4 ms·

The idea is that bitcoin is decentralized. But common sense tells you that by having an append only system, eventually you will need a supercomputer to deal wit

The idea is that bitcoin is decentralized. But common sense tells you that by having an append only system, eventually you will need a supercomputer to deal with the blockchain. Naturally, the blockchain will have to be hierarchical, or find ways to compact it.

zhte415 9y ago
Cash is decentralised.
- gt_ 9y ago
  In America, cash production is only legally done by one centralized entity, and it is done by them on a regular basis. As far as I know, this holds true in all developed nations.
  - partycoder 9y ago
    However, let's say I run a dollar store. Can I afford to do an exhaustive verification for every dollar I receive? maybe not. There are security features built into each bill that help me establishing if it's legit or not, that I can do myself without third parties.
  - eecc 9y ago
    Banks are allowed to emit debt at the condition that a fraction of it is backed by some liquidity. Banks decide where and when to create this debt so yes, money production is indeed decentralized. If you’re referring to actual coinage and notes then yes, facilities are indeed centralized, but with electronic debit cards the lines are blurred further
asddkk 9y ago
This has always seemed to me to be the elephant in the room to me. I brought this up with some ethereum devs at one point in a forum a year or two ago, and they addressed the questions pretty openly and graciously, but I still wonder about it. Essentially, they talked about the presumption of branches, subchains, etc. When that happens it seems the system would be a little more complicated than the classic crypto-libertarian currency model. These chains get pretty large, and the overhead in terms of time and storage space seems well beyond what most people are used to now. Maybe people just have to get used to 15 minute payments, and having an extra drive or computer for financial transactions, as the cost of decentralized finance and economics, but from the current vantage point it seems burdensome to me, and only more so as adoption would grow.
- garmaine 9y ago
  Lightning allows instantaneous and cheap payments without the need to hit the chain. Mimblewimble allows for pruning of block chain history down to a few dozen bytes per historical transaction in the limit. If it is still an issue then at some point we could decide to checkpoint the UTXO set of a block sufficiently buried in the past (say, 1+ year) and just sync from there.
  - tom_mellior 9y ago
    > Lightning allows instantaneous and cheap payments without the need to hit the chain. But a transaction to set up a channel needs to hit the chain first. What's the intended workflow for occasionally (once every few weeks) instantaneously ordering pizza via Lightning? Would I have to set up a very long lived channel with my preferred pizza place in advance?
    - garmaine 9y ago
      No no no, that's just payment channels. Lightning is multi-hop payment protocol on top of the traditional channel idea. Alice has a channel with Bob, and Bob has a channel with the Pizza company, and the Pizza company has a channel with its employees for salary and tips. On delivery, Alice pays Bob, Bob pays the Pizza company, and the Pizza company keeps some and forwards on just Alice's tip to the delivery guy. This is coordinated and setup in less than a second of communication among the parties along the route (meaning Bob and the Pizza company need to have their nodes online; but there can be non-routable offline signers at the endpoints; the Lightning routing is an online protocol however), but the payment is atomic -- intermediaries like Bob only get their money if they forward it on (minus a reasonable micro-fee), and either all channel updates happen or none at all. It only takes about the same amount of time as a point of sale credit card transaction to setup and sign, and completion of the atomic signing step is finality of settlement, so long as the parties remain online to check for preemptive or fraudulent channel closure. Lightning allows sub-second transactions with finality of settlement for any routable path along a peer-to-peer network of payment channels. Because of the six degrees of Kevin Bacon idea, anyone can pay anyone else on the network with a dozen or so different hops in the worst case, each hop taking a really small fee, like a hundredth of a percent. A true peer-to-peer electronic cash system, if you will. As in this example, your long-lived channels as an individual are likely to be with your employer or payroll company, your mortgage company, your telco, your favorite restaurants, and other other regular payments you make or sources of income. As an organization your channels will probably be with your suppliers (accounts payable) and major customers (accounts receivable), as well as payroll and office supply companies, etc. But you're able to reach everyone on the network through these starting points. This allows you to do things like proactively "pull" in funds from accounts receivable to cover expenses, by routing through that part of the network. It maps pretty well onto both the individual and corporate use cases. It also allows micropayment in bitcoin again.. the cost of a payment is basically the cost of maintaining these online routable nodes, and the message communication, which is probably sub-cents in the limit. It's possible you can, e.g., do a 5-cent payment per article view on a news website, or per song played on a future Spotify, etc.
moondev 9y ago
Dosen't the blockchain already compact? Correct me if i'm wrong but once a block is hashed and confirmed as the most recent block, why would you need to store the previous blocks?
- snodnipper 9y ago
  https://en.bitcoin.it/wiki/Full_node https://en.bitcoin.it/wiki/Full_node "Full nodes are the most secure way to use Bitcoin, they do not suffer from many attacks that affect lightweight wallets...[a]s explained previously, full nodes enforce the consensus rules no matter what. However, lightweight nodes do not do this."
- nikkwong 9y ago
  Full nodes must store the hashes that are generated from previous blocks. If they didn't, how would you know that the most recent hash really is the generated as a result of hashing the previous block recursively? A malicious attacker could forge a previous block, act as the next one is the current block based on the hash of the forged block, and there would be no way to verify that the hash of this block is valid or invalid.
  - moondev 9y ago
    I was under the impression that attack would not work because as long as the majority of nodes are "honest" there would be no way for an attacking node to outpace the chain. Each node votes with their cpu so the combined compute/speed of honest nodes would always outpace the introduction of a corrupt block. Since the the most recent "confirmed block" is a hash of the previous successful block it just continues. However I admit much of my assumptions are just from reading the original paper so i'm probably over-simplifying. I guess it would make sense they need some history but not sure why they would need all of it since the previous blocks are confirmed?
- SeoxyS 9y ago
  "Balances" are stored on every block. If you threw away the beginning of the chain, you could keep generating blocks, but you couldn't spend any transaction outputs that are stored in the thrown away parts of the chain. Unless I'm missing something something pretty big about how Bitcoin's blockchain works.
snodnipper 9y ago
pretty sure there are parallels with DNS. If BTC is the root of trust networks then that is probably enough for one protocol - and good enough to survive. Root node back others networks (think top level domains and delegated subdomains). IMO, it needs to be _very_ expensive to keep minimal blockchain size and avoid spam.
mungojelly 9y ago
Common sense does tell you that, but knowledge of computer science tells you otherwise (and of course where it contradicts common sense is the only place science is useful). The appending grows linearly, the price of storage falls exponentially, so storage wins in the end (regardless of the multipliers). Too bad really, as it would be more interesting if chains got too heavy and we had to constantly switch to new ones.
Tepix 9y ago
This is a mostly solved issue.

