. How are checksums more effective and secure than Filecoin's "challenge" system?
ZenFS design is old. the current thoughts are to make it Tahoe-LAFS based (yet torrent compatible).
|
Tau-Chain & Agoras
|
|
|
|
|