This paper was already discussed extensively in this subforum: https://bitcointalksearch.org/topic/on-the-longest-chain-rule-and-programmed-self-destruction-of-crypto-currencies-600436
It exhibits a number of serious technical misunderstandings of the Bitcoin system which moot some of its arguments.
In the future please search before posting.
Topic: Programmed self-destruction and how to prevent it (Read 1088 times)
Well I haven't read that whole paper, but I did post this thread https://bitcointalksearch.org/topic/m.7388422 regarding the issue of how few full nodes there are and how that represents a vulnerability. So instead getting hash power you can gain effective control of bitcoin by controlling enough nodes.
As you (or the paper) mentions, without a node incentive, the number of full non-mining nodes running will likely decrease over time.
As you (or the paper) mentions, without a node incentive, the number of full non-mining nodes running will likely decrease over time.
Dear fellows!
I have read this http://arxiv.org/pdf/1405.0534.pdf , which was not a merry read, and despite the fact that i am an absolute fan of the blockchain technology, the distributed network idea, the empowerment of the individual, and almost full in BTC - financially and carrier wise, i am rather alarmed.
Even if the above paper has some bad arguments from time to time, the underlying ideas seems to be correct which are (TLDR):
1. centralization of mining through ASIC, and it's effect on entry level users, broader adaptation,
2. pooled mining, and the arise of need to trust on pool operators and "big money miners",
3. how diminishing return of block rewards will threaten the system - there is a case study analogy of Unobtanium, and Doge/LTC -,
4. the non-incentive way of running full nodes - hence their numbers begun to drop, again lowering security and decentralization,
5. the Longest Chain Rule and it's consequences on transaction speed (transactions vs. block creation, should require 2 different methods from an engineering standpoint).
These aforementioned arguments conclude the self destructing behavior of the whole network, that is the paper all about.
My thoughts:
Mining.
So, while from an economic stand point i love the idea of deflation monetary policy of bitcoin, from a (human) behavioral angle it is indeed not good for mining, follows:
- after awhile, only the most expensive and specialized machinery will be profitable to run, if ever to make a RoI, which is NOT the average user, hence the emerge of big cloud mining operations!
- this is not the original idea of "1 cpu, 1 vote", and goes against decentralization and network neutrality! Entry level mining - and so the incentive to adopt the network - would be impossible for the average user.
- at the end, only big corporations with huge capital (banks? o.0) - whom are tied to physical and known places, under a known jurisdiction!!! - will mining, and we have our central point(s) of failure situation again!
My questions to the way more "code minded" people :-)
- I am almost sure that ASIC mining has to go. It is a huge issue, given the capital brought into it. But what than? Is it possible, that the nodes themselves, given they are fully connected to the network (8+ connections) would create the new blocks, not using anything else, but their home PC (and yes, company bot nets by employees
), lets say CPU power? Is that "codeable", to keep the encryption, and preventing specialized hardware to emerge at the same time? How this will effect the block generation speed? The reward dispensation, the intentional inability to form pools?
- i understand, that this might be a radical, and or impossible to do proposition.
- this would also require the blockchain to be smaller, or reformed, the bloating is enormous. It might help fungibility, to loose old records, hence no coin would record "taint" in the long run!
- how about a negative incentive, an avoidance of cost reward system? Like: there is a built in transaction fee (dynamic with volumes!), but if you are running your node, you "collect" free transaction tokens/volume.
- the transaction fees either get "destroyed" -> deflation, which could be balanced out by nodes "mining" new blocks -> inflation. This is the seeming erasion of the hard cap!! in a controlled way, there will be always new coins, in the same rate as they gets destroyed by transaction fees. (so block reward would be a floating number).
- or they redistributed along the running nodes based on a rule (the hard cap can stay).
- i understand that peercoin has a hybrid PoW/PoS method, but it still has the huge diminishing return of block rewards, which causes hashing power to drop fasterer, than PoS system builds up, and or distribution gets screwed up (see the paper for data).
