4 ms·

Transaction throughput can easily be increased a hundredfold right now, with a simple increase in block size. As far as I know, this is the plan, once it become

Transaction throughput can easily be increased a hundredfold right now, with a simple increase in block size. As far as I know, this is the plan, once it becomes necessary (for example due to Lightning adoption picking up). It is only out of principle that bitcoin developers have so far chosen to keep the block size limit extremely low (incidentally limiting transaction throughput too).

Majromax 6y ago
> with a simple increase in block size. "Simple" changes to the bitcoin protocol aren't simple because it's a coordination problem. As long as more than 50% of mining power is behind the older protocol, the longest block history will remain under that protocol. Nodes which can switch will have incentives against doing so, because mining an embiggened block that does not become part of the official history orphans the mining reward. From another perspective: any breaking change to the bitcoin protocol involves politely asking miners to execute a 51% attack on themselves.
- aeternum 6y ago
  This is true, but breaking changes, aka hard forks have occurred in the past and blocksize has already been successfully increased from 1MB to 2MB. It will all come down to whether miners think they can make more from having fewer higher fee txns vs. more lower fee txns in a block. My prediction is that more lower fee txns will win since competition from other cryptocurrencies will cap the txn fee that people are willing to pay.
  - 1996 6y ago
    > My prediction is that more lower fee txns will win since competition from other cryptocurrencies will cap the txn fee that people are willing to pay Exactly my analysis too.
rspeele 6y ago
100x the block size would mean the blockchain would grow by around 6 TB every year, which would make running a full node impractical for regular consumers. And it would still only have the capacity for everyone in China to do about one on-chain transaction per month. It's unclear to me what the financial incentive to run a full node is at all -- from what I'm reading, you don't need one to mine, and you don't get any reward just from running a node. So if it becomes expensive (20 TB hard drives, big pipe for bandwidth) to run a full node, who will want to do it? I suppose the exchanges would run some nodes, if only to keep the coins they trade in functioning.
- johnyzee 6y ago
  It was never the intention that regular people would run nodes. Satoshi envisioned that eventually nodes would run in large datacenters: https://bitcointalk.org/index.php?topic=532.msg6306#msg6306 https://bitcointalk.org/index.php?topic=532.msg6306#msg6306 "The current system where every user is a network node is not the intended configuration for large scale. That would be like every Usenet user runs their own NNTP server. The design supports letting users just be users. The more burden it is to run a node, the fewer nodes there will be. Those few nodes will be big server farms. The rest will be client nodes that only do transactions and don't generate." 6TB is absolutely nothing in terms of storage. That is close to a single consumer device today, notwithstanding that storage is dropping in price constantly. For datacenters you could increase this by orders of magnitude and nobody would bat an eyelid.
  - rspeele 6y ago
    But why? Sure, it doesn't cost that much to store terabytes in a data center (plus bandwidth to distribute the chain to other peers, validate new blocks etc). But any cost is something, considering there is no reward for running a node. AWS S3 standard charges $138/mo to store 6 TB of data. Who gets $138/mo of value out of running their own full node, rather than just piggybacking off the existing network? As I mentioned in my first post, exchanges are the only ones I can think of.
    - 6y ago

Majromax 6y ago

> with a simple increase in block size. "Simple" changes to the bitcoin protocol aren't simple because it's a coordination problem. As long as more than 50% of mining power is behind the older protocol, the longest block history will remain under that protocol. Nodes which can switch will have incentives against doing so, because mining an embiggened block that does not become part of the official history orphans the mining reward. From another perspective: any breaking change to the bitcoin protocol involves politely asking miners to execute a 51% attack on themselves.

aeternum 6y ago
This is true, but breaking changes, aka hard forks have occurred in the past and blocksize has already been successfully increased from 1MB to 2MB. It will all come down to whether miners think they can make more from having fewer higher fee txns vs. more lower fee txns in a block. My prediction is that more lower fee txns will win since competition from other cryptocurrencies will cap the txn fee that people are willing to pay.
- 1996 6y ago
  > My prediction is that more lower fee txns will win since competition from other cryptocurrencies will cap the txn fee that people are willing to pay Exactly my analysis too.

rspeele 6y ago

100x the block size would mean the blockchain would grow by around 6 TB every year, which would make running a full node impractical for regular consumers. And it would still only have the capacity for everyone in China to do about one on-chain transaction per month. It's unclear to me what the financial incentive to run a full node is at all -- from what I'm reading, you don't need one to mine, and you don't get any reward just from running a node. So if it becomes expensive (20 TB hard drives, big pipe for bandwidth) to run a full node, who will want to do it? I suppose the exchanges would run some nodes, if only to keep the coins they trade in functioning.

johnyzee 6y ago
It was never the intention that regular people would run nodes. Satoshi envisioned that eventually nodes would run in large datacenters: https://bitcointalk.org/index.php?topic=532.msg6306#msg6306 https://bitcointalk.org/index.php?topic=532.msg6306#msg6306 "The current system where every user is a network node is not the intended configuration for large scale. That would be like every Usenet user runs their own NNTP server. The design supports letting users just be users. The more burden it is to run a node, the fewer nodes there will be. Those few nodes will be big server farms. The rest will be client nodes that only do transactions and don't generate." 6TB is absolutely nothing in terms of storage. That is close to a single consumer device today, notwithstanding that storage is dropping in price constantly. For datacenters you could increase this by orders of magnitude and nobody would bat an eyelid.
- rspeele 6y ago
  But why? Sure, it doesn't cost that much to store terabytes in a data center (plus bandwidth to distribute the chain to other peers, validate new blocks etc). But any cost is something, considering there is no reward for running a node. AWS S3 standard charges $138/mo to store 6 TB of data. Who gets $138/mo of value out of running their own full node, rather than just piggybacking off the existing network? As I mentioned in my first post, exchanges are the only ones I can think of.
  - 6y ago