zhte415 9y ago

Cash is decentralised.

gt_ 9y ago
In America, cash production is only legally done by one centralized entity, and it is done by them on a regular basis. As far as I know, this holds true in all developed nations.
- partycoder 9y ago
  However, let's say I run a dollar store. Can I afford to do an exhaustive verification for every dollar I receive? maybe not. There are security features built into each bill that help me establishing if it's legit or not, that I can do myself without third parties.
- eecc 9y ago
  Banks are allowed to emit debt at the condition that a fraction of it is backed by some liquidity. Banks decide where and when to create this debt so yes, money production is indeed decentralized. If you’re referring to actual coinage and notes then yes, facilities are indeed centralized, but with electronic debit cards the lines are blurred further

asddkk 9y ago

This has always seemed to me to be the elephant in the room to me. I brought this up with some ethereum devs at one point in a forum a year or two ago, and they addressed the questions pretty openly and graciously, but I still wonder about it. Essentially, they talked about the presumption of branches, subchains, etc. When that happens it seems the system would be a little more complicated than the classic crypto-libertarian currency model. These chains get pretty large, and the overhead in terms of time and storage space seems well beyond what most people are used to now. Maybe people just have to get used to 15 minute payments, and having an extra drive or computer for financial transactions, as the cost of decentralized finance and economics, but from the current vantage point it seems burdensome to me, and only more so as adoption would grow.

garmaine 9y ago
Lightning allows instantaneous and cheap payments without the need to hit the chain. Mimblewimble allows for pruning of block chain history down to a few dozen bytes per historical transaction in the limit. If it is still an issue then at some point we could decide to checkpoint the UTXO set of a block sufficiently buried in the past (say, 1+ year) and just sync from there.
- tom_mellior 9y ago
  > Lightning allows instantaneous and cheap payments without the need to hit the chain. But a transaction to set up a channel needs to hit the chain first. What's the intended workflow for occasionally (once every few weeks) instantaneously ordering pizza via Lightning? Would I have to set up a very long lived channel with my preferred pizza place in advance?
  - garmaine 9y ago
    No no no, that's just payment channels. Lightning is multi-hop payment protocol on top of the traditional channel idea. Alice has a channel with Bob, and Bob has a channel with the Pizza company, and the Pizza company has a channel with its employees for salary and tips. On delivery, Alice pays Bob, Bob pays the Pizza company, and the Pizza company keeps some and forwards on just Alice's tip to the delivery guy. This is coordinated and setup in less than a second of communication among the parties along the route (meaning Bob and the Pizza company need to have their nodes online; but there can be non-routable offline signers at the endpoints; the Lightning routing is an online protocol however), but the payment is atomic -- intermediaries like Bob only get their money if they forward it on (minus a reasonable micro-fee), and either all channel updates happen or none at all. It only takes about the same amount of time as a point of sale credit card transaction to setup and sign, and completion of the atomic signing step is finality of settlement, so long as the parties remain online to check for preemptive or fraudulent channel closure. Lightning allows sub-second transactions with finality of settlement for any routable path along a peer-to-peer network of payment channels. Because of the six degrees of Kevin Bacon idea, anyone can pay anyone else on the network with a dozen or so different hops in the worst case, each hop taking a really small fee, like a hundredth of a percent. A true peer-to-peer electronic cash system, if you will. As in this example, your long-lived channels as an individual are likely to be with your employer or payroll company, your mortgage company, your telco, your favorite restaurants, and other other regular payments you make or sources of income. As an organization your channels will probably be with your suppliers (accounts payable) and major customers (accounts receivable), as well as payroll and office supply companies, etc. But you're able to reach everyone on the network through these starting points. This allows you to do things like proactively "pull" in funds from accounts receivable to cover expenses, by routing through that part of the network. It maps pretty well onto both the individual and corporate use cases. It also allows micropayment in bitcoin again.. the cost of a payment is basically the cost of maintaining these online routable nodes, and the message communication, which is probably sub-cents in the limit. It's possible you can, e.g., do a 5-cent payment per article view on a news website, or per song played on a future Spotify, etc.

moondev 9y ago

Dosen't the blockchain already compact? Correct me if i'm wrong but once a block is hashed and confirmed as the most recent block, why would you need to store the previous blocks?

snodnipper 9y ago
https://en.bitcoin.it/wiki/Full_node https://en.bitcoin.it/wiki/Full_node "Full nodes are the most secure way to use Bitcoin, they do not suffer from many attacks that affect lightweight wallets...[a]s explained previously, full nodes enforce the consensus rules no matter what. However, lightweight nodes do not do this."
nikkwong 9y ago
Full nodes must store the hashes that are generated from previous blocks. If they didn't, how would you know that the most recent hash really is the generated as a result of hashing the previous block recursively? A malicious attacker could forge a previous block, act as the next one is the current block based on the hash of the forged block, and there would be no way to verify that the hash of this block is valid or invalid.
- moondev 9y ago
  I was under the impression that attack would not work because as long as the majority of nodes are "honest" there would be no way for an attacking node to outpace the chain. Each node votes with their cpu so the combined compute/speed of honest nodes would always outpace the introduction of a corrupt block. Since the the most recent "confirmed block" is a hash of the previous successful block it just continues. However I admit much of my assumptions are just from reading the original paper so i'm probably over-simplifying. I guess it would make sense they need some history but not sure why they would need all of it since the previous blocks are confirmed?
SeoxyS 9y ago
"Balances" are stored on every block. If you threw away the beginning of the chain, you could keep generating blocks, but you couldn't spend any transaction outputs that are stored in the thrown away parts of the chain. Unless I'm missing something something pretty big about how Bitcoin's blockchain works.

snodnipper 9y ago

pretty sure there are parallels with DNS. If BTC is the root of trust networks then that is probably enough for one protocol - and good enough to survive. Root node back others networks (think top level domains and delegated subdomains). IMO, it needs to be _very_ expensive to keep minimal blockchain size and avoid spam.

mungojelly 9y ago

Common sense does tell you that, but knowledge of computer science tells you otherwise (and of course where it contradicts common sense is the only place science is useful). The appending grows linearly, the price of storage falls exponentially, so storage wins in the end (regardless of the multipliers). Too bad really, as it would be more interesting if chains got too heavy and we had to constantly switch to new ones.

Tepix 9y ago

This is a mostly solved issue.