TLDR:
Separating miners and "average currency users" undermines - pun intended:) the idea of network neutrality, ALL participants who run a full node with the full blockchain should be automatically be "miners", hence an average home computer could contribute and get rewarded, the entry level SHOULD be for the average Joe. Without this, we are going to see central points of failure, and disinterest from the general user of the currency, for they are can not be real contributors - and lack "voting power" - on the networking!
I am waiting for suggestions, or explanations why this whole argument is wrong/correct
I have read this http://arxiv.org/pdf/1405.0534.pdf , which was not a merry read, and despite the fact that i am an absolute fan of the blockchain technology, the distributed network idea, the empowerment of the individual, and almost full in BTC - financially and carrier wise, i am rather alarmed.
Even if the above paper has some bad arguments from time to time, the underlying ideas seems to be correct which are (TLDR):
1. centralization of mining through ASIC, and it's effect on entry level users, broader adaptation,
2. pooled mining, and the arise of need to trust on pool operators and "big money miners",
3. how diminishing return of block rewards will threaten the system - there is a case study analogy of Unobtanium, and Doge/LTC -,
4. the non-incentive way of running full nodes - hence their numbers begun to drop, again lowering security and decentralization,
5. the Longest Chain Rule and it's consequences on transaction speed (transactions vs. block creation, should require 2 different methods from an engineering standpoint).
These aforementioned arguments conclude the self destructing behavior of the whole network, that is the paper all about.
My thoughts:
Mining.
So, while from an economic stand point i love the idea of deflation monetary policy of bitcoin, from a (human) behavioral angle it is indeed not good for mining, follows:
- after awhile, only the most expensive and specialized machinery will be profitable to run, if ever to make a RoI, which is NOT the average user, hence the emerge of big cloud mining operations!
- this is not the original idea of "1 cpu, 1 vote", and goes against decentralization and network neutrality! Entry level mining - and so the incentive to adopt the network - would be impossible for the average user.
- at the end, only big corporations with huge capital (banks? o.0) - whom are tied to physical and known places, under a known jurisdiction!!! - will mining, and we have our central point(s) of failure situation again!
My questions to the way more "code minded" people :-)
- I am almost sure that ASIC mining has to go. It is a huge issue, given the capital brought into it. But what than? Is it possible, that the nodes themselves, given they are fully connected to the network (8+ connections) would create the new blocks, not using anything else, but their home PC (and yes, company bot nets by employees
), lets say CPU power? Is that "codeable", to keep the encryption, and preventing specialized hardware to emerge at the same time? How this will effect the block generation speed? The reward dispensation, the intentional inability to form pools?- i understand, that this might be a radical, and or impossible to do proposition.
- this would also require the blockchain to be smaller, or reformed, the bloating is enormous. It might help fungibility, to loose old records, hence no coin would record "taint" in the long run!
- how about a negative incentive, an avoidance of cost reward system? Like: there is a built in transaction fee (dynamic with volumes!), but if you are running your node, you "collect" free transaction tokens/volume.
- the transaction fees either get "destroyed" -> deflation, which could be balanced out by nodes "mining" new blocks -> inflation. This is the seeming erasion of the hard cap!! in a controlled way, there will be always new coins, in the same rate as they gets destroyed by transaction fees. (so block reward would be a floating number).
- or they redistributed along the running nodes based on a rule (the hard cap can stay).
- i understand that peercoin has a hybrid PoW/PoS method, but it still has the huge diminishing return of block rewards, which causes hashing power to drop fasterer, than PoS system builds up, and or distribution gets screwed up (see the paper for data).
TLDR:
Separating miners and "average currency users" undermines - pun intended:) the idea of network neutrality, ALL participants who run a full node with the full blockchain should be automatically be "miners", hence an average home computer could contribute and get rewarded, the entry level SHOULD be for the average Joe. Without this, we are going to see central points of failure, and disinterest from the general user of the currency, for they are can not be real contributors - and lack "voting power" - on the networking!
I am waiting for suggestions, or explanations why this whole argument is wrong/correct
Jump to: